[llvm] c1b5e36 - [CostModel][X86] Add CostKinds handling for fcmp ops

Simon Pilgrim via llvm-commits llvm-commits at lists.llvm.org
Tue Sep 6 02:35:17 PDT 2022


Author: Simon Pilgrim
Date: 2022-09-06T10:34:53+01:00
New Revision: c1b5e36d74a3f604bcb55277e22c39f18fc637ce

URL: https://github.com/llvm/llvm-project/commit/c1b5e36d74a3f604bcb55277e22c39f18fc637ce
DIFF: https://github.com/llvm/llvm-project/commit/c1b5e36d74a3f604bcb55277e22c39f18fc637ce.diff

LOG: [CostModel][X86] Add CostKinds handling for fcmp ops

This was achieved with an updated version of the 'cost-tables vs llvm-mca' script D103695 (although it still struggles with avx512 predicate numbers which had to be done manually)

SSE numbers are still too low for FCMP_ONE/FCMP_UEQ cases which expand to a more complex sequence than the existing 'ExtraCost' system can manage.

Added: 
    

Modified: 
    llvm/lib/Target/X86/X86TargetTransformInfo.cpp
    llvm/test/Analysis/CostModel/X86/fcmp-codesize.ll
    llvm/test/Analysis/CostModel/X86/fcmp-latency.ll
    llvm/test/Analysis/CostModel/X86/fcmp-sizelatency.ll
    llvm/test/Analysis/CostModel/X86/fcmp.ll
    llvm/test/Analysis/CostModel/X86/fptoi_sat.ll

Removed: 
    


################################################################################
diff  --git a/llvm/lib/Target/X86/X86TargetTransformInfo.cpp b/llvm/lib/Target/X86/X86TargetTransformInfo.cpp
index 8b2b5471edd46..8b737b6b8e649 100644
--- a/llvm/lib/Target/X86/X86TargetTransformInfo.cpp
+++ b/llvm/lib/Target/X86/X86TargetTransformInfo.cpp
@@ -2867,6 +2867,7 @@ InstructionCost X86TTIImpl::getCmpSelInstrCost(unsigned Opcode, Type *ValTy,
                 Pred == CmpInst::BAD_FCMP_PREDICATE))
         Pred = cast<CmpInst>(I)->getPredicate();
 
+      // TODO: Handle pre-AVX FCMP_ONE/FCMP_UEQ slow cases.
       switch (Pred) {
       case CmpInst::Predicate::ICMP_NE:
         // xor(cmpeq(x,y),-1)
@@ -2927,10 +2928,13 @@ InstructionCost X86TTIImpl::getCmpSelInstrCost(unsigned Opcode, Type *ValTy,
   };
 
   static const CostKindTblEntry AVX512CostTbl[] = {
+    { ISD::SETCC,   MVT::v8f64,   { 1, 4, 1, 1 } },
+    { ISD::SETCC,   MVT::v4f64,   { 1, 4, 1, 1 } },
+    { ISD::SETCC,   MVT::v16f32,  { 1, 4, 1, 1 } },
+    { ISD::SETCC,   MVT::v8f32,   { 1, 4, 1, 1 } },
+
     { ISD::SETCC,   MVT::v8i64,   { 1 } },
     { ISD::SETCC,   MVT::v16i32,  { 1 } },
-    { ISD::SETCC,   MVT::v8f64,   { 1 } },
-    { ISD::SETCC,   MVT::v16f32,  { 1 } },
 
     { ISD::SELECT,  MVT::v8i64,   { 1, 1, 1, 1 } },
     { ISD::SELECT,  MVT::v4i64,   { 1, 1, 1, 1 } },
@@ -2959,6 +2963,13 @@ InstructionCost X86TTIImpl::getCmpSelInstrCost(unsigned Opcode, Type *ValTy,
   };
 
   static const CostKindTblEntry AVX2CostTbl[] = {
+    { ISD::SETCC,   MVT::v4f64,   { 1, 4, 1, 2 } },
+    { ISD::SETCC,   MVT::v2f64,   { 1, 4, 1, 1 } },
+    { ISD::SETCC,   MVT::f64,     { 1, 4, 1, 1 } },
+    { ISD::SETCC,   MVT::v8f32,   { 1, 4, 1, 2 } },
+    { ISD::SETCC,   MVT::v4f32,   { 1, 4, 1, 1 } },
+    { ISD::SETCC,   MVT::f32,     { 1, 4, 1, 1 } },
+
     { ISD::SETCC,   MVT::v4i64,   { 1 } },
     { ISD::SETCC,   MVT::v8i32,   { 1 } },
     { ISD::SETCC,   MVT::v16i16,  { 1 } },
@@ -2973,8 +2984,13 @@ InstructionCost X86TTIImpl::getCmpSelInstrCost(unsigned Opcode, Type *ValTy,
   };
 
   static const CostKindTblEntry AVX1CostTbl[] = {
-    { ISD::SETCC,   MVT::v4f64,   { 1 } },
-    { ISD::SETCC,   MVT::v8f32,   { 1 } },
+    { ISD::SETCC,   MVT::v4f64,   { 2, 3, 1, 2 } },
+    { ISD::SETCC,   MVT::v2f64,   { 1, 3, 1, 1 } },
+    { ISD::SETCC,   MVT::f64,     { 1, 3, 1, 1 } },
+    { ISD::SETCC,   MVT::v8f32,   { 2, 3, 1, 2 } },
+    { ISD::SETCC,   MVT::v4f32,   { 1, 3, 1, 1 } },
+    { ISD::SETCC,   MVT::f32,     { 1, 3, 1, 1 } },
+
     // AVX1 does not support 8-wide integer compare.
     { ISD::SETCC,   MVT::v4i64,   { 4 } },
     { ISD::SETCC,   MVT::v8i32,   { 4 } },
@@ -2994,8 +3010,8 @@ InstructionCost X86TTIImpl::getCmpSelInstrCost(unsigned Opcode, Type *ValTy,
   };
 
   static const CostKindTblEntry SSE41CostTbl[] = {
-    { ISD::SETCC,   MVT::v2f64,   { 1 } },
-    { ISD::SETCC,   MVT::v4f32,   { 1 } },
+    { ISD::SETCC,   MVT::v2f64,   { 1, 5, 1, 1 } },
+    { ISD::SETCC,   MVT::v4f32,   { 1, 5, 1, 1 } },
 
     { ISD::SELECT,  MVT::v2f64,   { 2, 2, 1, 2 } }, // blendvpd
     { ISD::SELECT,  MVT::f64,     { 2, 2, 1, 2 } }, // blendvpd
@@ -3008,8 +3024,9 @@ InstructionCost X86TTIImpl::getCmpSelInstrCost(unsigned Opcode, Type *ValTy,
   };
 
   static const CostKindTblEntry SSE2CostTbl[] = {
-    { ISD::SETCC,   MVT::v2f64,   { 2 } },
-    { ISD::SETCC,   MVT::f64,     { 1 } },
+    { ISD::SETCC,   MVT::v2f64,   { 2, 5, 1, 1 } },
+    { ISD::SETCC,   MVT::f64,     { 1, 5, 1, 1 } },
+
     { ISD::SETCC,   MVT::v2i64,   { 5 } }, // pcmpeqd/pcmpgtd expansion
     { ISD::SETCC,   MVT::v4i32,   { 1 } },
     { ISD::SETCC,   MVT::v8i16,   { 1 } },
@@ -3024,8 +3041,8 @@ InstructionCost X86TTIImpl::getCmpSelInstrCost(unsigned Opcode, Type *ValTy,
   };
 
   static const CostKindTblEntry SSE1CostTbl[] = {
-    { ISD::SETCC,   MVT::v4f32,   { 2 } },
-    { ISD::SETCC,   MVT::f32,     { 1 } },
+    { ISD::SETCC,   MVT::v4f32,   { 2, 5, 1, 1 } },
+    { ISD::SETCC,   MVT::f32,     { 1, 5, 1, 1 } },
 
     { ISD::SELECT,  MVT::v4f32,   { 2, 2, 3, 3 } }, // andps + andnps + orps
     { ISD::SELECT,  MVT::f32,     { 2, 2, 3, 3 } }, // andps + andnps + orps

diff  --git a/llvm/test/Analysis/CostModel/X86/fcmp-codesize.ll b/llvm/test/Analysis/CostModel/X86/fcmp-codesize.ll
index 1ae24ed05f9cc..e1e8ff8f6ab85 100644
--- a/llvm/test/Analysis/CostModel/X86/fcmp-codesize.ll
+++ b/llvm/test/Analysis/CostModel/X86/fcmp-codesize.ll
@@ -1,31 +1,57 @@
 ; NOTE: Assertions have been autogenerated by utils/update_analyze_test_checks.py
-; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mattr=+sse2 | FileCheck %s
-; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mattr=+sse3 | FileCheck %s
-; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mattr=+ssse3 | FileCheck %s
-; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mattr=+sse4.1 | FileCheck %s
-; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mattr=+sse4.2 | FileCheck %s
-; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mattr=+avx | FileCheck %s
-; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mattr=+avx2 | FileCheck %s
-; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mattr=+avx512f | FileCheck %s
-; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mattr=+avx512f,+avx512bw | FileCheck %s
-;
-; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mcpu=slm | FileCheck %s
-; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mcpu=goldmont | FileCheck %s
-; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mcpu=btver2 | FileCheck %s
+; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mattr=+sse2 | FileCheck %s --check-prefixes=SSE
+; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mattr=+sse3 | FileCheck %s --check-prefixes=SSE
+; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mattr=+ssse3 | FileCheck %s --check-prefixes=SSE
+; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mattr=+sse4.1 | FileCheck %s --check-prefixes=SSE
+; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mattr=+sse4.2 | FileCheck %s --check-prefixes=SSE
+; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mattr=+avx | FileCheck %s --check-prefixes=AVX
+; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mattr=+avx2 | FileCheck %s --check-prefixes=AVX
+; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mattr=+avx512f | FileCheck %s --check-prefixes=AVX512
+; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefixes=AVX512
+;
+; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mcpu=slm | FileCheck %s --check-prefixes=SSE
+; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mcpu=goldmont | FileCheck %s --check-prefixes=SSE
+; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=code-size -mcpu=btver2 | FileCheck %s --check-prefixes=AVX
 
 define i32 @cmp_float_oeq(i32 %arg) {
-; CHECK-LABEL: 'cmp_float_oeq'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp oeq float undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp oeq <2 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp oeq <4 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp oeq <8 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp oeq <16 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp oeq double undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp oeq <2 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp oeq <4 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp oeq <8 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F64 = fcmp oeq <16 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE-LABEL: 'cmp_float_oeq'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp oeq float undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp oeq <2 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp oeq <4 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp oeq <8 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp oeq <16 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp oeq double undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp oeq <2 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp oeq <4 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp oeq <8 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp oeq <16 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX-LABEL: 'cmp_float_oeq'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp oeq float undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp oeq <2 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp oeq <4 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp oeq <8 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F32 = fcmp oeq <16 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp oeq double undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp oeq <2 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp oeq <4 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F64 = fcmp oeq <8 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F64 = fcmp oeq <16 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512-LABEL: 'cmp_float_oeq'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp oeq float undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp oeq <2 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp oeq <4 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp oeq <8 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp oeq <16 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp oeq double undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp oeq <2 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp oeq <4 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp oeq <8 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F64 = fcmp oeq <16 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %F32 = fcmp oeq float undef, undef
   %V2F32 = fcmp oeq <2 x float> undef, undef
@@ -43,18 +69,44 @@ define i32 @cmp_float_oeq(i32 %arg) {
 }
 
 define i32 @cmp_float_one(i32 %arg) {
-; CHECK-LABEL: 'cmp_float_one'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp one float undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp one <2 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp one <4 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp one <8 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp one <16 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp one double undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp one <2 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp one <4 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp one <8 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F64 = fcmp one <16 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE-LABEL: 'cmp_float_one'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp one float undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp one <2 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp one <4 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp one <8 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp one <16 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp one double undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp one <2 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp one <4 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp one <8 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp one <16 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX-LABEL: 'cmp_float_one'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp one float undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp one <2 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp one <4 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp one <8 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F32 = fcmp one <16 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp one double undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp one <2 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp one <4 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F64 = fcmp one <8 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F64 = fcmp one <16 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512-LABEL: 'cmp_float_one'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp one float undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp one <2 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp one <4 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp one <8 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp one <16 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp one double undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp one <2 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp one <4 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp one <8 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F64 = fcmp one <16 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %F32 = fcmp one float undef, undef
   %V2F32 = fcmp one <2 x float> undef, undef
@@ -72,18 +124,44 @@ define i32 @cmp_float_one(i32 %arg) {
 }
 
 define i32 @cmp_float_ord(i32 %arg) {
-; CHECK-LABEL: 'cmp_float_ord'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ord float undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ord <2 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ord <4 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ord <8 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp ord <16 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ord double undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ord <2 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ord <4 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp ord <8 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F64 = fcmp ord <16 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE-LABEL: 'cmp_float_ord'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ord float undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ord <2 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ord <4 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp ord <8 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp ord <16 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ord double undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ord <2 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp ord <4 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp ord <8 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp ord <16 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX-LABEL: 'cmp_float_ord'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ord float undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ord <2 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ord <4 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ord <8 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F32 = fcmp ord <16 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ord double undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ord <2 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ord <4 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F64 = fcmp ord <8 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F64 = fcmp ord <16 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512-LABEL: 'cmp_float_ord'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ord float undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ord <2 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ord <4 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ord <8 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp ord <16 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ord double undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ord <2 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ord <4 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp ord <8 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F64 = fcmp ord <16 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %F32 = fcmp ord float undef, undef
   %V2F32 = fcmp ord <2 x float> undef, undef
@@ -101,18 +179,44 @@ define i32 @cmp_float_ord(i32 %arg) {
 }
 
 define i32 @cmp_float_oge(i32 %arg) {
-; CHECK-LABEL: 'cmp_float_oge'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp oge float undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp oge <2 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp oge <4 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp oge <8 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp oge <16 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp oge double undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp oge <2 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp oge <4 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp oge <8 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F64 = fcmp oge <16 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE-LABEL: 'cmp_float_oge'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp oge float undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp oge <2 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp oge <4 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp oge <8 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp oge <16 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp oge double undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp oge <2 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp oge <4 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp oge <8 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp oge <16 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX-LABEL: 'cmp_float_oge'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp oge float undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp oge <2 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp oge <4 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp oge <8 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F32 = fcmp oge <16 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp oge double undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp oge <2 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp oge <4 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F64 = fcmp oge <8 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F64 = fcmp oge <16 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512-LABEL: 'cmp_float_oge'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp oge float undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp oge <2 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp oge <4 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp oge <8 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp oge <16 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp oge double undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp oge <2 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp oge <4 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp oge <8 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F64 = fcmp oge <16 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %F32 = fcmp oge float undef, undef
   %V2F32 = fcmp oge <2 x float> undef, undef
@@ -130,18 +234,44 @@ define i32 @cmp_float_oge(i32 %arg) {
 }
 
 define i32 @cmp_float_ogt(i32 %arg) {
-; CHECK-LABEL: 'cmp_float_ogt'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ogt float undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ogt <2 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ogt <4 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ogt <8 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp ogt <16 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ogt double undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ogt <2 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ogt <4 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp ogt <8 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F64 = fcmp ogt <16 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE-LABEL: 'cmp_float_ogt'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ogt float undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ogt <2 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ogt <4 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp ogt <8 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp ogt <16 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ogt double undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ogt <2 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp ogt <4 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp ogt <8 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp ogt <16 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX-LABEL: 'cmp_float_ogt'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ogt float undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ogt <2 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ogt <4 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ogt <8 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F32 = fcmp ogt <16 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ogt double undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ogt <2 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ogt <4 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F64 = fcmp ogt <8 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F64 = fcmp ogt <16 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512-LABEL: 'cmp_float_ogt'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ogt float undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ogt <2 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ogt <4 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ogt <8 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp ogt <16 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ogt double undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ogt <2 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ogt <4 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp ogt <8 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F64 = fcmp ogt <16 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %F32 = fcmp ogt float undef, undef
   %V2F32 = fcmp ogt <2 x float> undef, undef
@@ -159,18 +289,44 @@ define i32 @cmp_float_ogt(i32 %arg) {
 }
 
 define i32 @cmp_float_ole(i32 %arg) {
-; CHECK-LABEL: 'cmp_float_ole'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ole float undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ole <2 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ole <4 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ole <8 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp ole <16 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ole double undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ole <2 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ole <4 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp ole <8 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F64 = fcmp ole <16 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE-LABEL: 'cmp_float_ole'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ole float undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ole <2 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ole <4 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp ole <8 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp ole <16 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ole double undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ole <2 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp ole <4 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp ole <8 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp ole <16 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX-LABEL: 'cmp_float_ole'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ole float undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ole <2 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ole <4 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ole <8 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F32 = fcmp ole <16 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ole double undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ole <2 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ole <4 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F64 = fcmp ole <8 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F64 = fcmp ole <16 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512-LABEL: 'cmp_float_ole'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ole float undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ole <2 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ole <4 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ole <8 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp ole <16 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ole double undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ole <2 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ole <4 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp ole <8 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F64 = fcmp ole <16 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %F32 = fcmp ole float undef, undef
   %V2F32 = fcmp ole <2 x float> undef, undef
@@ -188,18 +344,44 @@ define i32 @cmp_float_ole(i32 %arg) {
 }
 
 define i32 @cmp_float_olt(i32 %arg) {
-; CHECK-LABEL: 'cmp_float_olt'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp olt float undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp olt <2 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp olt <4 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp olt <8 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp olt <16 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp olt double undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp olt <2 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp olt <4 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp olt <8 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F64 = fcmp olt <16 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE-LABEL: 'cmp_float_olt'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp olt float undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp olt <2 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp olt <4 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp olt <8 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp olt <16 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp olt double undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp olt <2 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp olt <4 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp olt <8 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp olt <16 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX-LABEL: 'cmp_float_olt'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp olt float undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp olt <2 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp olt <4 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp olt <8 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F32 = fcmp olt <16 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp olt double undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp olt <2 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp olt <4 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F64 = fcmp olt <8 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F64 = fcmp olt <16 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512-LABEL: 'cmp_float_olt'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp olt float undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp olt <2 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp olt <4 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp olt <8 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp olt <16 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp olt double undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp olt <2 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp olt <4 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp olt <8 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F64 = fcmp olt <16 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %F32 = fcmp olt float undef, undef
   %V2F32 = fcmp olt <2 x float> undef, undef
@@ -217,18 +399,44 @@ define i32 @cmp_float_olt(i32 %arg) {
 }
 
 define i32 @cmp_float_ueq(i32 %arg) {
-; CHECK-LABEL: 'cmp_float_ueq'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ueq float undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ueq <2 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ueq <4 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ueq <8 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp ueq <16 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ueq double undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ueq <2 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ueq <4 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp ueq <8 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F64 = fcmp ueq <16 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE-LABEL: 'cmp_float_ueq'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ueq float undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ueq <2 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ueq <4 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp ueq <8 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp ueq <16 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ueq double undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ueq <2 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp ueq <4 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp ueq <8 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp ueq <16 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX-LABEL: 'cmp_float_ueq'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ueq float undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ueq <2 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ueq <4 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ueq <8 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F32 = fcmp ueq <16 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ueq double undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ueq <2 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ueq <4 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F64 = fcmp ueq <8 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F64 = fcmp ueq <16 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512-LABEL: 'cmp_float_ueq'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ueq float undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ueq <2 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ueq <4 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ueq <8 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp ueq <16 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ueq double undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ueq <2 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ueq <4 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp ueq <8 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F64 = fcmp ueq <16 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %F32 = fcmp ueq float undef, undef
   %V2F32 = fcmp ueq <2 x float> undef, undef
@@ -246,18 +454,44 @@ define i32 @cmp_float_ueq(i32 %arg) {
 }
 
 define i32 @cmp_float_une(i32 %arg) {
-; CHECK-LABEL: 'cmp_float_une'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp une float undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp une <2 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp une <4 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp une <8 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp une <16 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp une double undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp une <2 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp une <4 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp une <8 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F64 = fcmp une <16 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE-LABEL: 'cmp_float_une'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp une float undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp une <2 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp une <4 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp une <8 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp une <16 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp une double undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp une <2 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp une <4 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp une <8 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp une <16 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX-LABEL: 'cmp_float_une'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp une float undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp une <2 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp une <4 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp une <8 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F32 = fcmp une <16 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp une double undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp une <2 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp une <4 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F64 = fcmp une <8 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F64 = fcmp une <16 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512-LABEL: 'cmp_float_une'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp une float undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp une <2 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp une <4 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp une <8 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp une <16 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp une double undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp une <2 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp une <4 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp une <8 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F64 = fcmp une <16 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %F32 = fcmp une float undef, undef
   %V2F32 = fcmp une <2 x float> undef, undef
@@ -275,18 +509,44 @@ define i32 @cmp_float_une(i32 %arg) {
 }
 
 define i32 @cmp_float_uno(i32 %arg) {
-; CHECK-LABEL: 'cmp_float_uno'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp uno float undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp uno <2 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp uno <4 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp uno <8 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp uno <16 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp uno double undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp uno <2 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp uno <4 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp uno <8 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F64 = fcmp uno <16 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE-LABEL: 'cmp_float_uno'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp uno float undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp uno <2 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp uno <4 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp uno <8 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp uno <16 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp uno double undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp uno <2 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp uno <4 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp uno <8 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp uno <16 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX-LABEL: 'cmp_float_uno'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp uno float undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp uno <2 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp uno <4 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp uno <8 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F32 = fcmp uno <16 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp uno double undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp uno <2 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp uno <4 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F64 = fcmp uno <8 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F64 = fcmp uno <16 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512-LABEL: 'cmp_float_uno'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp uno float undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp uno <2 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp uno <4 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp uno <8 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp uno <16 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp uno double undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp uno <2 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp uno <4 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp uno <8 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F64 = fcmp uno <16 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %F32 = fcmp uno float undef, undef
   %V2F32 = fcmp uno <2 x float> undef, undef
@@ -304,18 +564,44 @@ define i32 @cmp_float_uno(i32 %arg) {
 }
 
 define i32 @cmp_float_uge(i32 %arg) {
-; CHECK-LABEL: 'cmp_float_uge'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp uge float undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp uge <2 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp uge <4 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp uge <8 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp uge <16 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp uge double undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp uge <2 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp uge <4 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp uge <8 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F64 = fcmp uge <16 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE-LABEL: 'cmp_float_uge'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp uge float undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp uge <2 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp uge <4 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp uge <8 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp uge <16 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp uge double undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp uge <2 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp uge <4 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp uge <8 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp uge <16 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX-LABEL: 'cmp_float_uge'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp uge float undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp uge <2 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp uge <4 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp uge <8 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F32 = fcmp uge <16 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp uge double undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp uge <2 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp uge <4 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F64 = fcmp uge <8 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F64 = fcmp uge <16 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512-LABEL: 'cmp_float_uge'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp uge float undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp uge <2 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp uge <4 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp uge <8 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp uge <16 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp uge double undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp uge <2 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp uge <4 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp uge <8 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F64 = fcmp uge <16 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %F32 = fcmp uge float undef, undef
   %V2F32 = fcmp uge <2 x float> undef, undef
@@ -333,18 +619,44 @@ define i32 @cmp_float_uge(i32 %arg) {
 }
 
 define i32 @cmp_float_ugt(i32 %arg) {
-; CHECK-LABEL: 'cmp_float_ugt'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ugt float undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ugt <2 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ugt <4 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ugt <8 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp ugt <16 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ugt double undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ugt <2 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ugt <4 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp ugt <8 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F64 = fcmp ugt <16 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE-LABEL: 'cmp_float_ugt'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ugt float undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ugt <2 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ugt <4 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp ugt <8 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp ugt <16 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ugt double undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ugt <2 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp ugt <4 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp ugt <8 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp ugt <16 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX-LABEL: 'cmp_float_ugt'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ugt float undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ugt <2 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ugt <4 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ugt <8 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F32 = fcmp ugt <16 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ugt double undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ugt <2 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ugt <4 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F64 = fcmp ugt <8 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F64 = fcmp ugt <16 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512-LABEL: 'cmp_float_ugt'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ugt float undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ugt <2 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ugt <4 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ugt <8 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp ugt <16 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ugt double undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ugt <2 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ugt <4 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp ugt <8 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F64 = fcmp ugt <16 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %F32 = fcmp ugt float undef, undef
   %V2F32 = fcmp ugt <2 x float> undef, undef
@@ -362,18 +674,44 @@ define i32 @cmp_float_ugt(i32 %arg) {
 }
 
 define i32 @cmp_float_ule(i32 %arg) {
-; CHECK-LABEL: 'cmp_float_ule'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ule float undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ule <2 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ule <4 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ule <8 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp ule <16 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ule double undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ule <2 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ule <4 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp ule <8 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F64 = fcmp ule <16 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE-LABEL: 'cmp_float_ule'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ule float undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ule <2 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ule <4 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp ule <8 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp ule <16 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ule double undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ule <2 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp ule <4 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp ule <8 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp ule <16 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX-LABEL: 'cmp_float_ule'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ule float undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ule <2 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ule <4 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ule <8 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F32 = fcmp ule <16 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ule double undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ule <2 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ule <4 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F64 = fcmp ule <8 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F64 = fcmp ule <16 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512-LABEL: 'cmp_float_ule'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ule float undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ule <2 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ule <4 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ule <8 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp ule <16 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ule double undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ule <2 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ule <4 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp ule <8 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F64 = fcmp ule <16 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %F32 = fcmp ule float undef, undef
   %V2F32 = fcmp ule <2 x float> undef, undef
@@ -391,18 +729,44 @@ define i32 @cmp_float_ule(i32 %arg) {
 }
 
 define i32 @cmp_float_ult(i32 %arg) {
-; CHECK-LABEL: 'cmp_float_ult'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ult float undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ult <2 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ult <4 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ult <8 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp ult <16 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ult double undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ult <2 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ult <4 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp ult <8 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F64 = fcmp ult <16 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE-LABEL: 'cmp_float_ult'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ult float undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ult <2 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ult <4 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp ult <8 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp ult <16 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ult double undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ult <2 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp ult <4 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp ult <8 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp ult <16 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX-LABEL: 'cmp_float_ult'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ult float undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ult <2 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ult <4 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ult <8 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F32 = fcmp ult <16 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ult double undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ult <2 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ult <4 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F64 = fcmp ult <8 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F64 = fcmp ult <16 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512-LABEL: 'cmp_float_ult'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ult float undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ult <2 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ult <4 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ult <8 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp ult <16 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ult double undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ult <2 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ult <4 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp ult <8 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F64 = fcmp ult <16 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %F32 = fcmp ult float undef, undef
   %V2F32 = fcmp ult <2 x float> undef, undef
@@ -420,18 +784,44 @@ define i32 @cmp_float_ult(i32 %arg) {
 }
 
 define i32 @cmp_float_false(i32 %arg) {
-; CHECK-LABEL: 'cmp_float_false'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp false float undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp false <2 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp false <4 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp false <8 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp false <16 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp false double undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp false <2 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp false <4 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp false <8 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F64 = fcmp false <16 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE-LABEL: 'cmp_float_false'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp false float undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp false <2 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp false <4 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp false <8 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp false <16 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp false double undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp false <2 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp false <4 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp false <8 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp false <16 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX-LABEL: 'cmp_float_false'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp false float undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp false <2 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp false <4 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp false <8 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F32 = fcmp false <16 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp false double undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp false <2 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp false <4 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F64 = fcmp false <8 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F64 = fcmp false <16 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512-LABEL: 'cmp_float_false'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp false float undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp false <2 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp false <4 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp false <8 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp false <16 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp false double undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp false <2 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp false <4 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp false <8 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F64 = fcmp false <16 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %F32 = fcmp false float undef, undef
   %V2F32 = fcmp false <2 x float> undef, undef
@@ -449,18 +839,44 @@ define i32 @cmp_float_false(i32 %arg) {
 }
 
 define i32 @cmp_float_true(i32 %arg) {
-; CHECK-LABEL: 'cmp_float_true'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp true float undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp true <2 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp true <4 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp true <8 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp true <16 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp true double undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp true <2 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp true <4 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp true <8 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F64 = fcmp true <16 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE-LABEL: 'cmp_float_true'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp true float undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp true <2 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp true <4 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp true <8 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp true <16 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp true double undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp true <2 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp true <4 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp true <8 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp true <16 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX-LABEL: 'cmp_float_true'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp true float undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp true <2 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp true <4 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp true <8 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F32 = fcmp true <16 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp true double undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp true <2 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp true <4 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F64 = fcmp true <8 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F64 = fcmp true <16 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512-LABEL: 'cmp_float_true'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp true float undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp true <2 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp true <4 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp true <8 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp true <16 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp true double undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp true <2 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp true <4 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp true <8 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F64 = fcmp true <16 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %F32 = fcmp true float undef, undef
   %V2F32 = fcmp true <2 x float> undef, undef

diff  --git a/llvm/test/Analysis/CostModel/X86/fcmp-latency.ll b/llvm/test/Analysis/CostModel/X86/fcmp-latency.ll
index 113913e2a8aaa..a3d8b51f28af1 100644
--- a/llvm/test/Analysis/CostModel/X86/fcmp-latency.ll
+++ b/llvm/test/Analysis/CostModel/X86/fcmp-latency.ll
@@ -1,31 +1,70 @@
 ; NOTE: Assertions have been autogenerated by utils/update_analyze_test_checks.py
-; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mattr=+sse2 | FileCheck %s
-; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mattr=+sse3 | FileCheck %s
-; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mattr=+ssse3 | FileCheck %s
-; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mattr=+sse4.1 | FileCheck %s
-; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mattr=+sse4.2 | FileCheck %s
-; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mattr=+avx | FileCheck %s
-; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mattr=+avx2 | FileCheck %s
-; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mattr=+avx512f | FileCheck %s
-; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mattr=+avx512f,+avx512bw | FileCheck %s
-;
-; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mcpu=slm | FileCheck %s
-; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mcpu=goldmont | FileCheck %s
-; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mcpu=btver2 | FileCheck %s
+; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mattr=+sse2 | FileCheck %s --check-prefixes=SSE
+; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mattr=+sse3 | FileCheck %s --check-prefixes=SSE
+; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mattr=+ssse3 | FileCheck %s --check-prefixes=SSE
+; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mattr=+sse4.1 | FileCheck %s --check-prefixes=SSE
+; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mattr=+sse4.2 | FileCheck %s --check-prefixes=SSE
+; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mattr=+avx | FileCheck %s --check-prefixes=AVX1
+; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mattr=+avx2 | FileCheck %s --check-prefixes=AVX2
+; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mattr=+avx512f | FileCheck %s --check-prefixes=AVX512
+; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefixes=AVX512
+;
+; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mcpu=slm | FileCheck %s --check-prefixes=SSE
+; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mcpu=goldmont | FileCheck %s --check-prefixes=SSE
+; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=latency -mcpu=btver2 | FileCheck %s --check-prefixes=AVX1
 
 define i32 @cmp_float_oeq(i32 %arg) {
-; CHECK-LABEL: 'cmp_float_oeq'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp oeq float undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp oeq <2 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp oeq <4 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp oeq <8 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp oeq <16 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp oeq double undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp oeq <2 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp oeq <4 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp oeq <8 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F64 = fcmp oeq <16 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE-LABEL: 'cmp_float_oeq'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %F32 = fcmp oeq float undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V2F32 = fcmp oeq <2 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V4F32 = fcmp oeq <4 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V8F32 = fcmp oeq <8 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V16F32 = fcmp oeq <16 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %F64 = fcmp oeq double undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V2F64 = fcmp oeq <2 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V4F64 = fcmp oeq <4 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V8F64 = fcmp oeq <8 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 40 for instruction: %V16F64 = fcmp oeq <16 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX1-LABEL: 'cmp_float_oeq'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %F32 = fcmp oeq float undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2F32 = fcmp oeq <2 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4F32 = fcmp oeq <4 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8F32 = fcmp oeq <8 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16F32 = fcmp oeq <16 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %F64 = fcmp oeq double undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2F64 = fcmp oeq <2 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4F64 = fcmp oeq <4 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8F64 = fcmp oeq <8 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V16F64 = fcmp oeq <16 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX2-LABEL: 'cmp_float_oeq'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F32 = fcmp oeq float undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F32 = fcmp oeq <2 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F32 = fcmp oeq <4 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp oeq <8 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = fcmp oeq <16 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F64 = fcmp oeq double undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F64 = fcmp oeq <2 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp oeq <4 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8F64 = fcmp oeq <8 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp oeq <16 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512-LABEL: 'cmp_float_oeq'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F32 = fcmp oeq float undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F32 = fcmp oeq <2 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F32 = fcmp oeq <4 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp oeq <8 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp oeq <16 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F64 = fcmp oeq double undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F64 = fcmp oeq <2 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp oeq <4 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp oeq <8 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp oeq <16 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %F32 = fcmp oeq float undef, undef
   %V2F32 = fcmp oeq <2 x float> undef, undef
@@ -43,18 +82,57 @@ define i32 @cmp_float_oeq(i32 %arg) {
 }
 
 define i32 @cmp_float_one(i32 %arg) {
-; CHECK-LABEL: 'cmp_float_one'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp one float undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp one <2 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp one <4 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp one <8 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp one <16 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp one double undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp one <2 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp one <4 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp one <8 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F64 = fcmp one <16 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE-LABEL: 'cmp_float_one'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %F32 = fcmp one float undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V2F32 = fcmp one <2 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V4F32 = fcmp one <4 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V8F32 = fcmp one <8 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V16F32 = fcmp one <16 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %F64 = fcmp one double undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V2F64 = fcmp one <2 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V4F64 = fcmp one <4 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V8F64 = fcmp one <8 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 40 for instruction: %V16F64 = fcmp one <16 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX1-LABEL: 'cmp_float_one'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %F32 = fcmp one float undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2F32 = fcmp one <2 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4F32 = fcmp one <4 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8F32 = fcmp one <8 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16F32 = fcmp one <16 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %F64 = fcmp one double undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2F64 = fcmp one <2 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4F64 = fcmp one <4 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8F64 = fcmp one <8 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V16F64 = fcmp one <16 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX2-LABEL: 'cmp_float_one'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F32 = fcmp one float undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F32 = fcmp one <2 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F32 = fcmp one <4 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp one <8 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = fcmp one <16 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F64 = fcmp one double undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F64 = fcmp one <2 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp one <4 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8F64 = fcmp one <8 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp one <16 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512-LABEL: 'cmp_float_one'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F32 = fcmp one float undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F32 = fcmp one <2 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F32 = fcmp one <4 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp one <8 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp one <16 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F64 = fcmp one double undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F64 = fcmp one <2 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp one <4 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp one <8 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp one <16 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %F32 = fcmp one float undef, undef
   %V2F32 = fcmp one <2 x float> undef, undef
@@ -72,18 +150,57 @@ define i32 @cmp_float_one(i32 %arg) {
 }
 
 define i32 @cmp_float_ord(i32 %arg) {
-; CHECK-LABEL: 'cmp_float_ord'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ord float undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ord <2 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ord <4 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ord <8 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp ord <16 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ord double undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ord <2 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ord <4 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp ord <8 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F64 = fcmp ord <16 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE-LABEL: 'cmp_float_ord'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %F32 = fcmp ord float undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V2F32 = fcmp ord <2 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V4F32 = fcmp ord <4 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V8F32 = fcmp ord <8 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V16F32 = fcmp ord <16 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %F64 = fcmp ord double undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V2F64 = fcmp ord <2 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V4F64 = fcmp ord <4 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V8F64 = fcmp ord <8 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 40 for instruction: %V16F64 = fcmp ord <16 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX1-LABEL: 'cmp_float_ord'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %F32 = fcmp ord float undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2F32 = fcmp ord <2 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4F32 = fcmp ord <4 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8F32 = fcmp ord <8 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16F32 = fcmp ord <16 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %F64 = fcmp ord double undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2F64 = fcmp ord <2 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4F64 = fcmp ord <4 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8F64 = fcmp ord <8 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V16F64 = fcmp ord <16 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX2-LABEL: 'cmp_float_ord'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F32 = fcmp ord float undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F32 = fcmp ord <2 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F32 = fcmp ord <4 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp ord <8 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = fcmp ord <16 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F64 = fcmp ord double undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F64 = fcmp ord <2 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp ord <4 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8F64 = fcmp ord <8 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp ord <16 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512-LABEL: 'cmp_float_ord'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F32 = fcmp ord float undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F32 = fcmp ord <2 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F32 = fcmp ord <4 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp ord <8 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp ord <16 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F64 = fcmp ord double undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F64 = fcmp ord <2 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp ord <4 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp ord <8 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp ord <16 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %F32 = fcmp ord float undef, undef
   %V2F32 = fcmp ord <2 x float> undef, undef
@@ -101,18 +218,57 @@ define i32 @cmp_float_ord(i32 %arg) {
 }
 
 define i32 @cmp_float_oge(i32 %arg) {
-; CHECK-LABEL: 'cmp_float_oge'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp oge float undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp oge <2 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp oge <4 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp oge <8 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp oge <16 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp oge double undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp oge <2 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp oge <4 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp oge <8 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F64 = fcmp oge <16 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE-LABEL: 'cmp_float_oge'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %F32 = fcmp oge float undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V2F32 = fcmp oge <2 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V4F32 = fcmp oge <4 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V8F32 = fcmp oge <8 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V16F32 = fcmp oge <16 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %F64 = fcmp oge double undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V2F64 = fcmp oge <2 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V4F64 = fcmp oge <4 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V8F64 = fcmp oge <8 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 40 for instruction: %V16F64 = fcmp oge <16 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX1-LABEL: 'cmp_float_oge'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %F32 = fcmp oge float undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2F32 = fcmp oge <2 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4F32 = fcmp oge <4 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8F32 = fcmp oge <8 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16F32 = fcmp oge <16 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %F64 = fcmp oge double undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2F64 = fcmp oge <2 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4F64 = fcmp oge <4 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8F64 = fcmp oge <8 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V16F64 = fcmp oge <16 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX2-LABEL: 'cmp_float_oge'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F32 = fcmp oge float undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F32 = fcmp oge <2 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F32 = fcmp oge <4 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp oge <8 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = fcmp oge <16 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F64 = fcmp oge double undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F64 = fcmp oge <2 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp oge <4 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8F64 = fcmp oge <8 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp oge <16 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512-LABEL: 'cmp_float_oge'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F32 = fcmp oge float undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F32 = fcmp oge <2 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F32 = fcmp oge <4 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp oge <8 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp oge <16 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F64 = fcmp oge double undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F64 = fcmp oge <2 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp oge <4 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp oge <8 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp oge <16 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %F32 = fcmp oge float undef, undef
   %V2F32 = fcmp oge <2 x float> undef, undef
@@ -130,18 +286,57 @@ define i32 @cmp_float_oge(i32 %arg) {
 }
 
 define i32 @cmp_float_ogt(i32 %arg) {
-; CHECK-LABEL: 'cmp_float_ogt'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ogt float undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ogt <2 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ogt <4 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ogt <8 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp ogt <16 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ogt double undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ogt <2 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ogt <4 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp ogt <8 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F64 = fcmp ogt <16 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE-LABEL: 'cmp_float_ogt'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %F32 = fcmp ogt float undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V2F32 = fcmp ogt <2 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V4F32 = fcmp ogt <4 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V8F32 = fcmp ogt <8 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V16F32 = fcmp ogt <16 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %F64 = fcmp ogt double undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V2F64 = fcmp ogt <2 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V4F64 = fcmp ogt <4 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V8F64 = fcmp ogt <8 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 40 for instruction: %V16F64 = fcmp ogt <16 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX1-LABEL: 'cmp_float_ogt'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %F32 = fcmp ogt float undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2F32 = fcmp ogt <2 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4F32 = fcmp ogt <4 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8F32 = fcmp ogt <8 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16F32 = fcmp ogt <16 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %F64 = fcmp ogt double undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2F64 = fcmp ogt <2 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4F64 = fcmp ogt <4 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8F64 = fcmp ogt <8 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V16F64 = fcmp ogt <16 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX2-LABEL: 'cmp_float_ogt'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F32 = fcmp ogt float undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F32 = fcmp ogt <2 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F32 = fcmp ogt <4 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp ogt <8 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = fcmp ogt <16 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F64 = fcmp ogt double undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F64 = fcmp ogt <2 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp ogt <4 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8F64 = fcmp ogt <8 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp ogt <16 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512-LABEL: 'cmp_float_ogt'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F32 = fcmp ogt float undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F32 = fcmp ogt <2 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F32 = fcmp ogt <4 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp ogt <8 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp ogt <16 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F64 = fcmp ogt double undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F64 = fcmp ogt <2 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp ogt <4 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp ogt <8 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp ogt <16 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %F32 = fcmp ogt float undef, undef
   %V2F32 = fcmp ogt <2 x float> undef, undef
@@ -159,18 +354,57 @@ define i32 @cmp_float_ogt(i32 %arg) {
 }
 
 define i32 @cmp_float_ole(i32 %arg) {
-; CHECK-LABEL: 'cmp_float_ole'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ole float undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ole <2 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ole <4 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ole <8 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp ole <16 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ole double undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ole <2 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ole <4 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp ole <8 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F64 = fcmp ole <16 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE-LABEL: 'cmp_float_ole'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %F32 = fcmp ole float undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V2F32 = fcmp ole <2 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V4F32 = fcmp ole <4 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V8F32 = fcmp ole <8 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V16F32 = fcmp ole <16 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %F64 = fcmp ole double undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V2F64 = fcmp ole <2 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V4F64 = fcmp ole <4 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V8F64 = fcmp ole <8 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 40 for instruction: %V16F64 = fcmp ole <16 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX1-LABEL: 'cmp_float_ole'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %F32 = fcmp ole float undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2F32 = fcmp ole <2 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4F32 = fcmp ole <4 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8F32 = fcmp ole <8 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16F32 = fcmp ole <16 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %F64 = fcmp ole double undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2F64 = fcmp ole <2 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4F64 = fcmp ole <4 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8F64 = fcmp ole <8 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V16F64 = fcmp ole <16 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX2-LABEL: 'cmp_float_ole'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F32 = fcmp ole float undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F32 = fcmp ole <2 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F32 = fcmp ole <4 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp ole <8 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = fcmp ole <16 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F64 = fcmp ole double undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F64 = fcmp ole <2 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp ole <4 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8F64 = fcmp ole <8 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp ole <16 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512-LABEL: 'cmp_float_ole'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F32 = fcmp ole float undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F32 = fcmp ole <2 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F32 = fcmp ole <4 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp ole <8 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp ole <16 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F64 = fcmp ole double undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F64 = fcmp ole <2 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp ole <4 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp ole <8 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp ole <16 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %F32 = fcmp ole float undef, undef
   %V2F32 = fcmp ole <2 x float> undef, undef
@@ -188,18 +422,57 @@ define i32 @cmp_float_ole(i32 %arg) {
 }
 
 define i32 @cmp_float_olt(i32 %arg) {
-; CHECK-LABEL: 'cmp_float_olt'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp olt float undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp olt <2 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp olt <4 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp olt <8 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp olt <16 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp olt double undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp olt <2 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp olt <4 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp olt <8 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F64 = fcmp olt <16 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE-LABEL: 'cmp_float_olt'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %F32 = fcmp olt float undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V2F32 = fcmp olt <2 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V4F32 = fcmp olt <4 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V8F32 = fcmp olt <8 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V16F32 = fcmp olt <16 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %F64 = fcmp olt double undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V2F64 = fcmp olt <2 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V4F64 = fcmp olt <4 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V8F64 = fcmp olt <8 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 40 for instruction: %V16F64 = fcmp olt <16 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX1-LABEL: 'cmp_float_olt'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %F32 = fcmp olt float undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2F32 = fcmp olt <2 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4F32 = fcmp olt <4 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8F32 = fcmp olt <8 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16F32 = fcmp olt <16 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %F64 = fcmp olt double undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2F64 = fcmp olt <2 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4F64 = fcmp olt <4 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8F64 = fcmp olt <8 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V16F64 = fcmp olt <16 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX2-LABEL: 'cmp_float_olt'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F32 = fcmp olt float undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F32 = fcmp olt <2 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F32 = fcmp olt <4 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp olt <8 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = fcmp olt <16 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F64 = fcmp olt double undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F64 = fcmp olt <2 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp olt <4 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8F64 = fcmp olt <8 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp olt <16 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512-LABEL: 'cmp_float_olt'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F32 = fcmp olt float undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F32 = fcmp olt <2 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F32 = fcmp olt <4 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp olt <8 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp olt <16 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F64 = fcmp olt double undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F64 = fcmp olt <2 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp olt <4 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp olt <8 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp olt <16 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %F32 = fcmp olt float undef, undef
   %V2F32 = fcmp olt <2 x float> undef, undef
@@ -217,18 +490,57 @@ define i32 @cmp_float_olt(i32 %arg) {
 }
 
 define i32 @cmp_float_ueq(i32 %arg) {
-; CHECK-LABEL: 'cmp_float_ueq'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ueq float undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ueq <2 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ueq <4 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ueq <8 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp ueq <16 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ueq double undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ueq <2 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ueq <4 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp ueq <8 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F64 = fcmp ueq <16 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE-LABEL: 'cmp_float_ueq'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %F32 = fcmp ueq float undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V2F32 = fcmp ueq <2 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V4F32 = fcmp ueq <4 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V8F32 = fcmp ueq <8 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V16F32 = fcmp ueq <16 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %F64 = fcmp ueq double undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V2F64 = fcmp ueq <2 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V4F64 = fcmp ueq <4 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V8F64 = fcmp ueq <8 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 40 for instruction: %V16F64 = fcmp ueq <16 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX1-LABEL: 'cmp_float_ueq'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %F32 = fcmp ueq float undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2F32 = fcmp ueq <2 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4F32 = fcmp ueq <4 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8F32 = fcmp ueq <8 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16F32 = fcmp ueq <16 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %F64 = fcmp ueq double undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2F64 = fcmp ueq <2 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4F64 = fcmp ueq <4 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8F64 = fcmp ueq <8 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V16F64 = fcmp ueq <16 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX2-LABEL: 'cmp_float_ueq'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F32 = fcmp ueq float undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F32 = fcmp ueq <2 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F32 = fcmp ueq <4 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp ueq <8 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = fcmp ueq <16 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F64 = fcmp ueq double undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F64 = fcmp ueq <2 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp ueq <4 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8F64 = fcmp ueq <8 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp ueq <16 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512-LABEL: 'cmp_float_ueq'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F32 = fcmp ueq float undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F32 = fcmp ueq <2 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F32 = fcmp ueq <4 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp ueq <8 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp ueq <16 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F64 = fcmp ueq double undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F64 = fcmp ueq <2 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp ueq <4 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp ueq <8 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp ueq <16 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %F32 = fcmp ueq float undef, undef
   %V2F32 = fcmp ueq <2 x float> undef, undef
@@ -246,18 +558,57 @@ define i32 @cmp_float_ueq(i32 %arg) {
 }
 
 define i32 @cmp_float_une(i32 %arg) {
-; CHECK-LABEL: 'cmp_float_une'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp une float undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp une <2 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp une <4 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp une <8 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp une <16 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp une double undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp une <2 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp une <4 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp une <8 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F64 = fcmp une <16 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE-LABEL: 'cmp_float_une'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %F32 = fcmp une float undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V2F32 = fcmp une <2 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V4F32 = fcmp une <4 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V8F32 = fcmp une <8 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V16F32 = fcmp une <16 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %F64 = fcmp une double undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V2F64 = fcmp une <2 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V4F64 = fcmp une <4 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V8F64 = fcmp une <8 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 40 for instruction: %V16F64 = fcmp une <16 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX1-LABEL: 'cmp_float_une'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %F32 = fcmp une float undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2F32 = fcmp une <2 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4F32 = fcmp une <4 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8F32 = fcmp une <8 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16F32 = fcmp une <16 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %F64 = fcmp une double undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2F64 = fcmp une <2 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4F64 = fcmp une <4 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8F64 = fcmp une <8 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V16F64 = fcmp une <16 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX2-LABEL: 'cmp_float_une'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F32 = fcmp une float undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F32 = fcmp une <2 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F32 = fcmp une <4 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp une <8 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = fcmp une <16 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F64 = fcmp une double undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F64 = fcmp une <2 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp une <4 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8F64 = fcmp une <8 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp une <16 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512-LABEL: 'cmp_float_une'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F32 = fcmp une float undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F32 = fcmp une <2 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F32 = fcmp une <4 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp une <8 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp une <16 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F64 = fcmp une double undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F64 = fcmp une <2 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp une <4 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp une <8 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp une <16 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %F32 = fcmp une float undef, undef
   %V2F32 = fcmp une <2 x float> undef, undef
@@ -275,18 +626,57 @@ define i32 @cmp_float_une(i32 %arg) {
 }
 
 define i32 @cmp_float_uno(i32 %arg) {
-; CHECK-LABEL: 'cmp_float_uno'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp uno float undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp uno <2 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp uno <4 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp uno <8 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp uno <16 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp uno double undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp uno <2 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp uno <4 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp uno <8 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F64 = fcmp uno <16 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE-LABEL: 'cmp_float_uno'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %F32 = fcmp uno float undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V2F32 = fcmp uno <2 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V4F32 = fcmp uno <4 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V8F32 = fcmp uno <8 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V16F32 = fcmp uno <16 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %F64 = fcmp uno double undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V2F64 = fcmp uno <2 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V4F64 = fcmp uno <4 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V8F64 = fcmp uno <8 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 40 for instruction: %V16F64 = fcmp uno <16 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX1-LABEL: 'cmp_float_uno'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %F32 = fcmp uno float undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2F32 = fcmp uno <2 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4F32 = fcmp uno <4 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8F32 = fcmp uno <8 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16F32 = fcmp uno <16 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %F64 = fcmp uno double undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2F64 = fcmp uno <2 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4F64 = fcmp uno <4 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8F64 = fcmp uno <8 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V16F64 = fcmp uno <16 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX2-LABEL: 'cmp_float_uno'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F32 = fcmp uno float undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F32 = fcmp uno <2 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F32 = fcmp uno <4 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp uno <8 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = fcmp uno <16 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F64 = fcmp uno double undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F64 = fcmp uno <2 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp uno <4 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8F64 = fcmp uno <8 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp uno <16 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512-LABEL: 'cmp_float_uno'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F32 = fcmp uno float undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F32 = fcmp uno <2 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F32 = fcmp uno <4 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp uno <8 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp uno <16 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F64 = fcmp uno double undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F64 = fcmp uno <2 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp uno <4 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp uno <8 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp uno <16 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %F32 = fcmp uno float undef, undef
   %V2F32 = fcmp uno <2 x float> undef, undef
@@ -304,18 +694,57 @@ define i32 @cmp_float_uno(i32 %arg) {
 }
 
 define i32 @cmp_float_uge(i32 %arg) {
-; CHECK-LABEL: 'cmp_float_uge'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp uge float undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp uge <2 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp uge <4 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp uge <8 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp uge <16 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp uge double undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp uge <2 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp uge <4 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp uge <8 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F64 = fcmp uge <16 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE-LABEL: 'cmp_float_uge'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %F32 = fcmp uge float undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V2F32 = fcmp uge <2 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V4F32 = fcmp uge <4 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V8F32 = fcmp uge <8 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V16F32 = fcmp uge <16 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %F64 = fcmp uge double undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V2F64 = fcmp uge <2 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V4F64 = fcmp uge <4 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V8F64 = fcmp uge <8 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 40 for instruction: %V16F64 = fcmp uge <16 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX1-LABEL: 'cmp_float_uge'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %F32 = fcmp uge float undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2F32 = fcmp uge <2 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4F32 = fcmp uge <4 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8F32 = fcmp uge <8 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16F32 = fcmp uge <16 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %F64 = fcmp uge double undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2F64 = fcmp uge <2 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4F64 = fcmp uge <4 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8F64 = fcmp uge <8 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V16F64 = fcmp uge <16 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX2-LABEL: 'cmp_float_uge'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F32 = fcmp uge float undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F32 = fcmp uge <2 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F32 = fcmp uge <4 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp uge <8 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = fcmp uge <16 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F64 = fcmp uge double undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F64 = fcmp uge <2 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp uge <4 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8F64 = fcmp uge <8 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp uge <16 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512-LABEL: 'cmp_float_uge'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F32 = fcmp uge float undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F32 = fcmp uge <2 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F32 = fcmp uge <4 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp uge <8 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp uge <16 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F64 = fcmp uge double undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F64 = fcmp uge <2 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp uge <4 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp uge <8 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp uge <16 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %F32 = fcmp uge float undef, undef
   %V2F32 = fcmp uge <2 x float> undef, undef
@@ -333,18 +762,57 @@ define i32 @cmp_float_uge(i32 %arg) {
 }
 
 define i32 @cmp_float_ugt(i32 %arg) {
-; CHECK-LABEL: 'cmp_float_ugt'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ugt float undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ugt <2 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ugt <4 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ugt <8 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp ugt <16 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ugt double undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ugt <2 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ugt <4 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp ugt <8 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F64 = fcmp ugt <16 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE-LABEL: 'cmp_float_ugt'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %F32 = fcmp ugt float undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V2F32 = fcmp ugt <2 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V4F32 = fcmp ugt <4 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V8F32 = fcmp ugt <8 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V16F32 = fcmp ugt <16 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %F64 = fcmp ugt double undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V2F64 = fcmp ugt <2 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V4F64 = fcmp ugt <4 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V8F64 = fcmp ugt <8 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 40 for instruction: %V16F64 = fcmp ugt <16 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX1-LABEL: 'cmp_float_ugt'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %F32 = fcmp ugt float undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2F32 = fcmp ugt <2 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4F32 = fcmp ugt <4 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8F32 = fcmp ugt <8 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16F32 = fcmp ugt <16 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %F64 = fcmp ugt double undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2F64 = fcmp ugt <2 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4F64 = fcmp ugt <4 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8F64 = fcmp ugt <8 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V16F64 = fcmp ugt <16 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX2-LABEL: 'cmp_float_ugt'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F32 = fcmp ugt float undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F32 = fcmp ugt <2 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F32 = fcmp ugt <4 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp ugt <8 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = fcmp ugt <16 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F64 = fcmp ugt double undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F64 = fcmp ugt <2 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp ugt <4 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8F64 = fcmp ugt <8 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp ugt <16 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512-LABEL: 'cmp_float_ugt'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F32 = fcmp ugt float undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F32 = fcmp ugt <2 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F32 = fcmp ugt <4 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp ugt <8 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp ugt <16 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F64 = fcmp ugt double undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F64 = fcmp ugt <2 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp ugt <4 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp ugt <8 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp ugt <16 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %F32 = fcmp ugt float undef, undef
   %V2F32 = fcmp ugt <2 x float> undef, undef
@@ -362,18 +830,57 @@ define i32 @cmp_float_ugt(i32 %arg) {
 }
 
 define i32 @cmp_float_ule(i32 %arg) {
-; CHECK-LABEL: 'cmp_float_ule'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ule float undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ule <2 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ule <4 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ule <8 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp ule <16 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ule double undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ule <2 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ule <4 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp ule <8 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F64 = fcmp ule <16 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE-LABEL: 'cmp_float_ule'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %F32 = fcmp ule float undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V2F32 = fcmp ule <2 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V4F32 = fcmp ule <4 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V8F32 = fcmp ule <8 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V16F32 = fcmp ule <16 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %F64 = fcmp ule double undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V2F64 = fcmp ule <2 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V4F64 = fcmp ule <4 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V8F64 = fcmp ule <8 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 40 for instruction: %V16F64 = fcmp ule <16 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX1-LABEL: 'cmp_float_ule'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %F32 = fcmp ule float undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2F32 = fcmp ule <2 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4F32 = fcmp ule <4 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8F32 = fcmp ule <8 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16F32 = fcmp ule <16 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %F64 = fcmp ule double undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2F64 = fcmp ule <2 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4F64 = fcmp ule <4 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8F64 = fcmp ule <8 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V16F64 = fcmp ule <16 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX2-LABEL: 'cmp_float_ule'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F32 = fcmp ule float undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F32 = fcmp ule <2 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F32 = fcmp ule <4 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp ule <8 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = fcmp ule <16 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F64 = fcmp ule double undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F64 = fcmp ule <2 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp ule <4 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8F64 = fcmp ule <8 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp ule <16 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512-LABEL: 'cmp_float_ule'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F32 = fcmp ule float undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F32 = fcmp ule <2 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F32 = fcmp ule <4 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp ule <8 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp ule <16 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F64 = fcmp ule double undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F64 = fcmp ule <2 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp ule <4 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp ule <8 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp ule <16 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %F32 = fcmp ule float undef, undef
   %V2F32 = fcmp ule <2 x float> undef, undef
@@ -391,18 +898,57 @@ define i32 @cmp_float_ule(i32 %arg) {
 }
 
 define i32 @cmp_float_ult(i32 %arg) {
-; CHECK-LABEL: 'cmp_float_ult'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ult float undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ult <2 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ult <4 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ult <8 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp ult <16 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ult double undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ult <2 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ult <4 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp ult <8 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F64 = fcmp ult <16 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE-LABEL: 'cmp_float_ult'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %F32 = fcmp ult float undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V2F32 = fcmp ult <2 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V4F32 = fcmp ult <4 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V8F32 = fcmp ult <8 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V16F32 = fcmp ult <16 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %F64 = fcmp ult double undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V2F64 = fcmp ult <2 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V4F64 = fcmp ult <4 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V8F64 = fcmp ult <8 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 40 for instruction: %V16F64 = fcmp ult <16 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX1-LABEL: 'cmp_float_ult'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %F32 = fcmp ult float undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2F32 = fcmp ult <2 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4F32 = fcmp ult <4 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8F32 = fcmp ult <8 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16F32 = fcmp ult <16 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %F64 = fcmp ult double undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2F64 = fcmp ult <2 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4F64 = fcmp ult <4 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8F64 = fcmp ult <8 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V16F64 = fcmp ult <16 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX2-LABEL: 'cmp_float_ult'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F32 = fcmp ult float undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F32 = fcmp ult <2 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F32 = fcmp ult <4 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp ult <8 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = fcmp ult <16 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F64 = fcmp ult double undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F64 = fcmp ult <2 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp ult <4 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8F64 = fcmp ult <8 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp ult <16 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512-LABEL: 'cmp_float_ult'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F32 = fcmp ult float undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F32 = fcmp ult <2 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F32 = fcmp ult <4 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp ult <8 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp ult <16 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F64 = fcmp ult double undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F64 = fcmp ult <2 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp ult <4 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp ult <8 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp ult <16 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %F32 = fcmp ult float undef, undef
   %V2F32 = fcmp ult <2 x float> undef, undef
@@ -420,18 +966,57 @@ define i32 @cmp_float_ult(i32 %arg) {
 }
 
 define i32 @cmp_float_false(i32 %arg) {
-; CHECK-LABEL: 'cmp_float_false'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp false float undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp false <2 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp false <4 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp false <8 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp false <16 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp false double undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp false <2 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp false <4 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp false <8 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F64 = fcmp false <16 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE-LABEL: 'cmp_float_false'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %F32 = fcmp false float undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V2F32 = fcmp false <2 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V4F32 = fcmp false <4 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V8F32 = fcmp false <8 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V16F32 = fcmp false <16 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %F64 = fcmp false double undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V2F64 = fcmp false <2 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V4F64 = fcmp false <4 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V8F64 = fcmp false <8 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 40 for instruction: %V16F64 = fcmp false <16 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX1-LABEL: 'cmp_float_false'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %F32 = fcmp false float undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2F32 = fcmp false <2 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4F32 = fcmp false <4 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8F32 = fcmp false <8 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16F32 = fcmp false <16 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %F64 = fcmp false double undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2F64 = fcmp false <2 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4F64 = fcmp false <4 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8F64 = fcmp false <8 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V16F64 = fcmp false <16 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX2-LABEL: 'cmp_float_false'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F32 = fcmp false float undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F32 = fcmp false <2 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F32 = fcmp false <4 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp false <8 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = fcmp false <16 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F64 = fcmp false double undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F64 = fcmp false <2 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp false <4 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8F64 = fcmp false <8 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp false <16 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512-LABEL: 'cmp_float_false'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F32 = fcmp false float undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F32 = fcmp false <2 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F32 = fcmp false <4 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp false <8 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp false <16 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F64 = fcmp false double undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F64 = fcmp false <2 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp false <4 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp false <8 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp false <16 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %F32 = fcmp false float undef, undef
   %V2F32 = fcmp false <2 x float> undef, undef
@@ -449,18 +1034,57 @@ define i32 @cmp_float_false(i32 %arg) {
 }
 
 define i32 @cmp_float_true(i32 %arg) {
-; CHECK-LABEL: 'cmp_float_true'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp true float undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp true <2 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp true <4 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp true <8 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp true <16 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp true double undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp true <2 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp true <4 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp true <8 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F64 = fcmp true <16 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE-LABEL: 'cmp_float_true'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %F32 = fcmp true float undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V2F32 = fcmp true <2 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V4F32 = fcmp true <4 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V8F32 = fcmp true <8 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V16F32 = fcmp true <16 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %F64 = fcmp true double undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 5 for instruction: %V2F64 = fcmp true <2 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V4F64 = fcmp true <4 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V8F64 = fcmp true <8 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 40 for instruction: %V16F64 = fcmp true <16 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX1-LABEL: 'cmp_float_true'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %F32 = fcmp true float undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2F32 = fcmp true <2 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4F32 = fcmp true <4 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V8F32 = fcmp true <8 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V16F32 = fcmp true <16 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %F64 = fcmp true double undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2F64 = fcmp true <2 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V4F64 = fcmp true <4 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8F64 = fcmp true <8 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V16F64 = fcmp true <16 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX2-LABEL: 'cmp_float_true'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F32 = fcmp true float undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F32 = fcmp true <2 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F32 = fcmp true <4 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp true <8 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = fcmp true <16 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F64 = fcmp true double undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F64 = fcmp true <2 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp true <4 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8F64 = fcmp true <8 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp true <16 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512-LABEL: 'cmp_float_true'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F32 = fcmp true float undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F32 = fcmp true <2 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F32 = fcmp true <4 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp true <8 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp true <16 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %F64 = fcmp true double undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2F64 = fcmp true <2 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp true <4 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp true <8 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp true <16 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %F32 = fcmp true float undef, undef
   %V2F32 = fcmp true <2 x float> undef, undef

diff  --git a/llvm/test/Analysis/CostModel/X86/fcmp-sizelatency.ll b/llvm/test/Analysis/CostModel/X86/fcmp-sizelatency.ll
index cf2111fc433f3..e07a834dfb76d 100644
--- a/llvm/test/Analysis/CostModel/X86/fcmp-sizelatency.ll
+++ b/llvm/test/Analysis/CostModel/X86/fcmp-sizelatency.ll
@@ -1,31 +1,57 @@
 ; NOTE: Assertions have been autogenerated by utils/update_analyze_test_checks.py
-; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mattr=+sse2 | FileCheck %s
-; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mattr=+sse3 | FileCheck %s
-; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mattr=+ssse3 | FileCheck %s
-; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mattr=+sse4.1 | FileCheck %s
-; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mattr=+sse4.2 | FileCheck %s
-; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mattr=+avx | FileCheck %s
-; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mattr=+avx2 | FileCheck %s
-; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mattr=+avx512f | FileCheck %s
-; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mattr=+avx512f,+avx512bw | FileCheck %s
-;
-; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mcpu=slm | FileCheck %s
-; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mcpu=goldmont | FileCheck %s
-; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mcpu=btver2 | FileCheck %s
+; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mattr=+sse2 | FileCheck %s --check-prefixes=SSE
+; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mattr=+sse3 | FileCheck %s --check-prefixes=SSE
+; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mattr=+ssse3 | FileCheck %s --check-prefixes=SSE
+; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mattr=+sse4.1 | FileCheck %s --check-prefixes=SSE
+; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mattr=+sse4.2 | FileCheck %s --check-prefixes=SSE
+; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mattr=+avx | FileCheck %s --check-prefixes=AVX
+; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mattr=+avx2 | FileCheck %s --check-prefixes=AVX
+; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mattr=+avx512f | FileCheck %s --check-prefixes=AVX512
+; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefixes=AVX512
+;
+; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mcpu=slm | FileCheck %s --check-prefixes=SSE
+; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mcpu=goldmont | FileCheck %s --check-prefixes=SSE
+; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -cost-kind=size-latency -mcpu=btver2 | FileCheck %s --check-prefixes=AVX
 
 define i32 @cmp_float_oeq(i32 %arg) {
-; CHECK-LABEL: 'cmp_float_oeq'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp oeq float undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp oeq <2 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp oeq <4 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp oeq <8 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp oeq <16 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp oeq double undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp oeq <2 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp oeq <4 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp oeq <8 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F64 = fcmp oeq <16 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE-LABEL: 'cmp_float_oeq'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp oeq float undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp oeq <2 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp oeq <4 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp oeq <8 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp oeq <16 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp oeq double undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp oeq <2 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp oeq <4 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp oeq <8 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp oeq <16 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX-LABEL: 'cmp_float_oeq'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp oeq float undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp oeq <2 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp oeq <4 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp oeq <8 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp oeq <16 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp oeq double undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp oeq <2 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp oeq <4 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp oeq <8 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp oeq <16 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512-LABEL: 'cmp_float_oeq'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp oeq float undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp oeq <2 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp oeq <4 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp oeq <8 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp oeq <16 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp oeq double undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp oeq <2 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp oeq <4 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp oeq <8 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F64 = fcmp oeq <16 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %F32 = fcmp oeq float undef, undef
   %V2F32 = fcmp oeq <2 x float> undef, undef
@@ -43,18 +69,44 @@ define i32 @cmp_float_oeq(i32 %arg) {
 }
 
 define i32 @cmp_float_one(i32 %arg) {
-; CHECK-LABEL: 'cmp_float_one'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp one float undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp one <2 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp one <4 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp one <8 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp one <16 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp one double undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp one <2 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp one <4 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp one <8 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F64 = fcmp one <16 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE-LABEL: 'cmp_float_one'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp one float undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp one <2 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp one <4 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp one <8 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp one <16 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp one double undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp one <2 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp one <4 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp one <8 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp one <16 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX-LABEL: 'cmp_float_one'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp one float undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp one <2 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp one <4 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp one <8 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp one <16 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp one double undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp one <2 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp one <4 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp one <8 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp one <16 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512-LABEL: 'cmp_float_one'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp one float undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp one <2 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp one <4 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp one <8 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp one <16 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp one double undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp one <2 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp one <4 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp one <8 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F64 = fcmp one <16 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %F32 = fcmp one float undef, undef
   %V2F32 = fcmp one <2 x float> undef, undef
@@ -72,18 +124,44 @@ define i32 @cmp_float_one(i32 %arg) {
 }
 
 define i32 @cmp_float_ord(i32 %arg) {
-; CHECK-LABEL: 'cmp_float_ord'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ord float undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ord <2 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ord <4 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ord <8 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp ord <16 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ord double undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ord <2 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ord <4 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp ord <8 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F64 = fcmp ord <16 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE-LABEL: 'cmp_float_ord'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ord float undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ord <2 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ord <4 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp ord <8 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp ord <16 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ord double undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ord <2 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp ord <4 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp ord <8 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp ord <16 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX-LABEL: 'cmp_float_ord'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ord float undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ord <2 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ord <4 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp ord <8 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp ord <16 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ord double undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ord <2 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp ord <4 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp ord <8 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp ord <16 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512-LABEL: 'cmp_float_ord'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ord float undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ord <2 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ord <4 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ord <8 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp ord <16 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ord double undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ord <2 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ord <4 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp ord <8 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F64 = fcmp ord <16 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %F32 = fcmp ord float undef, undef
   %V2F32 = fcmp ord <2 x float> undef, undef
@@ -101,18 +179,44 @@ define i32 @cmp_float_ord(i32 %arg) {
 }
 
 define i32 @cmp_float_oge(i32 %arg) {
-; CHECK-LABEL: 'cmp_float_oge'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp oge float undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp oge <2 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp oge <4 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp oge <8 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp oge <16 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp oge double undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp oge <2 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp oge <4 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp oge <8 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F64 = fcmp oge <16 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE-LABEL: 'cmp_float_oge'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp oge float undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp oge <2 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp oge <4 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp oge <8 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp oge <16 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp oge double undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp oge <2 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp oge <4 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp oge <8 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp oge <16 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX-LABEL: 'cmp_float_oge'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp oge float undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp oge <2 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp oge <4 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp oge <8 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp oge <16 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp oge double undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp oge <2 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp oge <4 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp oge <8 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp oge <16 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512-LABEL: 'cmp_float_oge'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp oge float undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp oge <2 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp oge <4 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp oge <8 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp oge <16 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp oge double undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp oge <2 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp oge <4 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp oge <8 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F64 = fcmp oge <16 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %F32 = fcmp oge float undef, undef
   %V2F32 = fcmp oge <2 x float> undef, undef
@@ -130,18 +234,44 @@ define i32 @cmp_float_oge(i32 %arg) {
 }
 
 define i32 @cmp_float_ogt(i32 %arg) {
-; CHECK-LABEL: 'cmp_float_ogt'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ogt float undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ogt <2 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ogt <4 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ogt <8 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp ogt <16 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ogt double undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ogt <2 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ogt <4 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp ogt <8 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F64 = fcmp ogt <16 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE-LABEL: 'cmp_float_ogt'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ogt float undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ogt <2 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ogt <4 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp ogt <8 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp ogt <16 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ogt double undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ogt <2 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp ogt <4 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp ogt <8 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp ogt <16 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX-LABEL: 'cmp_float_ogt'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ogt float undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ogt <2 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ogt <4 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp ogt <8 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp ogt <16 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ogt double undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ogt <2 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp ogt <4 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp ogt <8 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp ogt <16 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512-LABEL: 'cmp_float_ogt'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ogt float undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ogt <2 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ogt <4 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ogt <8 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp ogt <16 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ogt double undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ogt <2 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ogt <4 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp ogt <8 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F64 = fcmp ogt <16 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %F32 = fcmp ogt float undef, undef
   %V2F32 = fcmp ogt <2 x float> undef, undef
@@ -159,18 +289,44 @@ define i32 @cmp_float_ogt(i32 %arg) {
 }
 
 define i32 @cmp_float_ole(i32 %arg) {
-; CHECK-LABEL: 'cmp_float_ole'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ole float undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ole <2 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ole <4 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ole <8 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp ole <16 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ole double undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ole <2 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ole <4 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp ole <8 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F64 = fcmp ole <16 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE-LABEL: 'cmp_float_ole'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ole float undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ole <2 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ole <4 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp ole <8 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp ole <16 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ole double undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ole <2 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp ole <4 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp ole <8 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp ole <16 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX-LABEL: 'cmp_float_ole'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ole float undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ole <2 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ole <4 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp ole <8 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp ole <16 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ole double undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ole <2 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp ole <4 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp ole <8 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp ole <16 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512-LABEL: 'cmp_float_ole'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ole float undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ole <2 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ole <4 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ole <8 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp ole <16 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ole double undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ole <2 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ole <4 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp ole <8 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F64 = fcmp ole <16 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %F32 = fcmp ole float undef, undef
   %V2F32 = fcmp ole <2 x float> undef, undef
@@ -188,18 +344,44 @@ define i32 @cmp_float_ole(i32 %arg) {
 }
 
 define i32 @cmp_float_olt(i32 %arg) {
-; CHECK-LABEL: 'cmp_float_olt'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp olt float undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp olt <2 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp olt <4 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp olt <8 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp olt <16 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp olt double undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp olt <2 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp olt <4 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp olt <8 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F64 = fcmp olt <16 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE-LABEL: 'cmp_float_olt'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp olt float undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp olt <2 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp olt <4 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp olt <8 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp olt <16 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp olt double undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp olt <2 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp olt <4 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp olt <8 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp olt <16 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX-LABEL: 'cmp_float_olt'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp olt float undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp olt <2 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp olt <4 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp olt <8 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp olt <16 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp olt double undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp olt <2 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp olt <4 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp olt <8 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp olt <16 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512-LABEL: 'cmp_float_olt'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp olt float undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp olt <2 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp olt <4 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp olt <8 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp olt <16 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp olt double undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp olt <2 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp olt <4 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp olt <8 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F64 = fcmp olt <16 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %F32 = fcmp olt float undef, undef
   %V2F32 = fcmp olt <2 x float> undef, undef
@@ -217,18 +399,44 @@ define i32 @cmp_float_olt(i32 %arg) {
 }
 
 define i32 @cmp_float_ueq(i32 %arg) {
-; CHECK-LABEL: 'cmp_float_ueq'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ueq float undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ueq <2 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ueq <4 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ueq <8 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp ueq <16 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ueq double undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ueq <2 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ueq <4 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp ueq <8 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F64 = fcmp ueq <16 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE-LABEL: 'cmp_float_ueq'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ueq float undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ueq <2 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ueq <4 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp ueq <8 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp ueq <16 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ueq double undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ueq <2 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp ueq <4 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp ueq <8 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp ueq <16 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX-LABEL: 'cmp_float_ueq'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ueq float undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ueq <2 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ueq <4 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp ueq <8 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp ueq <16 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ueq double undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ueq <2 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp ueq <4 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp ueq <8 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp ueq <16 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512-LABEL: 'cmp_float_ueq'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ueq float undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ueq <2 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ueq <4 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ueq <8 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp ueq <16 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ueq double undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ueq <2 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ueq <4 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp ueq <8 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F64 = fcmp ueq <16 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %F32 = fcmp ueq float undef, undef
   %V2F32 = fcmp ueq <2 x float> undef, undef
@@ -246,18 +454,44 @@ define i32 @cmp_float_ueq(i32 %arg) {
 }
 
 define i32 @cmp_float_une(i32 %arg) {
-; CHECK-LABEL: 'cmp_float_une'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp une float undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp une <2 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp une <4 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp une <8 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp une <16 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp une double undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp une <2 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp une <4 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp une <8 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F64 = fcmp une <16 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE-LABEL: 'cmp_float_une'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp une float undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp une <2 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp une <4 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp une <8 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp une <16 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp une double undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp une <2 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp une <4 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp une <8 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp une <16 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX-LABEL: 'cmp_float_une'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp une float undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp une <2 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp une <4 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp une <8 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp une <16 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp une double undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp une <2 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp une <4 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp une <8 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp une <16 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512-LABEL: 'cmp_float_une'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp une float undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp une <2 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp une <4 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp une <8 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp une <16 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp une double undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp une <2 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp une <4 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp une <8 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F64 = fcmp une <16 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %F32 = fcmp une float undef, undef
   %V2F32 = fcmp une <2 x float> undef, undef
@@ -275,18 +509,44 @@ define i32 @cmp_float_une(i32 %arg) {
 }
 
 define i32 @cmp_float_uno(i32 %arg) {
-; CHECK-LABEL: 'cmp_float_uno'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp uno float undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp uno <2 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp uno <4 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp uno <8 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp uno <16 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp uno double undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp uno <2 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp uno <4 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp uno <8 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F64 = fcmp uno <16 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE-LABEL: 'cmp_float_uno'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp uno float undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp uno <2 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp uno <4 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp uno <8 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp uno <16 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp uno double undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp uno <2 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp uno <4 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp uno <8 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp uno <16 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX-LABEL: 'cmp_float_uno'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp uno float undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp uno <2 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp uno <4 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp uno <8 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp uno <16 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp uno double undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp uno <2 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp uno <4 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp uno <8 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp uno <16 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512-LABEL: 'cmp_float_uno'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp uno float undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp uno <2 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp uno <4 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp uno <8 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp uno <16 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp uno double undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp uno <2 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp uno <4 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp uno <8 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F64 = fcmp uno <16 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %F32 = fcmp uno float undef, undef
   %V2F32 = fcmp uno <2 x float> undef, undef
@@ -304,18 +564,44 @@ define i32 @cmp_float_uno(i32 %arg) {
 }
 
 define i32 @cmp_float_uge(i32 %arg) {
-; CHECK-LABEL: 'cmp_float_uge'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp uge float undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp uge <2 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp uge <4 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp uge <8 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp uge <16 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp uge double undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp uge <2 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp uge <4 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp uge <8 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F64 = fcmp uge <16 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE-LABEL: 'cmp_float_uge'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp uge float undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp uge <2 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp uge <4 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp uge <8 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp uge <16 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp uge double undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp uge <2 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp uge <4 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp uge <8 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp uge <16 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX-LABEL: 'cmp_float_uge'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp uge float undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp uge <2 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp uge <4 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp uge <8 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp uge <16 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp uge double undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp uge <2 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp uge <4 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp uge <8 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp uge <16 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512-LABEL: 'cmp_float_uge'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp uge float undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp uge <2 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp uge <4 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp uge <8 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp uge <16 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp uge double undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp uge <2 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp uge <4 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp uge <8 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F64 = fcmp uge <16 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %F32 = fcmp uge float undef, undef
   %V2F32 = fcmp uge <2 x float> undef, undef
@@ -333,18 +619,44 @@ define i32 @cmp_float_uge(i32 %arg) {
 }
 
 define i32 @cmp_float_ugt(i32 %arg) {
-; CHECK-LABEL: 'cmp_float_ugt'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ugt float undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ugt <2 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ugt <4 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ugt <8 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp ugt <16 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ugt double undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ugt <2 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ugt <4 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp ugt <8 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F64 = fcmp ugt <16 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE-LABEL: 'cmp_float_ugt'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ugt float undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ugt <2 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ugt <4 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp ugt <8 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp ugt <16 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ugt double undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ugt <2 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp ugt <4 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp ugt <8 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp ugt <16 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX-LABEL: 'cmp_float_ugt'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ugt float undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ugt <2 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ugt <4 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp ugt <8 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp ugt <16 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ugt double undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ugt <2 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp ugt <4 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp ugt <8 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp ugt <16 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512-LABEL: 'cmp_float_ugt'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ugt float undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ugt <2 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ugt <4 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ugt <8 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp ugt <16 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ugt double undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ugt <2 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ugt <4 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp ugt <8 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F64 = fcmp ugt <16 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %F32 = fcmp ugt float undef, undef
   %V2F32 = fcmp ugt <2 x float> undef, undef
@@ -362,18 +674,44 @@ define i32 @cmp_float_ugt(i32 %arg) {
 }
 
 define i32 @cmp_float_ule(i32 %arg) {
-; CHECK-LABEL: 'cmp_float_ule'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ule float undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ule <2 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ule <4 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ule <8 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp ule <16 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ule double undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ule <2 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ule <4 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp ule <8 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F64 = fcmp ule <16 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE-LABEL: 'cmp_float_ule'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ule float undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ule <2 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ule <4 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp ule <8 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp ule <16 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ule double undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ule <2 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp ule <4 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp ule <8 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp ule <16 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX-LABEL: 'cmp_float_ule'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ule float undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ule <2 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ule <4 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp ule <8 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp ule <16 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ule double undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ule <2 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp ule <4 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp ule <8 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp ule <16 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512-LABEL: 'cmp_float_ule'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ule float undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ule <2 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ule <4 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ule <8 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp ule <16 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ule double undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ule <2 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ule <4 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp ule <8 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F64 = fcmp ule <16 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %F32 = fcmp ule float undef, undef
   %V2F32 = fcmp ule <2 x float> undef, undef
@@ -391,18 +729,44 @@ define i32 @cmp_float_ule(i32 %arg) {
 }
 
 define i32 @cmp_float_ult(i32 %arg) {
-; CHECK-LABEL: 'cmp_float_ult'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ult float undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ult <2 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ult <4 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ult <8 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp ult <16 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ult double undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ult <2 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ult <4 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp ult <8 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F64 = fcmp ult <16 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE-LABEL: 'cmp_float_ult'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ult float undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ult <2 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ult <4 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp ult <8 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp ult <16 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ult double undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ult <2 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp ult <4 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp ult <8 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp ult <16 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX-LABEL: 'cmp_float_ult'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ult float undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ult <2 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ult <4 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp ult <8 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp ult <16 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ult double undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ult <2 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp ult <4 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp ult <8 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp ult <16 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512-LABEL: 'cmp_float_ult'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ult float undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ult <2 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ult <4 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ult <8 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp ult <16 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ult double undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ult <2 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ult <4 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp ult <8 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F64 = fcmp ult <16 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %F32 = fcmp ult float undef, undef
   %V2F32 = fcmp ult <2 x float> undef, undef
@@ -420,18 +784,44 @@ define i32 @cmp_float_ult(i32 %arg) {
 }
 
 define i32 @cmp_float_false(i32 %arg) {
-; CHECK-LABEL: 'cmp_float_false'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp false float undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp false <2 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp false <4 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp false <8 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp false <16 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp false double undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp false <2 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp false <4 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp false <8 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F64 = fcmp false <16 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE-LABEL: 'cmp_float_false'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp false float undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp false <2 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp false <4 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp false <8 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp false <16 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp false double undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp false <2 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp false <4 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp false <8 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp false <16 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX-LABEL: 'cmp_float_false'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp false float undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp false <2 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp false <4 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp false <8 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp false <16 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp false double undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp false <2 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp false <4 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp false <8 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp false <16 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512-LABEL: 'cmp_float_false'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp false float undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp false <2 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp false <4 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp false <8 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp false <16 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp false double undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp false <2 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp false <4 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp false <8 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F64 = fcmp false <16 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %F32 = fcmp false float undef, undef
   %V2F32 = fcmp false <2 x float> undef, undef
@@ -449,18 +839,44 @@ define i32 @cmp_float_false(i32 %arg) {
 }
 
 define i32 @cmp_float_true(i32 %arg) {
-; CHECK-LABEL: 'cmp_float_true'
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp true float undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp true <2 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp true <4 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp true <8 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp true <16 x float> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp true double undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp true <2 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp true <4 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp true <8 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F64 = fcmp true <16 x double> undef, undef
-; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+; SSE-LABEL: 'cmp_float_true'
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp true float undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp true <2 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp true <4 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp true <8 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp true <16 x float> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp true double undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp true <2 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp true <4 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp true <8 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp true <16 x double> undef, undef
+; SSE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX-LABEL: 'cmp_float_true'
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp true float undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp true <2 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp true <4 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp true <8 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp true <16 x float> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp true double undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp true <2 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp true <4 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp true <8 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp true <16 x double> undef, undef
+; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; AVX512-LABEL: 'cmp_float_true'
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp true float undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp true <2 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp true <4 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp true <8 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = fcmp true <16 x float> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp true double undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp true <2 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp true <4 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = fcmp true <8 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F64 = fcmp true <16 x double> undef, undef
+; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
 ;
   %F32 = fcmp true float undef, undef
   %V2F32 = fcmp true <2 x float> undef, undef

diff  --git a/llvm/test/Analysis/CostModel/X86/fcmp.ll b/llvm/test/Analysis/CostModel/X86/fcmp.ll
index 0103b63a0e6c9..2a9fbfbf1027f 100644
--- a/llvm/test/Analysis/CostModel/X86/fcmp.ll
+++ b/llvm/test/Analysis/CostModel/X86/fcmp.ll
@@ -1,17 +1,17 @@
 ; NOTE: Assertions have been autogenerated by utils/update_analyze_test_checks.py
 ; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -mattr=+sse2 | FileCheck %s --check-prefixes=SSE2
-; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -mattr=+sse3 | FileCheck %s --check-prefixes=SSE3
-; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -mattr=+ssse3 | FileCheck %s --check-prefixes=SSSE3
+; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -mattr=+sse3 | FileCheck %s --check-prefixes=SSE2
+; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -mattr=+ssse3 | FileCheck %s --check-prefixes=SSE2
 ; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -mattr=+sse4.1 | FileCheck %s --check-prefixes=SSE41
-; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -mattr=+sse4.2 | FileCheck %s --check-prefixes=SSE42
-; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -mattr=+avx | FileCheck %s --check-prefixes=AVX
-; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -mattr=+avx2 | FileCheck %s --check-prefixes=AVX
+; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -mattr=+sse4.2 | FileCheck %s --check-prefixes=SSE41
+; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -mattr=+avx | FileCheck %s --check-prefixes=AVX1
+; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -mattr=+avx2 | FileCheck %s --check-prefixes=AVX2
 ; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -mattr=+avx512f | FileCheck %s --check-prefixes=AVX512
 ; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefixes=AVX512
 ;
-; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -mcpu=slm | FileCheck %s --check-prefixes=SSE42
-; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -mcpu=goldmont | FileCheck %s --check-prefixes=SSE42
-; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -mcpu=btver2 | FileCheck %s --check-prefixes=AVX
+; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -mcpu=slm | FileCheck %s --check-prefixes=SSE41
+; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -mcpu=goldmont | FileCheck %s --check-prefixes=SSE41
+; RUN: opt < %s -mtriple=x86_64-- -passes="print<cost-model>" 2>&1 -disable-output -mcpu=btver2 | FileCheck %s --check-prefixes=AVX1
 
 define i32 @cmp_float_oeq(i32 %arg) {
 ; SSE2-LABEL: 'cmp_float_oeq'
@@ -27,32 +27,6 @@ define i32 @cmp_float_oeq(i32 %arg) {
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp oeq <16 x double> undef, undef
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
-; SSE3-LABEL: 'cmp_float_oeq'
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp oeq float undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F32 = fcmp oeq <2 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F32 = fcmp oeq <4 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp oeq <8 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = fcmp oeq <16 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp oeq double undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F64 = fcmp oeq <2 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp oeq <4 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8F64 = fcmp oeq <8 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp oeq <16 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
-; SSSE3-LABEL: 'cmp_float_oeq'
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp oeq float undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F32 = fcmp oeq <2 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F32 = fcmp oeq <4 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp oeq <8 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = fcmp oeq <16 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp oeq double undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F64 = fcmp oeq <2 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp oeq <4 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8F64 = fcmp oeq <8 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp oeq <16 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
 ; SSE41-LABEL: 'cmp_float_oeq'
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp oeq float undef, undef
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp oeq <2 x float> undef, undef
@@ -66,31 +40,31 @@ define i32 @cmp_float_oeq(i32 %arg) {
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp oeq <16 x double> undef, undef
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
-; SSE42-LABEL: 'cmp_float_oeq'
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp oeq float undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp oeq <2 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp oeq <4 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp oeq <8 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp oeq <16 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp oeq double undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp oeq <2 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp oeq <4 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp oeq <8 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp oeq <16 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
-; AVX-LABEL: 'cmp_float_oeq'
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp oeq float undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp oeq <2 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp oeq <4 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp oeq <8 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F32 = fcmp oeq <16 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp oeq double undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp oeq <2 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp oeq <4 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F64 = fcmp oeq <8 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F64 = fcmp oeq <16 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+; AVX1-LABEL: 'cmp_float_oeq'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp oeq float undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp oeq <2 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp oeq <4 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp oeq <8 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp oeq <16 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp oeq double undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp oeq <2 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp oeq <4 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp oeq <8 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp oeq <16 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+;
+; AVX2-LABEL: 'cmp_float_oeq'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp oeq float undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp oeq <2 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp oeq <4 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp oeq <8 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F32 = fcmp oeq <16 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp oeq double undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp oeq <2 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp oeq <4 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F64 = fcmp oeq <8 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F64 = fcmp oeq <16 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; AVX512-LABEL: 'cmp_float_oeq'
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp oeq float undef, undef
@@ -134,32 +108,6 @@ define i32 @cmp_float_one(i32 %arg) {
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp one <16 x double> undef, undef
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
-; SSE3-LABEL: 'cmp_float_one'
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp one float undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F32 = fcmp one <2 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F32 = fcmp one <4 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp one <8 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = fcmp one <16 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp one double undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F64 = fcmp one <2 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp one <4 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8F64 = fcmp one <8 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp one <16 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
-; SSSE3-LABEL: 'cmp_float_one'
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp one float undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F32 = fcmp one <2 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F32 = fcmp one <4 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp one <8 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = fcmp one <16 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp one double undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F64 = fcmp one <2 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp one <4 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8F64 = fcmp one <8 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp one <16 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
 ; SSE41-LABEL: 'cmp_float_one'
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp one float undef, undef
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp one <2 x float> undef, undef
@@ -173,31 +121,31 @@ define i32 @cmp_float_one(i32 %arg) {
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp one <16 x double> undef, undef
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
-; SSE42-LABEL: 'cmp_float_one'
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp one float undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp one <2 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp one <4 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp one <8 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp one <16 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp one double undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp one <2 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp one <4 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp one <8 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp one <16 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
-; AVX-LABEL: 'cmp_float_one'
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp one float undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp one <2 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp one <4 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp one <8 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F32 = fcmp one <16 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp one double undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp one <2 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp one <4 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F64 = fcmp one <8 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F64 = fcmp one <16 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+; AVX1-LABEL: 'cmp_float_one'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp one float undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp one <2 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp one <4 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp one <8 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp one <16 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp one double undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp one <2 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp one <4 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp one <8 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp one <16 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+;
+; AVX2-LABEL: 'cmp_float_one'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp one float undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp one <2 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp one <4 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp one <8 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F32 = fcmp one <16 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp one double undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp one <2 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp one <4 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F64 = fcmp one <8 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F64 = fcmp one <16 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; AVX512-LABEL: 'cmp_float_one'
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp one float undef, undef
@@ -241,32 +189,6 @@ define i32 @cmp_float_ord(i32 %arg) {
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp ord <16 x double> undef, undef
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
-; SSE3-LABEL: 'cmp_float_ord'
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ord float undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F32 = fcmp ord <2 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F32 = fcmp ord <4 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp ord <8 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = fcmp ord <16 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ord double undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F64 = fcmp ord <2 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp ord <4 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8F64 = fcmp ord <8 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp ord <16 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
-; SSSE3-LABEL: 'cmp_float_ord'
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ord float undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F32 = fcmp ord <2 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F32 = fcmp ord <4 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp ord <8 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = fcmp ord <16 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ord double undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F64 = fcmp ord <2 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp ord <4 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8F64 = fcmp ord <8 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp ord <16 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
 ; SSE41-LABEL: 'cmp_float_ord'
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ord float undef, undef
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ord <2 x float> undef, undef
@@ -280,31 +202,31 @@ define i32 @cmp_float_ord(i32 %arg) {
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp ord <16 x double> undef, undef
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
-; SSE42-LABEL: 'cmp_float_ord'
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ord float undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ord <2 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ord <4 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp ord <8 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp ord <16 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ord double undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ord <2 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp ord <4 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp ord <8 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp ord <16 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
-; AVX-LABEL: 'cmp_float_ord'
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ord float undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ord <2 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ord <4 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ord <8 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F32 = fcmp ord <16 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ord double undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ord <2 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ord <4 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F64 = fcmp ord <8 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F64 = fcmp ord <16 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+; AVX1-LABEL: 'cmp_float_ord'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ord float undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ord <2 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ord <4 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp ord <8 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp ord <16 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ord double undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ord <2 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp ord <4 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp ord <8 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp ord <16 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+;
+; AVX2-LABEL: 'cmp_float_ord'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ord float undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ord <2 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ord <4 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ord <8 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F32 = fcmp ord <16 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ord double undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ord <2 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ord <4 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F64 = fcmp ord <8 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F64 = fcmp ord <16 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; AVX512-LABEL: 'cmp_float_ord'
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ord float undef, undef
@@ -348,32 +270,6 @@ define i32 @cmp_float_oge(i32 %arg) {
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp oge <16 x double> undef, undef
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
-; SSE3-LABEL: 'cmp_float_oge'
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp oge float undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F32 = fcmp oge <2 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F32 = fcmp oge <4 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp oge <8 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = fcmp oge <16 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp oge double undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F64 = fcmp oge <2 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp oge <4 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8F64 = fcmp oge <8 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp oge <16 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
-; SSSE3-LABEL: 'cmp_float_oge'
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp oge float undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F32 = fcmp oge <2 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F32 = fcmp oge <4 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp oge <8 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = fcmp oge <16 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp oge double undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F64 = fcmp oge <2 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp oge <4 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8F64 = fcmp oge <8 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp oge <16 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
 ; SSE41-LABEL: 'cmp_float_oge'
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp oge float undef, undef
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp oge <2 x float> undef, undef
@@ -387,31 +283,31 @@ define i32 @cmp_float_oge(i32 %arg) {
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp oge <16 x double> undef, undef
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
-; SSE42-LABEL: 'cmp_float_oge'
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp oge float undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp oge <2 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp oge <4 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp oge <8 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp oge <16 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp oge double undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp oge <2 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp oge <4 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp oge <8 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp oge <16 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
-; AVX-LABEL: 'cmp_float_oge'
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp oge float undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp oge <2 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp oge <4 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp oge <8 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F32 = fcmp oge <16 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp oge double undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp oge <2 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp oge <4 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F64 = fcmp oge <8 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F64 = fcmp oge <16 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+; AVX1-LABEL: 'cmp_float_oge'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp oge float undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp oge <2 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp oge <4 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp oge <8 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp oge <16 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp oge double undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp oge <2 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp oge <4 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp oge <8 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp oge <16 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+;
+; AVX2-LABEL: 'cmp_float_oge'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp oge float undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp oge <2 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp oge <4 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp oge <8 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F32 = fcmp oge <16 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp oge double undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp oge <2 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp oge <4 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F64 = fcmp oge <8 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F64 = fcmp oge <16 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; AVX512-LABEL: 'cmp_float_oge'
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp oge float undef, undef
@@ -455,32 +351,6 @@ define i32 @cmp_float_ogt(i32 %arg) {
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp ogt <16 x double> undef, undef
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
-; SSE3-LABEL: 'cmp_float_ogt'
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ogt float undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F32 = fcmp ogt <2 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F32 = fcmp ogt <4 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp ogt <8 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = fcmp ogt <16 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ogt double undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F64 = fcmp ogt <2 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp ogt <4 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8F64 = fcmp ogt <8 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp ogt <16 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
-; SSSE3-LABEL: 'cmp_float_ogt'
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ogt float undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F32 = fcmp ogt <2 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F32 = fcmp ogt <4 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp ogt <8 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = fcmp ogt <16 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ogt double undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F64 = fcmp ogt <2 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp ogt <4 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8F64 = fcmp ogt <8 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp ogt <16 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
 ; SSE41-LABEL: 'cmp_float_ogt'
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ogt float undef, undef
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ogt <2 x float> undef, undef
@@ -494,31 +364,31 @@ define i32 @cmp_float_ogt(i32 %arg) {
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp ogt <16 x double> undef, undef
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
-; SSE42-LABEL: 'cmp_float_ogt'
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ogt float undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ogt <2 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ogt <4 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp ogt <8 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp ogt <16 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ogt double undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ogt <2 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp ogt <4 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp ogt <8 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp ogt <16 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
-; AVX-LABEL: 'cmp_float_ogt'
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ogt float undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ogt <2 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ogt <4 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ogt <8 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F32 = fcmp ogt <16 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ogt double undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ogt <2 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ogt <4 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F64 = fcmp ogt <8 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F64 = fcmp ogt <16 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+; AVX1-LABEL: 'cmp_float_ogt'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ogt float undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ogt <2 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ogt <4 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp ogt <8 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp ogt <16 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ogt double undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ogt <2 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp ogt <4 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp ogt <8 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp ogt <16 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+;
+; AVX2-LABEL: 'cmp_float_ogt'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ogt float undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ogt <2 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ogt <4 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ogt <8 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F32 = fcmp ogt <16 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ogt double undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ogt <2 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ogt <4 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F64 = fcmp ogt <8 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F64 = fcmp ogt <16 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; AVX512-LABEL: 'cmp_float_ogt'
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ogt float undef, undef
@@ -562,32 +432,6 @@ define i32 @cmp_float_ole(i32 %arg) {
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp ole <16 x double> undef, undef
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
-; SSE3-LABEL: 'cmp_float_ole'
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ole float undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F32 = fcmp ole <2 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F32 = fcmp ole <4 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp ole <8 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = fcmp ole <16 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ole double undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F64 = fcmp ole <2 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp ole <4 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8F64 = fcmp ole <8 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp ole <16 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
-; SSSE3-LABEL: 'cmp_float_ole'
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ole float undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F32 = fcmp ole <2 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F32 = fcmp ole <4 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp ole <8 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = fcmp ole <16 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ole double undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F64 = fcmp ole <2 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp ole <4 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8F64 = fcmp ole <8 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp ole <16 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
 ; SSE41-LABEL: 'cmp_float_ole'
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ole float undef, undef
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ole <2 x float> undef, undef
@@ -601,31 +445,31 @@ define i32 @cmp_float_ole(i32 %arg) {
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp ole <16 x double> undef, undef
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
-; SSE42-LABEL: 'cmp_float_ole'
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ole float undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ole <2 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ole <4 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp ole <8 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp ole <16 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ole double undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ole <2 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp ole <4 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp ole <8 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp ole <16 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
-; AVX-LABEL: 'cmp_float_ole'
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ole float undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ole <2 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ole <4 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ole <8 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F32 = fcmp ole <16 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ole double undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ole <2 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ole <4 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F64 = fcmp ole <8 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F64 = fcmp ole <16 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+; AVX1-LABEL: 'cmp_float_ole'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ole float undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ole <2 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ole <4 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp ole <8 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp ole <16 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ole double undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ole <2 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp ole <4 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp ole <8 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp ole <16 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+;
+; AVX2-LABEL: 'cmp_float_ole'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ole float undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ole <2 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ole <4 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ole <8 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F32 = fcmp ole <16 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ole double undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ole <2 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ole <4 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F64 = fcmp ole <8 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F64 = fcmp ole <16 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; AVX512-LABEL: 'cmp_float_ole'
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ole float undef, undef
@@ -669,32 +513,6 @@ define i32 @cmp_float_olt(i32 %arg) {
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp olt <16 x double> undef, undef
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
-; SSE3-LABEL: 'cmp_float_olt'
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp olt float undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F32 = fcmp olt <2 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F32 = fcmp olt <4 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp olt <8 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = fcmp olt <16 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp olt double undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F64 = fcmp olt <2 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp olt <4 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8F64 = fcmp olt <8 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp olt <16 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
-; SSSE3-LABEL: 'cmp_float_olt'
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp olt float undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F32 = fcmp olt <2 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F32 = fcmp olt <4 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp olt <8 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = fcmp olt <16 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp olt double undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F64 = fcmp olt <2 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp olt <4 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8F64 = fcmp olt <8 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp olt <16 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
 ; SSE41-LABEL: 'cmp_float_olt'
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp olt float undef, undef
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp olt <2 x float> undef, undef
@@ -708,31 +526,31 @@ define i32 @cmp_float_olt(i32 %arg) {
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp olt <16 x double> undef, undef
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
-; SSE42-LABEL: 'cmp_float_olt'
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp olt float undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp olt <2 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp olt <4 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp olt <8 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp olt <16 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp olt double undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp olt <2 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp olt <4 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp olt <8 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp olt <16 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
-; AVX-LABEL: 'cmp_float_olt'
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp olt float undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp olt <2 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp olt <4 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp olt <8 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F32 = fcmp olt <16 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp olt double undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp olt <2 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp olt <4 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F64 = fcmp olt <8 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F64 = fcmp olt <16 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+; AVX1-LABEL: 'cmp_float_olt'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp olt float undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp olt <2 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp olt <4 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp olt <8 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp olt <16 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp olt double undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp olt <2 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp olt <4 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp olt <8 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp olt <16 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+;
+; AVX2-LABEL: 'cmp_float_olt'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp olt float undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp olt <2 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp olt <4 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp olt <8 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F32 = fcmp olt <16 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp olt double undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp olt <2 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp olt <4 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F64 = fcmp olt <8 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F64 = fcmp olt <16 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; AVX512-LABEL: 'cmp_float_olt'
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp olt float undef, undef
@@ -776,32 +594,6 @@ define i32 @cmp_float_ueq(i32 %arg) {
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp ueq <16 x double> undef, undef
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
-; SSE3-LABEL: 'cmp_float_ueq'
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ueq float undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F32 = fcmp ueq <2 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F32 = fcmp ueq <4 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp ueq <8 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = fcmp ueq <16 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ueq double undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F64 = fcmp ueq <2 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp ueq <4 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8F64 = fcmp ueq <8 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp ueq <16 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
-; SSSE3-LABEL: 'cmp_float_ueq'
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ueq float undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F32 = fcmp ueq <2 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F32 = fcmp ueq <4 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp ueq <8 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = fcmp ueq <16 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ueq double undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F64 = fcmp ueq <2 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp ueq <4 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8F64 = fcmp ueq <8 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp ueq <16 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
 ; SSE41-LABEL: 'cmp_float_ueq'
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ueq float undef, undef
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ueq <2 x float> undef, undef
@@ -815,31 +607,31 @@ define i32 @cmp_float_ueq(i32 %arg) {
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp ueq <16 x double> undef, undef
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
-; SSE42-LABEL: 'cmp_float_ueq'
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ueq float undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ueq <2 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ueq <4 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp ueq <8 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp ueq <16 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ueq double undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ueq <2 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp ueq <4 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp ueq <8 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp ueq <16 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
-; AVX-LABEL: 'cmp_float_ueq'
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ueq float undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ueq <2 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ueq <4 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ueq <8 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F32 = fcmp ueq <16 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ueq double undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ueq <2 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ueq <4 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F64 = fcmp ueq <8 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F64 = fcmp ueq <16 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+; AVX1-LABEL: 'cmp_float_ueq'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ueq float undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ueq <2 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ueq <4 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp ueq <8 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp ueq <16 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ueq double undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ueq <2 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp ueq <4 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp ueq <8 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp ueq <16 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+;
+; AVX2-LABEL: 'cmp_float_ueq'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ueq float undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ueq <2 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ueq <4 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ueq <8 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F32 = fcmp ueq <16 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ueq double undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ueq <2 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ueq <4 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F64 = fcmp ueq <8 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F64 = fcmp ueq <16 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; AVX512-LABEL: 'cmp_float_ueq'
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ueq float undef, undef
@@ -883,32 +675,6 @@ define i32 @cmp_float_une(i32 %arg) {
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp une <16 x double> undef, undef
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
-; SSE3-LABEL: 'cmp_float_une'
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp une float undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F32 = fcmp une <2 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F32 = fcmp une <4 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp une <8 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = fcmp une <16 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp une double undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F64 = fcmp une <2 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp une <4 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8F64 = fcmp une <8 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp une <16 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
-; SSSE3-LABEL: 'cmp_float_une'
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp une float undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F32 = fcmp une <2 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F32 = fcmp une <4 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp une <8 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = fcmp une <16 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp une double undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F64 = fcmp une <2 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp une <4 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8F64 = fcmp une <8 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp une <16 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
 ; SSE41-LABEL: 'cmp_float_une'
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp une float undef, undef
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp une <2 x float> undef, undef
@@ -922,31 +688,31 @@ define i32 @cmp_float_une(i32 %arg) {
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp une <16 x double> undef, undef
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
-; SSE42-LABEL: 'cmp_float_une'
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp une float undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp une <2 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp une <4 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp une <8 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp une <16 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp une double undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp une <2 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp une <4 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp une <8 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp une <16 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
-; AVX-LABEL: 'cmp_float_une'
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp une float undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp une <2 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp une <4 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp une <8 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F32 = fcmp une <16 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp une double undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp une <2 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp une <4 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F64 = fcmp une <8 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F64 = fcmp une <16 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+; AVX1-LABEL: 'cmp_float_une'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp une float undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp une <2 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp une <4 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp une <8 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp une <16 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp une double undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp une <2 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp une <4 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp une <8 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp une <16 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+;
+; AVX2-LABEL: 'cmp_float_une'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp une float undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp une <2 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp une <4 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp une <8 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F32 = fcmp une <16 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp une double undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp une <2 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp une <4 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F64 = fcmp une <8 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F64 = fcmp une <16 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; AVX512-LABEL: 'cmp_float_une'
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp une float undef, undef
@@ -990,32 +756,6 @@ define i32 @cmp_float_uno(i32 %arg) {
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp uno <16 x double> undef, undef
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
-; SSE3-LABEL: 'cmp_float_uno'
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp uno float undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F32 = fcmp uno <2 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F32 = fcmp uno <4 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp uno <8 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = fcmp uno <16 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp uno double undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F64 = fcmp uno <2 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp uno <4 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8F64 = fcmp uno <8 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp uno <16 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
-; SSSE3-LABEL: 'cmp_float_uno'
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp uno float undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F32 = fcmp uno <2 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F32 = fcmp uno <4 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp uno <8 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = fcmp uno <16 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp uno double undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F64 = fcmp uno <2 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp uno <4 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8F64 = fcmp uno <8 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp uno <16 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
 ; SSE41-LABEL: 'cmp_float_uno'
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp uno float undef, undef
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp uno <2 x float> undef, undef
@@ -1029,31 +769,31 @@ define i32 @cmp_float_uno(i32 %arg) {
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp uno <16 x double> undef, undef
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
-; SSE42-LABEL: 'cmp_float_uno'
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp uno float undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp uno <2 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp uno <4 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp uno <8 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp uno <16 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp uno double undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp uno <2 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp uno <4 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp uno <8 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp uno <16 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
-; AVX-LABEL: 'cmp_float_uno'
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp uno float undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp uno <2 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp uno <4 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp uno <8 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F32 = fcmp uno <16 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp uno double undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp uno <2 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp uno <4 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F64 = fcmp uno <8 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F64 = fcmp uno <16 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+; AVX1-LABEL: 'cmp_float_uno'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp uno float undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp uno <2 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp uno <4 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp uno <8 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp uno <16 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp uno double undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp uno <2 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp uno <4 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp uno <8 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp uno <16 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+;
+; AVX2-LABEL: 'cmp_float_uno'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp uno float undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp uno <2 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp uno <4 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp uno <8 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F32 = fcmp uno <16 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp uno double undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp uno <2 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp uno <4 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F64 = fcmp uno <8 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F64 = fcmp uno <16 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; AVX512-LABEL: 'cmp_float_uno'
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp uno float undef, undef
@@ -1097,32 +837,6 @@ define i32 @cmp_float_uge(i32 %arg) {
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp uge <16 x double> undef, undef
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
-; SSE3-LABEL: 'cmp_float_uge'
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp uge float undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F32 = fcmp uge <2 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F32 = fcmp uge <4 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp uge <8 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = fcmp uge <16 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp uge double undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F64 = fcmp uge <2 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp uge <4 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8F64 = fcmp uge <8 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp uge <16 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
-; SSSE3-LABEL: 'cmp_float_uge'
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp uge float undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F32 = fcmp uge <2 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F32 = fcmp uge <4 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp uge <8 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = fcmp uge <16 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp uge double undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F64 = fcmp uge <2 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp uge <4 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8F64 = fcmp uge <8 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp uge <16 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
 ; SSE41-LABEL: 'cmp_float_uge'
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp uge float undef, undef
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp uge <2 x float> undef, undef
@@ -1136,31 +850,31 @@ define i32 @cmp_float_uge(i32 %arg) {
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp uge <16 x double> undef, undef
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
-; SSE42-LABEL: 'cmp_float_uge'
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp uge float undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp uge <2 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp uge <4 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp uge <8 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp uge <16 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp uge double undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp uge <2 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp uge <4 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp uge <8 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp uge <16 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
-; AVX-LABEL: 'cmp_float_uge'
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp uge float undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp uge <2 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp uge <4 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp uge <8 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F32 = fcmp uge <16 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp uge double undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp uge <2 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp uge <4 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F64 = fcmp uge <8 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F64 = fcmp uge <16 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+; AVX1-LABEL: 'cmp_float_uge'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp uge float undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp uge <2 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp uge <4 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp uge <8 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp uge <16 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp uge double undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp uge <2 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp uge <4 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp uge <8 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp uge <16 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+;
+; AVX2-LABEL: 'cmp_float_uge'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp uge float undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp uge <2 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp uge <4 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp uge <8 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F32 = fcmp uge <16 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp uge double undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp uge <2 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp uge <4 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F64 = fcmp uge <8 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F64 = fcmp uge <16 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; AVX512-LABEL: 'cmp_float_uge'
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp uge float undef, undef
@@ -1204,32 +918,6 @@ define i32 @cmp_float_ugt(i32 %arg) {
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp ugt <16 x double> undef, undef
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
-; SSE3-LABEL: 'cmp_float_ugt'
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ugt float undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F32 = fcmp ugt <2 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F32 = fcmp ugt <4 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp ugt <8 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = fcmp ugt <16 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ugt double undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F64 = fcmp ugt <2 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp ugt <4 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8F64 = fcmp ugt <8 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp ugt <16 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
-; SSSE3-LABEL: 'cmp_float_ugt'
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ugt float undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F32 = fcmp ugt <2 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F32 = fcmp ugt <4 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp ugt <8 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = fcmp ugt <16 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ugt double undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F64 = fcmp ugt <2 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp ugt <4 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8F64 = fcmp ugt <8 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp ugt <16 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
 ; SSE41-LABEL: 'cmp_float_ugt'
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ugt float undef, undef
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ugt <2 x float> undef, undef
@@ -1243,31 +931,31 @@ define i32 @cmp_float_ugt(i32 %arg) {
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp ugt <16 x double> undef, undef
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
-; SSE42-LABEL: 'cmp_float_ugt'
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ugt float undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ugt <2 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ugt <4 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp ugt <8 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp ugt <16 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ugt double undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ugt <2 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp ugt <4 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp ugt <8 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp ugt <16 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
-; AVX-LABEL: 'cmp_float_ugt'
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ugt float undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ugt <2 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ugt <4 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ugt <8 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F32 = fcmp ugt <16 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ugt double undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ugt <2 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ugt <4 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F64 = fcmp ugt <8 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F64 = fcmp ugt <16 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+; AVX1-LABEL: 'cmp_float_ugt'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ugt float undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ugt <2 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ugt <4 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp ugt <8 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp ugt <16 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ugt double undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ugt <2 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp ugt <4 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp ugt <8 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp ugt <16 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+;
+; AVX2-LABEL: 'cmp_float_ugt'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ugt float undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ugt <2 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ugt <4 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ugt <8 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F32 = fcmp ugt <16 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ugt double undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ugt <2 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ugt <4 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F64 = fcmp ugt <8 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F64 = fcmp ugt <16 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; AVX512-LABEL: 'cmp_float_ugt'
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ugt float undef, undef
@@ -1311,32 +999,6 @@ define i32 @cmp_float_ule(i32 %arg) {
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp ule <16 x double> undef, undef
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
-; SSE3-LABEL: 'cmp_float_ule'
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ule float undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F32 = fcmp ule <2 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F32 = fcmp ule <4 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp ule <8 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = fcmp ule <16 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ule double undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F64 = fcmp ule <2 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp ule <4 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8F64 = fcmp ule <8 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp ule <16 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
-; SSSE3-LABEL: 'cmp_float_ule'
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ule float undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F32 = fcmp ule <2 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F32 = fcmp ule <4 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp ule <8 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = fcmp ule <16 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ule double undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F64 = fcmp ule <2 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp ule <4 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8F64 = fcmp ule <8 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp ule <16 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
 ; SSE41-LABEL: 'cmp_float_ule'
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ule float undef, undef
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ule <2 x float> undef, undef
@@ -1350,31 +1012,31 @@ define i32 @cmp_float_ule(i32 %arg) {
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp ule <16 x double> undef, undef
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
-; SSE42-LABEL: 'cmp_float_ule'
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ule float undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ule <2 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ule <4 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp ule <8 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp ule <16 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ule double undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ule <2 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp ule <4 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp ule <8 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp ule <16 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
-; AVX-LABEL: 'cmp_float_ule'
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ule float undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ule <2 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ule <4 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ule <8 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F32 = fcmp ule <16 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ule double undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ule <2 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ule <4 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F64 = fcmp ule <8 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F64 = fcmp ule <16 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+; AVX1-LABEL: 'cmp_float_ule'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ule float undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ule <2 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ule <4 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp ule <8 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp ule <16 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ule double undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ule <2 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp ule <4 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp ule <8 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp ule <16 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+;
+; AVX2-LABEL: 'cmp_float_ule'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ule float undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ule <2 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ule <4 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ule <8 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F32 = fcmp ule <16 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ule double undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ule <2 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ule <4 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F64 = fcmp ule <8 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F64 = fcmp ule <16 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; AVX512-LABEL: 'cmp_float_ule'
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ule float undef, undef
@@ -1418,32 +1080,6 @@ define i32 @cmp_float_ult(i32 %arg) {
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp ult <16 x double> undef, undef
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
-; SSE3-LABEL: 'cmp_float_ult'
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ult float undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F32 = fcmp ult <2 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F32 = fcmp ult <4 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp ult <8 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = fcmp ult <16 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ult double undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F64 = fcmp ult <2 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp ult <4 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8F64 = fcmp ult <8 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp ult <16 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
-; SSSE3-LABEL: 'cmp_float_ult'
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ult float undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F32 = fcmp ult <2 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F32 = fcmp ult <4 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp ult <8 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = fcmp ult <16 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ult double undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F64 = fcmp ult <2 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp ult <4 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8F64 = fcmp ult <8 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp ult <16 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
 ; SSE41-LABEL: 'cmp_float_ult'
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ult float undef, undef
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ult <2 x float> undef, undef
@@ -1457,31 +1093,31 @@ define i32 @cmp_float_ult(i32 %arg) {
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp ult <16 x double> undef, undef
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
-; SSE42-LABEL: 'cmp_float_ult'
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ult float undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ult <2 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ult <4 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp ult <8 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp ult <16 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ult double undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ult <2 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp ult <4 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp ult <8 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp ult <16 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
-; AVX-LABEL: 'cmp_float_ult'
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ult float undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ult <2 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ult <4 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ult <8 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F32 = fcmp ult <16 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ult double undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ult <2 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ult <4 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F64 = fcmp ult <8 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F64 = fcmp ult <16 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+; AVX1-LABEL: 'cmp_float_ult'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ult float undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ult <2 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ult <4 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp ult <8 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp ult <16 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ult double undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ult <2 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp ult <4 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp ult <8 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp ult <16 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+;
+; AVX2-LABEL: 'cmp_float_ult'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ult float undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp ult <2 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp ult <4 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp ult <8 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F32 = fcmp ult <16 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp ult double undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp ult <2 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp ult <4 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F64 = fcmp ult <8 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F64 = fcmp ult <16 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; AVX512-LABEL: 'cmp_float_ult'
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp ult float undef, undef
@@ -1525,32 +1161,6 @@ define i32 @cmp_float_false(i32 %arg) {
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp false <16 x double> undef, undef
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
-; SSE3-LABEL: 'cmp_float_false'
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp false float undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F32 = fcmp false <2 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F32 = fcmp false <4 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp false <8 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = fcmp false <16 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp false double undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F64 = fcmp false <2 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp false <4 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8F64 = fcmp false <8 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp false <16 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
-; SSSE3-LABEL: 'cmp_float_false'
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp false float undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F32 = fcmp false <2 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F32 = fcmp false <4 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp false <8 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = fcmp false <16 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp false double undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F64 = fcmp false <2 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp false <4 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8F64 = fcmp false <8 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp false <16 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
 ; SSE41-LABEL: 'cmp_float_false'
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp false float undef, undef
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp false <2 x float> undef, undef
@@ -1564,31 +1174,31 @@ define i32 @cmp_float_false(i32 %arg) {
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp false <16 x double> undef, undef
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
-; SSE42-LABEL: 'cmp_float_false'
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp false float undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp false <2 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp false <4 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp false <8 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp false <16 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp false double undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp false <2 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp false <4 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp false <8 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp false <16 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
-; AVX-LABEL: 'cmp_float_false'
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp false float undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp false <2 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp false <4 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp false <8 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F32 = fcmp false <16 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp false double undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp false <2 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp false <4 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F64 = fcmp false <8 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F64 = fcmp false <16 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+; AVX1-LABEL: 'cmp_float_false'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp false float undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp false <2 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp false <4 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp false <8 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp false <16 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp false double undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp false <2 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp false <4 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp false <8 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp false <16 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+;
+; AVX2-LABEL: 'cmp_float_false'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp false float undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp false <2 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp false <4 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp false <8 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F32 = fcmp false <16 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp false double undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp false <2 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp false <4 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F64 = fcmp false <8 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F64 = fcmp false <16 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; AVX512-LABEL: 'cmp_float_false'
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp false float undef, undef
@@ -1632,32 +1242,6 @@ define i32 @cmp_float_true(i32 %arg) {
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp true <16 x double> undef, undef
 ; SSE2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
-; SSE3-LABEL: 'cmp_float_true'
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp true float undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F32 = fcmp true <2 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F32 = fcmp true <4 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp true <8 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = fcmp true <16 x float> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp true double undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F64 = fcmp true <2 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp true <4 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8F64 = fcmp true <8 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp true <16 x double> undef, undef
-; SSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
-; SSSE3-LABEL: 'cmp_float_true'
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp true float undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F32 = fcmp true <2 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F32 = fcmp true <4 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F32 = fcmp true <8 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F32 = fcmp true <16 x float> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp true double undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V2F64 = fcmp true <2 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4F64 = fcmp true <4 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V8F64 = fcmp true <8 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V16F64 = fcmp true <16 x double> undef, undef
-; SSSE3-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
 ; SSE41-LABEL: 'cmp_float_true'
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp true float undef, undef
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp true <2 x float> undef, undef
@@ -1671,31 +1255,31 @@ define i32 @cmp_float_true(i32 %arg) {
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp true <16 x double> undef, undef
 ; SSE41-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
-; SSE42-LABEL: 'cmp_float_true'
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp true float undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp true <2 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp true <4 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp true <8 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp true <16 x float> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp true double undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp true <2 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp true <4 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp true <8 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp true <16 x double> undef, undef
-; SSE42-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
-;
-; AVX-LABEL: 'cmp_float_true'
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp true float undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp true <2 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp true <4 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp true <8 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F32 = fcmp true <16 x float> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp true double undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp true <2 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp true <4 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F64 = fcmp true <8 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F64 = fcmp true <16 x double> undef, undef
-; AVX-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+; AVX1-LABEL: 'cmp_float_true'
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp true float undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp true <2 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp true <4 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = fcmp true <8 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = fcmp true <16 x float> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp true double undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp true <2 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = fcmp true <4 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = fcmp true <8 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V16F64 = fcmp true <16 x double> undef, undef
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+;
+; AVX2-LABEL: 'cmp_float_true'
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp true float undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F32 = fcmp true <2 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = fcmp true <4 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = fcmp true <8 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V16F32 = fcmp true <16 x float> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F64 = fcmp true double undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = fcmp true <2 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = fcmp true <4 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %V8F64 = fcmp true <8 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V16F64 = fcmp true <16 x double> undef, undef
+; AVX2-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
 ;
 ; AVX512-LABEL: 'cmp_float_true'
 ; AVX512-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %F32 = fcmp true float undef, undef

diff  --git a/llvm/test/Analysis/CostModel/X86/fptoi_sat.ll b/llvm/test/Analysis/CostModel/X86/fptoi_sat.ll
index 71b9b82652985..c89e32df1aa9d 100644
--- a/llvm/test/Analysis/CostModel/X86/fptoi_sat.ll
+++ b/llvm/test/Analysis/CostModel/X86/fptoi_sat.ll
@@ -268,55 +268,55 @@ define void @casts() {
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 9 for instruction: %v4f32u32 = call <4 x i32> @llvm.fptoui.sat.v4i32.v4f32(<4 x float> undef)
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 25 for instruction: %v4f32s64 = call <4 x i64> @llvm.fptosi.sat.v4i64.v4f32(<4 x float> undef)
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 33 for instruction: %v4f32u64 = call <4 x i64> @llvm.fptoui.sat.v4i64.v4f32(<4 x float> undef)
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %v4f64s1 = call <4 x i1> @llvm.fptosi.sat.v4i1.v4f64(<4 x double> undef)
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %v4f64s1 = call <4 x i1> @llvm.fptosi.sat.v4i1.v4f64(<4 x double> undef)
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %v4f64u1 = call <4 x i1> @llvm.fptoui.sat.v4i1.v4f64(<4 x double> undef)
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %v4f64s8 = call <4 x i8> @llvm.fptosi.sat.v4i8.v4f64(<4 x double> undef)
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %v4f64s8 = call <4 x i8> @llvm.fptosi.sat.v4i8.v4f64(<4 x double> undef)
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %v4f64u8 = call <4 x i8> @llvm.fptoui.sat.v4i8.v4f64(<4 x double> undef)
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %v4f64s16 = call <4 x i16> @llvm.fptosi.sat.v4i16.v4f64(<4 x double> undef)
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %v4f64s16 = call <4 x i16> @llvm.fptosi.sat.v4i16.v4f64(<4 x double> undef)
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %v4f64u16 = call <4 x i16> @llvm.fptoui.sat.v4i16.v4f64(<4 x double> undef)
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %v4f64s32 = call <4 x i32> @llvm.fptosi.sat.v4i32.v4f64(<4 x double> undef)
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %v4f64s32 = call <4 x i32> @llvm.fptosi.sat.v4i32.v4f64(<4 x double> undef)
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %v4f64u32 = call <4 x i32> @llvm.fptoui.sat.v4i32.v4f64(<4 x double> undef)
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 29 for instruction: %v4f64s64 = call <4 x i64> @llvm.fptosi.sat.v4i64.v4f64(<4 x double> undef)
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 30 for instruction: %v4f64s64 = call <4 x i64> @llvm.fptosi.sat.v4i64.v4f64(<4 x double> undef)
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 37 for instruction: %v4f64u64 = call <4 x i64> @llvm.fptoui.sat.v4i64.v4f64(<4 x double> undef)
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %v8f32s1 = call <8 x i1> @llvm.fptosi.sat.v8i1.v8f32(<8 x float> undef)
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %v8f32s1 = call <8 x i1> @llvm.fptosi.sat.v8i1.v8f32(<8 x float> undef)
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %v8f32u1 = call <8 x i1> @llvm.fptoui.sat.v8i1.v8f32(<8 x float> undef)
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %v8f32s8 = call <8 x i8> @llvm.fptosi.sat.v8i8.v8f32(<8 x float> undef)
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %v8f32s8 = call <8 x i8> @llvm.fptosi.sat.v8i8.v8f32(<8 x float> undef)
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %v8f32u8 = call <8 x i8> @llvm.fptoui.sat.v8i8.v8f32(<8 x float> undef)
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 15 for instruction: %v8f32s16 = call <8 x i16> @llvm.fptosi.sat.v8i16.v8f32(<8 x float> undef)
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %v8f32s16 = call <8 x i16> @llvm.fptosi.sat.v8i16.v8f32(<8 x float> undef)
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %v8f32u16 = call <8 x i16> @llvm.fptoui.sat.v8i16.v8f32(<8 x float> undef)
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %v8f32s32 = call <8 x i32> @llvm.fptosi.sat.v8i32.v8f32(<8 x float> undef)
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 17 for instruction: %v8f32s32 = call <8 x i32> @llvm.fptosi.sat.v8i32.v8f32(<8 x float> undef)
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 17 for instruction: %v8f32u32 = call <8 x i32> @llvm.fptoui.sat.v8i32.v8f32(<8 x float> undef)
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 48 for instruction: %v8f32s64 = call <8 x i64> @llvm.fptosi.sat.v8i64.v8f32(<8 x float> undef)
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 49 for instruction: %v8f32s64 = call <8 x i64> @llvm.fptosi.sat.v8i64.v8f32(<8 x float> undef)
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 65 for instruction: %v8f32u64 = call <8 x i64> @llvm.fptoui.sat.v8i64.v8f32(<8 x float> undef)
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %v8f64s1 = call <8 x i1> @llvm.fptosi.sat.v8i1.v8f64(<8 x double> undef)
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 30 for instruction: %v8f64s1 = call <8 x i1> @llvm.fptosi.sat.v8i1.v8f64(<8 x double> undef)
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %v8f64u1 = call <8 x i1> @llvm.fptoui.sat.v8i1.v8f64(<8 x double> undef)
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %v8f64s8 = call <8 x i8> @llvm.fptosi.sat.v8i8.v8f64(<8 x double> undef)
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 30 for instruction: %v8f64s8 = call <8 x i8> @llvm.fptosi.sat.v8i8.v8f64(<8 x double> undef)
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %v8f64u8 = call <8 x i8> @llvm.fptoui.sat.v8i8.v8f64(<8 x double> undef)
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %v8f64s16 = call <8 x i16> @llvm.fptosi.sat.v8i16.v8f64(<8 x double> undef)
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 30 for instruction: %v8f64s16 = call <8 x i16> @llvm.fptosi.sat.v8i16.v8f64(<8 x double> undef)
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %v8f64u16 = call <8 x i16> @llvm.fptoui.sat.v8i16.v8f64(<8 x double> undef)
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 30 for instruction: %v8f64s32 = call <8 x i32> @llvm.fptosi.sat.v8i32.v8f64(<8 x double> undef)
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 32 for instruction: %v8f64s32 = call <8 x i32> @llvm.fptosi.sat.v8i32.v8f64(<8 x double> undef)
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 34 for instruction: %v8f64u32 = call <8 x i32> @llvm.fptoui.sat.v8i32.v8f64(<8 x double> undef)
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 58 for instruction: %v8f64s64 = call <8 x i64> @llvm.fptosi.sat.v8i64.v8f64(<8 x double> undef)
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 60 for instruction: %v8f64s64 = call <8 x i64> @llvm.fptosi.sat.v8i64.v8f64(<8 x double> undef)
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 74 for instruction: %v8f64u64 = call <8 x i64> @llvm.fptoui.sat.v8i64.v8f64(<8 x double> undef)
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %v16f32s1 = call <16 x i1> @llvm.fptosi.sat.v16i1.v16f32(<16 x float> undef)
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 30 for instruction: %v16f32s1 = call <16 x i1> @llvm.fptosi.sat.v16i1.v16f32(<16 x float> undef)
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %v16f32u1 = call <16 x i1> @llvm.fptoui.sat.v16i1.v16f32(<16 x float> undef)
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 28 for instruction: %v16f32s8 = call <16 x i8> @llvm.fptosi.sat.v16i8.v16f32(<16 x float> undef)
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 30 for instruction: %v16f32s8 = call <16 x i8> @llvm.fptosi.sat.v16i8.v16f32(<16 x float> undef)
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %v16f32u8 = call <16 x i8> @llvm.fptoui.sat.v16i8.v16f32(<16 x float> undef)
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 29 for instruction: %v16f32s16 = call <16 x i16> @llvm.fptosi.sat.v16i16.v16f32(<16 x float> undef)
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 31 for instruction: %v16f32s16 = call <16 x i16> @llvm.fptosi.sat.v16i16.v16f32(<16 x float> undef)
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %v16f32u16 = call <16 x i16> @llvm.fptoui.sat.v16i16.v16f32(<16 x float> undef)
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 32 for instruction: %v16f32s32 = call <16 x i32> @llvm.fptosi.sat.v16i32.v16f32(<16 x float> undef)
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 34 for instruction: %v16f32s32 = call <16 x i32> @llvm.fptosi.sat.v16i32.v16f32(<16 x float> undef)
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 34 for instruction: %v16f32u32 = call <16 x i32> @llvm.fptoui.sat.v16i32.v16f32(<16 x float> undef)
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 96 for instruction: %v16f32s64 = call <16 x i64> @llvm.fptosi.sat.v16i64.v16f32(<16 x float> undef)
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 98 for instruction: %v16f32s64 = call <16 x i64> @llvm.fptosi.sat.v16i64.v16f32(<16 x float> undef)
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 130 for instruction: %v16f32u64 = call <16 x i64> @llvm.fptoui.sat.v16i64.v16f32(<16 x float> undef)
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 54 for instruction: %v16f64s1 = call <16 x i1> @llvm.fptosi.sat.v16i1.v16f64(<16 x double> undef)
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 58 for instruction: %v16f64s1 = call <16 x i1> @llvm.fptosi.sat.v16i1.v16f64(<16 x double> undef)
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 48 for instruction: %v16f64u1 = call <16 x i1> @llvm.fptoui.sat.v16i1.v16f64(<16 x double> undef)
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 54 for instruction: %v16f64s8 = call <16 x i8> @llvm.fptosi.sat.v16i8.v16f64(<16 x double> undef)
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 58 for instruction: %v16f64s8 = call <16 x i8> @llvm.fptosi.sat.v16i8.v16f64(<16 x double> undef)
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 48 for instruction: %v16f64u8 = call <16 x i8> @llvm.fptoui.sat.v16i8.v16f64(<16 x double> undef)
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 55 for instruction: %v16f64s16 = call <16 x i16> @llvm.fptosi.sat.v16i16.v16f64(<16 x double> undef)
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 59 for instruction: %v16f64s16 = call <16 x i16> @llvm.fptosi.sat.v16i16.v16f64(<16 x double> undef)
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 48 for instruction: %v16f64u16 = call <16 x i16> @llvm.fptoui.sat.v16i16.v16f64(<16 x double> undef)
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 60 for instruction: %v16f64s32 = call <16 x i32> @llvm.fptosi.sat.v16i32.v16f64(<16 x double> undef)
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 64 for instruction: %v16f64s32 = call <16 x i32> @llvm.fptosi.sat.v16i32.v16f64(<16 x double> undef)
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 68 for instruction: %v16f64u32 = call <16 x i32> @llvm.fptoui.sat.v16i32.v16f64(<16 x double> undef)
-; AVX1-NEXT:  Cost Model: Found an estimated cost of 116 for instruction: %v16f64s64 = call <16 x i64> @llvm.fptosi.sat.v16i64.v16f64(<16 x double> undef)
+; AVX1-NEXT:  Cost Model: Found an estimated cost of 120 for instruction: %v16f64s64 = call <16 x i64> @llvm.fptosi.sat.v16i64.v16f64(<16 x double> undef)
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 148 for instruction: %v16f64u64 = call <16 x i64> @llvm.fptoui.sat.v16i64.v16f64(<16 x double> undef)
 ; AVX1-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret void
 ;


        


More information about the llvm-commits mailing list