[llvm] c296474 - [CostModel][X86] Reduce cost of vector selects on SSE2/AVX1 targets
Simon Pilgrim via llvm-commits
llvm-commits at lists.llvm.org
Sun May 1 01:32:35 PDT 2022
Author: Simon Pilgrim
Date: 2022-05-01T09:32:14+01:00
New Revision: c2964746e3391360f1cbbfbfdb82e4c1e661c436
URL: https://github.com/llvm/llvm-project/commit/c2964746e3391360f1cbbfbfdb82e4c1e661c436
DIFF: https://github.com/llvm/llvm-project/commit/c2964746e3391360f1cbbfbfdb82e4c1e661c436.diff
LOG: [CostModel][X86] Reduce cost of vector selects on SSE2/AVX1 targets
Based off the script from D103695, we were exaggerating the cost of the OR(AND(X,M),AND(Y,~M)) expansion using instruction count instead of effective throughput
Added:
Modified:
llvm/lib/Target/X86/X86TargetTransformInfo.cpp
llvm/test/Analysis/CostModel/X86/arith-sminmax.ll
llvm/test/Analysis/CostModel/X86/arith-ssat.ll
llvm/test/Analysis/CostModel/X86/arith-uminmax.ll
llvm/test/Analysis/CostModel/X86/arith-usat.ll
llvm/test/Analysis/CostModel/X86/fptoi_sat.ll
llvm/test/Analysis/CostModel/X86/fshl.ll
llvm/test/Analysis/CostModel/X86/fshr.ll
llvm/test/Analysis/CostModel/X86/intrinsic-cost-kinds.ll
llvm/test/Analysis/CostModel/X86/reduce-smax.ll
llvm/test/Analysis/CostModel/X86/reduce-smin.ll
llvm/test/Analysis/CostModel/X86/reduce-umax.ll
llvm/test/Analysis/CostModel/X86/reduce-umin.ll
llvm/test/Analysis/CostModel/X86/vselect-cost.ll
llvm/test/Transforms/LoopVectorize/X86/imprecise-through-phis.ll
llvm/test/Transforms/SLPVectorizer/X86/crash_cmpop.ll
Removed:
################################################################################
diff --git a/llvm/lib/Target/X86/X86TargetTransformInfo.cpp b/llvm/lib/Target/X86/X86TargetTransformInfo.cpp
index d6291a1bb0ce2..5bc8078907d40 100644
--- a/llvm/lib/Target/X86/X86TargetTransformInfo.cpp
+++ b/llvm/lib/Target/X86/X86TargetTransformInfo.cpp
@@ -2700,8 +2700,8 @@ InstructionCost X86TTIImpl::getCmpSelInstrCost(unsigned Opcode, Type *ValTy,
{ ISD::SETCC, MVT::v32i16, 2 }, // FIXME: should probably be 4
{ ISD::SETCC, MVT::v64i8, 2 }, // FIXME: should probably be 4
- { ISD::SELECT, MVT::v32i16, 2 }, // FIXME: should be 3
- { ISD::SELECT, MVT::v64i8, 2 }, // FIXME: should be 3
+ { ISD::SELECT, MVT::v32i16, 2 },
+ { ISD::SELECT, MVT::v64i8, 2 },
};
static const CostTblEntry AVX2CostTbl[] = {
@@ -2729,8 +2729,8 @@ InstructionCost X86TTIImpl::getCmpSelInstrCost(unsigned Opcode, Type *ValTy,
{ ISD::SELECT, MVT::v8f32, 1 }, // vblendvps
{ ISD::SELECT, MVT::v4i64, 1 }, // vblendvpd
{ ISD::SELECT, MVT::v8i32, 1 }, // vblendvps
- { ISD::SELECT, MVT::v16i16, 3 }, // vandps + vandnps + vorps
- { ISD::SELECT, MVT::v32i8, 3 }, // vandps + vandnps + vorps
+ { ISD::SELECT, MVT::v16i16, 2 }, // vandps + vandnps + vorps
+ { ISD::SELECT, MVT::v32i8, 2 }, // vandps + vandnps + vorps
};
static const CostTblEntry SSE42CostTbl[] = {
@@ -2756,18 +2756,18 @@ InstructionCost X86TTIImpl::getCmpSelInstrCost(unsigned Opcode, Type *ValTy,
{ ISD::SETCC, MVT::v8i16, 1 },
{ ISD::SETCC, MVT::v16i8, 1 },
- { ISD::SELECT, MVT::v2f64, 3 }, // andpd + andnpd + orpd
- { ISD::SELECT, MVT::v2i64, 3 }, // pand + pandn + por
- { ISD::SELECT, MVT::v4i32, 3 }, // pand + pandn + por
- { ISD::SELECT, MVT::v8i16, 3 }, // pand + pandn + por
- { ISD::SELECT, MVT::v16i8, 3 }, // pand + pandn + por
+ { ISD::SELECT, MVT::v2f64, 2 }, // andpd + andnpd + orpd
+ { ISD::SELECT, MVT::v2i64, 2 }, // pand + pandn + por
+ { ISD::SELECT, MVT::v4i32, 2 }, // pand + pandn + por
+ { ISD::SELECT, MVT::v8i16, 2 }, // pand + pandn + por
+ { ISD::SELECT, MVT::v16i8, 2 }, // pand + pandn + por
};
static const CostTblEntry SSE1CostTbl[] = {
{ ISD::SETCC, MVT::v4f32, 2 },
{ ISD::SETCC, MVT::f32, 1 },
- { ISD::SELECT, MVT::v4f32, 3 }, // andps + andnps + orps
+ { ISD::SELECT, MVT::v4f32, 2 }, // andps + andnps + orps
};
if (ST->useSLMArithCosts())
diff --git a/llvm/test/Analysis/CostModel/X86/arith-sminmax.ll b/llvm/test/Analysis/CostModel/X86/arith-sminmax.ll
index 2d74f07ba9b41..851a2762f80a2 100644
--- a/llvm/test/Analysis/CostModel/X86/arith-sminmax.ll
+++ b/llvm/test/Analysis/CostModel/X86/arith-sminmax.ll
@@ -31,40 +31,40 @@ declare <64 x i8> @llvm.smax.v64i8(<64 x i8>, <64 x i8>)
define i32 @smax(i32 %arg) {
; SSE2-LABEL: 'smax'
; SSE2-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %I64 = call i64 @llvm.smax.i64(i64 undef, i64 undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V2I64 = call <2 x i64> @llvm.smax.v2i64(<2 x i64> undef, <2 x i64> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V4I64 = call <4 x i64> @llvm.smax.v4i64(<4 x i64> undef, <4 x i64> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V8I64 = call <8 x i64> @llvm.smax.v8i64(<8 x i64> undef, <8 x i64> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V2I64 = call <2 x i64> @llvm.smax.v2i64(<2 x i64> undef, <2 x i64> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 14 for instruction: %V4I64 = call <4 x i64> @llvm.smax.v4i64(<4 x i64> undef, <4 x i64> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 28 for instruction: %V8I64 = call <8 x i64> @llvm.smax.v8i64(<8 x i64> undef, <8 x i64> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %I32 = call i32 @llvm.smax.i32(i32 undef, i32 undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V4I32 = call <4 x i32> @llvm.smax.v4i32(<4 x i32> undef, <4 x i32> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V8I32 = call <8 x i32> @llvm.smax.v8i32(<8 x i32> undef, <8 x i32> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V16I32 = call <16 x i32> @llvm.smax.v16i32(<16 x i32> undef, <16 x i32> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V4I32 = call <4 x i32> @llvm.smax.v4i32(<4 x i32> undef, <4 x i32> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V8I32 = call <8 x i32> @llvm.smax.v8i32(<8 x i32> undef, <8 x i32> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %V16I32 = call <16 x i32> @llvm.smax.v16i32(<16 x i32> undef, <16 x i32> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %I16 = call i16 @llvm.smax.i16(i16 undef, i16 undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V8I16 = call <8 x i16> @llvm.smax.v8i16(<8 x i16> undef, <8 x i16> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V16I16 = call <16 x i16> @llvm.smax.v16i16(<16 x i16> undef, <16 x i16> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V32I16 = call <32 x i16> @llvm.smax.v32i16(<32 x i16> undef, <32 x i16> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %I8 = call i8 @llvm.smax.i8(i8 undef, i8 undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V16I8 = call <16 x i8> @llvm.smax.v16i8(<16 x i8> undef, <16 x i8> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V32I8 = call <32 x i8> @llvm.smax.v32i8(<32 x i8> undef, <32 x i8> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V64I8 = call <64 x i8> @llvm.smax.v64i8(<64 x i8> undef, <64 x i8> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V16I8 = call <16 x i8> @llvm.smax.v16i8(<16 x i8> undef, <16 x i8> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V32I8 = call <32 x i8> @llvm.smax.v32i8(<32 x i8> undef, <32 x i8> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %V64I8 = call <64 x i8> @llvm.smax.v64i8(<64 x i8> undef, <64 x i8> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; SSSE3-LABEL: 'smax'
; SSSE3-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %I64 = call i64 @llvm.smax.i64(i64 undef, i64 undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V2I64 = call <2 x i64> @llvm.smax.v2i64(<2 x i64> undef, <2 x i64> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V4I64 = call <4 x i64> @llvm.smax.v4i64(<4 x i64> undef, <4 x i64> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V8I64 = call <8 x i64> @llvm.smax.v8i64(<8 x i64> undef, <8 x i64> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V2I64 = call <2 x i64> @llvm.smax.v2i64(<2 x i64> undef, <2 x i64> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 14 for instruction: %V4I64 = call <4 x i64> @llvm.smax.v4i64(<4 x i64> undef, <4 x i64> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 28 for instruction: %V8I64 = call <8 x i64> @llvm.smax.v8i64(<8 x i64> undef, <8 x i64> undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %I32 = call i32 @llvm.smax.i32(i32 undef, i32 undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V4I32 = call <4 x i32> @llvm.smax.v4i32(<4 x i32> undef, <4 x i32> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V8I32 = call <8 x i32> @llvm.smax.v8i32(<8 x i32> undef, <8 x i32> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V16I32 = call <16 x i32> @llvm.smax.v16i32(<16 x i32> undef, <16 x i32> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V4I32 = call <4 x i32> @llvm.smax.v4i32(<4 x i32> undef, <4 x i32> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V8I32 = call <8 x i32> @llvm.smax.v8i32(<8 x i32> undef, <8 x i32> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %V16I32 = call <16 x i32> @llvm.smax.v16i32(<16 x i32> undef, <16 x i32> undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %I16 = call i16 @llvm.smax.i16(i16 undef, i16 undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V8I16 = call <8 x i16> @llvm.smax.v8i16(<8 x i16> undef, <8 x i16> undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V16I16 = call <16 x i16> @llvm.smax.v16i16(<16 x i16> undef, <16 x i16> undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V32I16 = call <32 x i16> @llvm.smax.v32i16(<32 x i16> undef, <32 x i16> undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %I8 = call i8 @llvm.smax.i8(i8 undef, i8 undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V16I8 = call <16 x i8> @llvm.smax.v16i8(<16 x i8> undef, <16 x i8> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V32I8 = call <32 x i8> @llvm.smax.v32i8(<32 x i8> undef, <32 x i8> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V64I8 = call <64 x i8> @llvm.smax.v64i8(<64 x i8> undef, <64 x i8> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V16I8 = call <16 x i8> @llvm.smax.v16i8(<16 x i8> undef, <16 x i8> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V32I8 = call <32 x i8> @llvm.smax.v32i8(<32 x i8> undef, <32 x i8> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %V64I8 = call <64 x i8> @llvm.smax.v64i8(<64 x i8> undef, <64 x i8> undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; SSE42-LABEL: 'smax'
@@ -227,40 +227,40 @@ declare <64 x i8> @llvm.smin.v64i8(<64 x i8>, <64 x i8>)
define i32 @smin(i32 %arg) {
; SSE2-LABEL: 'smin'
; SSE2-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %I64 = call i64 @llvm.smin.i64(i64 undef, i64 undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V2I64 = call <2 x i64> @llvm.smin.v2i64(<2 x i64> undef, <2 x i64> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V4I64 = call <4 x i64> @llvm.smin.v4i64(<4 x i64> undef, <4 x i64> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V8I64 = call <8 x i64> @llvm.smin.v8i64(<8 x i64> undef, <8 x i64> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V2I64 = call <2 x i64> @llvm.smin.v2i64(<2 x i64> undef, <2 x i64> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 14 for instruction: %V4I64 = call <4 x i64> @llvm.smin.v4i64(<4 x i64> undef, <4 x i64> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 28 for instruction: %V8I64 = call <8 x i64> @llvm.smin.v8i64(<8 x i64> undef, <8 x i64> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %I32 = call i32 @llvm.smin.i32(i32 undef, i32 undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V4I32 = call <4 x i32> @llvm.smin.v4i32(<4 x i32> undef, <4 x i32> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V8I32 = call <8 x i32> @llvm.smin.v8i32(<8 x i32> undef, <8 x i32> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V16I32 = call <16 x i32> @llvm.smin.v16i32(<16 x i32> undef, <16 x i32> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V4I32 = call <4 x i32> @llvm.smin.v4i32(<4 x i32> undef, <4 x i32> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V8I32 = call <8 x i32> @llvm.smin.v8i32(<8 x i32> undef, <8 x i32> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %V16I32 = call <16 x i32> @llvm.smin.v16i32(<16 x i32> undef, <16 x i32> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %I16 = call i16 @llvm.smin.i16(i16 undef, i16 undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V8I16 = call <8 x i16> @llvm.smin.v8i16(<8 x i16> undef, <8 x i16> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V16I16 = call <16 x i16> @llvm.smin.v16i16(<16 x i16> undef, <16 x i16> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V32I16 = call <32 x i16> @llvm.smin.v32i16(<32 x i16> undef, <32 x i16> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %I8 = call i8 @llvm.smin.i8(i8 undef, i8 undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V16I8 = call <16 x i8> @llvm.smin.v16i8(<16 x i8> undef, <16 x i8> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V32I8 = call <32 x i8> @llvm.smin.v32i8(<32 x i8> undef, <32 x i8> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V64I8 = call <64 x i8> @llvm.smin.v64i8(<64 x i8> undef, <64 x i8> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V16I8 = call <16 x i8> @llvm.smin.v16i8(<16 x i8> undef, <16 x i8> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V32I8 = call <32 x i8> @llvm.smin.v32i8(<32 x i8> undef, <32 x i8> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %V64I8 = call <64 x i8> @llvm.smin.v64i8(<64 x i8> undef, <64 x i8> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; SSSE3-LABEL: 'smin'
; SSSE3-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %I64 = call i64 @llvm.smin.i64(i64 undef, i64 undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V2I64 = call <2 x i64> @llvm.smin.v2i64(<2 x i64> undef, <2 x i64> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V4I64 = call <4 x i64> @llvm.smin.v4i64(<4 x i64> undef, <4 x i64> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V8I64 = call <8 x i64> @llvm.smin.v8i64(<8 x i64> undef, <8 x i64> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V2I64 = call <2 x i64> @llvm.smin.v2i64(<2 x i64> undef, <2 x i64> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 14 for instruction: %V4I64 = call <4 x i64> @llvm.smin.v4i64(<4 x i64> undef, <4 x i64> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 28 for instruction: %V8I64 = call <8 x i64> @llvm.smin.v8i64(<8 x i64> undef, <8 x i64> undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %I32 = call i32 @llvm.smin.i32(i32 undef, i32 undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V4I32 = call <4 x i32> @llvm.smin.v4i32(<4 x i32> undef, <4 x i32> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V8I32 = call <8 x i32> @llvm.smin.v8i32(<8 x i32> undef, <8 x i32> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V16I32 = call <16 x i32> @llvm.smin.v16i32(<16 x i32> undef, <16 x i32> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V4I32 = call <4 x i32> @llvm.smin.v4i32(<4 x i32> undef, <4 x i32> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V8I32 = call <8 x i32> @llvm.smin.v8i32(<8 x i32> undef, <8 x i32> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %V16I32 = call <16 x i32> @llvm.smin.v16i32(<16 x i32> undef, <16 x i32> undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %I16 = call i16 @llvm.smin.i16(i16 undef, i16 undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V8I16 = call <8 x i16> @llvm.smin.v8i16(<8 x i16> undef, <8 x i16> undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V16I16 = call <16 x i16> @llvm.smin.v16i16(<16 x i16> undef, <16 x i16> undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V32I16 = call <32 x i16> @llvm.smin.v32i16(<32 x i16> undef, <32 x i16> undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %I8 = call i8 @llvm.smin.i8(i8 undef, i8 undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V16I8 = call <16 x i8> @llvm.smin.v16i8(<16 x i8> undef, <16 x i8> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V32I8 = call <32 x i8> @llvm.smin.v32i8(<32 x i8> undef, <32 x i8> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V64I8 = call <64 x i8> @llvm.smin.v64i8(<64 x i8> undef, <64 x i8> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V16I8 = call <16 x i8> @llvm.smin.v16i8(<16 x i8> undef, <16 x i8> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V32I8 = call <32 x i8> @llvm.smin.v32i8(<32 x i8> undef, <32 x i8> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %V64I8 = call <64 x i8> @llvm.smin.v64i8(<64 x i8> undef, <64 x i8> undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; SSE42-LABEL: 'smin'
diff --git a/llvm/test/Analysis/CostModel/X86/arith-ssat.ll b/llvm/test/Analysis/CostModel/X86/arith-ssat.ll
index 64d2cd4a2f28e..b6112e3e21929 100644
--- a/llvm/test/Analysis/CostModel/X86/arith-ssat.ll
+++ b/llvm/test/Analysis/CostModel/X86/arith-ssat.ll
@@ -37,13 +37,13 @@ declare <64 x i8> @llvm.sadd.sat.v64i8(<64 x i8>, <64 x i8>)
define i32 @add(i32 %arg) {
; SSSE3-LABEL: 'add'
; SSSE3-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I64 = call i64 @llvm.sadd.sat.i64(i64 undef, i64 undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 23 for instruction: %V2I64 = call <2 x i64> @llvm.sadd.sat.v2i64(<2 x i64> undef, <2 x i64> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 45 for instruction: %V4I64 = call <4 x i64> @llvm.sadd.sat.v4i64(<4 x i64> undef, <4 x i64> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 89 for instruction: %V8I64 = call <8 x i64> @llvm.sadd.sat.v8i64(<8 x i64> undef, <8 x i64> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 21 for instruction: %V2I64 = call <2 x i64> @llvm.sadd.sat.v2i64(<2 x i64> undef, <2 x i64> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 41 for instruction: %V4I64 = call <4 x i64> @llvm.sadd.sat.v4i64(<4 x i64> undef, <4 x i64> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 81 for instruction: %V8I64 = call <8 x i64> @llvm.sadd.sat.v8i64(<8 x i64> undef, <8 x i64> undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I32 = call i32 @llvm.sadd.sat.i32(i32 undef, i32 undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 11 for instruction: %V4I32 = call <4 x i32> @llvm.sadd.sat.v4i32(<4 x i32> undef, <4 x i32> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 21 for instruction: %V8I32 = call <8 x i32> @llvm.sadd.sat.v8i32(<8 x i32> undef, <8 x i32> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 41 for instruction: %V16I32 = call <16 x i32> @llvm.sadd.sat.v16i32(<16 x i32> undef, <16 x i32> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V4I32 = call <4 x i32> @llvm.sadd.sat.v4i32(<4 x i32> undef, <4 x i32> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 17 for instruction: %V8I32 = call <8 x i32> @llvm.sadd.sat.v8i32(<8 x i32> undef, <8 x i32> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 33 for instruction: %V16I32 = call <16 x i32> @llvm.sadd.sat.v16i32(<16 x i32> undef, <16 x i32> undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I16 = call i16 @llvm.sadd.sat.i16(i16 undef, i16 undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V8I16 = call <8 x i16> @llvm.sadd.sat.v8i16(<8 x i16> undef, <8 x i16> undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V16I16 = call <16 x i16> @llvm.sadd.sat.v16i16(<16 x i16> undef, <16 x i16> undef)
@@ -271,13 +271,13 @@ declare <64 x i8> @llvm.ssub.sat.v64i8(<64 x i8>, <64 x i8>)
define i32 @sub(i32 %arg) {
; SSSE3-LABEL: 'sub'
; SSSE3-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I64 = call i64 @llvm.ssub.sat.i64(i64 undef, i64 undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 23 for instruction: %V2I64 = call <2 x i64> @llvm.ssub.sat.v2i64(<2 x i64> undef, <2 x i64> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 45 for instruction: %V4I64 = call <4 x i64> @llvm.ssub.sat.v4i64(<4 x i64> undef, <4 x i64> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 89 for instruction: %V8I64 = call <8 x i64> @llvm.ssub.sat.v8i64(<8 x i64> undef, <8 x i64> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 21 for instruction: %V2I64 = call <2 x i64> @llvm.ssub.sat.v2i64(<2 x i64> undef, <2 x i64> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 41 for instruction: %V4I64 = call <4 x i64> @llvm.ssub.sat.v4i64(<4 x i64> undef, <4 x i64> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 81 for instruction: %V8I64 = call <8 x i64> @llvm.ssub.sat.v8i64(<8 x i64> undef, <8 x i64> undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I32 = call i32 @llvm.ssub.sat.i32(i32 undef, i32 undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 11 for instruction: %V4I32 = call <4 x i32> @llvm.ssub.sat.v4i32(<4 x i32> undef, <4 x i32> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 21 for instruction: %V8I32 = call <8 x i32> @llvm.ssub.sat.v8i32(<8 x i32> undef, <8 x i32> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 41 for instruction: %V16I32 = call <16 x i32> @llvm.ssub.sat.v16i32(<16 x i32> undef, <16 x i32> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V4I32 = call <4 x i32> @llvm.ssub.sat.v4i32(<4 x i32> undef, <4 x i32> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 17 for instruction: %V8I32 = call <8 x i32> @llvm.ssub.sat.v8i32(<8 x i32> undef, <8 x i32> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 33 for instruction: %V16I32 = call <16 x i32> @llvm.ssub.sat.v16i32(<16 x i32> undef, <16 x i32> undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I16 = call i16 @llvm.ssub.sat.i16(i16 undef, i16 undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V8I16 = call <8 x i16> @llvm.ssub.sat.v8i16(<8 x i16> undef, <8 x i16> undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V16I16 = call <16 x i16> @llvm.ssub.sat.v16i16(<16 x i16> undef, <16 x i16> undef)
diff --git a/llvm/test/Analysis/CostModel/X86/arith-uminmax.ll b/llvm/test/Analysis/CostModel/X86/arith-uminmax.ll
index 71cdb0d8ef498..0b54892254c9a 100644
--- a/llvm/test/Analysis/CostModel/X86/arith-uminmax.ll
+++ b/llvm/test/Analysis/CostModel/X86/arith-uminmax.ll
@@ -31,13 +31,13 @@ declare <64 x i8> @llvm.umax.v64i8(<64 x i8>, <64 x i8>)
define i32 @umax(i32 %arg) {
; SSE2-LABEL: 'umax'
; SSE2-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %I64 = call i64 @llvm.umax.i64(i64 undef, i64 undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 10 for instruction: %V2I64 = call <2 x i64> @llvm.umax.v2i64(<2 x i64> undef, <2 x i64> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 20 for instruction: %V4I64 = call <4 x i64> @llvm.umax.v4i64(<4 x i64> undef, <4 x i64> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 40 for instruction: %V8I64 = call <8 x i64> @llvm.umax.v8i64(<8 x i64> undef, <8 x i64> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V2I64 = call <2 x i64> @llvm.umax.v2i64(<2 x i64> undef, <2 x i64> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 18 for instruction: %V4I64 = call <4 x i64> @llvm.umax.v4i64(<4 x i64> undef, <4 x i64> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 36 for instruction: %V8I64 = call <8 x i64> @llvm.umax.v8i64(<8 x i64> undef, <8 x i64> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %I32 = call i32 @llvm.umax.i32(i32 undef, i32 undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V4I32 = call <4 x i32> @llvm.umax.v4i32(<4 x i32> undef, <4 x i32> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %V8I32 = call <8 x i32> @llvm.umax.v8i32(<8 x i32> undef, <8 x i32> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 24 for instruction: %V16I32 = call <16 x i32> @llvm.umax.v16i32(<16 x i32> undef, <16 x i32> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 5 for instruction: %V4I32 = call <4 x i32> @llvm.umax.v4i32(<4 x i32> undef, <4 x i32> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 10 for instruction: %V8I32 = call <8 x i32> @llvm.umax.v8i32(<8 x i32> undef, <8 x i32> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 20 for instruction: %V16I32 = call <16 x i32> @llvm.umax.v16i32(<16 x i32> undef, <16 x i32> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %I16 = call i16 @llvm.umax.i16(i16 undef, i16 undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V8I16 = call <8 x i16> @llvm.umax.v8i16(<8 x i16> undef, <8 x i16> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V16I16 = call <16 x i16> @llvm.umax.v16i16(<16 x i16> undef, <16 x i16> undef)
@@ -50,13 +50,13 @@ define i32 @umax(i32 %arg) {
;
; SSSE3-LABEL: 'umax'
; SSSE3-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %I64 = call i64 @llvm.umax.i64(i64 undef, i64 undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 10 for instruction: %V2I64 = call <2 x i64> @llvm.umax.v2i64(<2 x i64> undef, <2 x i64> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 20 for instruction: %V4I64 = call <4 x i64> @llvm.umax.v4i64(<4 x i64> undef, <4 x i64> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 40 for instruction: %V8I64 = call <8 x i64> @llvm.umax.v8i64(<8 x i64> undef, <8 x i64> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V2I64 = call <2 x i64> @llvm.umax.v2i64(<2 x i64> undef, <2 x i64> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 18 for instruction: %V4I64 = call <4 x i64> @llvm.umax.v4i64(<4 x i64> undef, <4 x i64> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 36 for instruction: %V8I64 = call <8 x i64> @llvm.umax.v8i64(<8 x i64> undef, <8 x i64> undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %I32 = call i32 @llvm.umax.i32(i32 undef, i32 undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V4I32 = call <4 x i32> @llvm.umax.v4i32(<4 x i32> undef, <4 x i32> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %V8I32 = call <8 x i32> @llvm.umax.v8i32(<8 x i32> undef, <8 x i32> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 24 for instruction: %V16I32 = call <16 x i32> @llvm.umax.v16i32(<16 x i32> undef, <16 x i32> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 5 for instruction: %V4I32 = call <4 x i32> @llvm.umax.v4i32(<4 x i32> undef, <4 x i32> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 10 for instruction: %V8I32 = call <8 x i32> @llvm.umax.v8i32(<8 x i32> undef, <8 x i32> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 20 for instruction: %V16I32 = call <16 x i32> @llvm.umax.v16i32(<16 x i32> undef, <16 x i32> undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %I16 = call i16 @llvm.umax.i16(i16 undef, i16 undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V8I16 = call <8 x i16> @llvm.umax.v8i16(<8 x i16> undef, <8 x i16> undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V16I16 = call <16 x i16> @llvm.umax.v16i16(<16 x i16> undef, <16 x i16> undef)
@@ -227,13 +227,13 @@ declare <64 x i8> @llvm.umin.v64i8(<64 x i8>, <64 x i8>)
define i32 @umin(i32 %arg) {
; SSE2-LABEL: 'umin'
; SSE2-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %I64 = call i64 @llvm.umin.i64(i64 undef, i64 undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 10 for instruction: %V2I64 = call <2 x i64> @llvm.umin.v2i64(<2 x i64> undef, <2 x i64> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 20 for instruction: %V4I64 = call <4 x i64> @llvm.umin.v4i64(<4 x i64> undef, <4 x i64> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 40 for instruction: %V8I64 = call <8 x i64> @llvm.umin.v8i64(<8 x i64> undef, <8 x i64> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V2I64 = call <2 x i64> @llvm.umin.v2i64(<2 x i64> undef, <2 x i64> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 18 for instruction: %V4I64 = call <4 x i64> @llvm.umin.v4i64(<4 x i64> undef, <4 x i64> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 36 for instruction: %V8I64 = call <8 x i64> @llvm.umin.v8i64(<8 x i64> undef, <8 x i64> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %I32 = call i32 @llvm.umin.i32(i32 undef, i32 undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V4I32 = call <4 x i32> @llvm.umin.v4i32(<4 x i32> undef, <4 x i32> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %V8I32 = call <8 x i32> @llvm.umin.v8i32(<8 x i32> undef, <8 x i32> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 24 for instruction: %V16I32 = call <16 x i32> @llvm.umin.v16i32(<16 x i32> undef, <16 x i32> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 5 for instruction: %V4I32 = call <4 x i32> @llvm.umin.v4i32(<4 x i32> undef, <4 x i32> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 10 for instruction: %V8I32 = call <8 x i32> @llvm.umin.v8i32(<8 x i32> undef, <8 x i32> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 20 for instruction: %V16I32 = call <16 x i32> @llvm.umin.v16i32(<16 x i32> undef, <16 x i32> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %I16 = call i16 @llvm.umin.i16(i16 undef, i16 undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V8I16 = call <8 x i16> @llvm.umin.v8i16(<8 x i16> undef, <8 x i16> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V16I16 = call <16 x i16> @llvm.umin.v16i16(<16 x i16> undef, <16 x i16> undef)
@@ -246,13 +246,13 @@ define i32 @umin(i32 %arg) {
;
; SSSE3-LABEL: 'umin'
; SSSE3-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %I64 = call i64 @llvm.umin.i64(i64 undef, i64 undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 10 for instruction: %V2I64 = call <2 x i64> @llvm.umin.v2i64(<2 x i64> undef, <2 x i64> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 20 for instruction: %V4I64 = call <4 x i64> @llvm.umin.v4i64(<4 x i64> undef, <4 x i64> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 40 for instruction: %V8I64 = call <8 x i64> @llvm.umin.v8i64(<8 x i64> undef, <8 x i64> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V2I64 = call <2 x i64> @llvm.umin.v2i64(<2 x i64> undef, <2 x i64> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 18 for instruction: %V4I64 = call <4 x i64> @llvm.umin.v4i64(<4 x i64> undef, <4 x i64> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 36 for instruction: %V8I64 = call <8 x i64> @llvm.umin.v8i64(<8 x i64> undef, <8 x i64> undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %I32 = call i32 @llvm.umin.i32(i32 undef, i32 undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V4I32 = call <4 x i32> @llvm.umin.v4i32(<4 x i32> undef, <4 x i32> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %V8I32 = call <8 x i32> @llvm.umin.v8i32(<8 x i32> undef, <8 x i32> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 24 for instruction: %V16I32 = call <16 x i32> @llvm.umin.v16i32(<16 x i32> undef, <16 x i32> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 5 for instruction: %V4I32 = call <4 x i32> @llvm.umin.v4i32(<4 x i32> undef, <4 x i32> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 10 for instruction: %V8I32 = call <8 x i32> @llvm.umin.v8i32(<8 x i32> undef, <8 x i32> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 20 for instruction: %V16I32 = call <16 x i32> @llvm.umin.v16i32(<16 x i32> undef, <16 x i32> undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %I16 = call i16 @llvm.umin.i16(i16 undef, i16 undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V8I16 = call <8 x i16> @llvm.umin.v8i16(<8 x i16> undef, <8 x i16> undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V16I16 = call <16 x i16> @llvm.umin.v16i16(<16 x i16> undef, <16 x i16> undef)
diff --git a/llvm/test/Analysis/CostModel/X86/arith-usat.ll b/llvm/test/Analysis/CostModel/X86/arith-usat.ll
index 398c12582ff3c..8f8a80aded868 100644
--- a/llvm/test/Analysis/CostModel/X86/arith-usat.ll
+++ b/llvm/test/Analysis/CostModel/X86/arith-usat.ll
@@ -38,13 +38,13 @@ declare <64 x i8> @llvm.uadd.sat.v64i8(<64 x i8>, <64 x i8>)
define i32 @add(i32 %arg) {
; SSSE3-LABEL: 'add'
; SSSE3-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %I64 = call i64 @llvm.uadd.sat.i64(i64 undef, i64 undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 11 for instruction: %V2I64 = call <2 x i64> @llvm.uadd.sat.v2i64(<2 x i64> undef, <2 x i64> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 22 for instruction: %V4I64 = call <4 x i64> @llvm.uadd.sat.v4i64(<4 x i64> undef, <4 x i64> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 44 for instruction: %V8I64 = call <8 x i64> @llvm.uadd.sat.v8i64(<8 x i64> undef, <8 x i64> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 10 for instruction: %V2I64 = call <2 x i64> @llvm.uadd.sat.v2i64(<2 x i64> undef, <2 x i64> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 20 for instruction: %V4I64 = call <4 x i64> @llvm.uadd.sat.v4i64(<4 x i64> undef, <4 x i64> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 40 for instruction: %V8I64 = call <8 x i64> @llvm.uadd.sat.v8i64(<8 x i64> undef, <8 x i64> undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %I32 = call i32 @llvm.uadd.sat.i32(i32 undef, i32 undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V4I32 = call <4 x i32> @llvm.uadd.sat.v4i32(<4 x i32> undef, <4 x i32> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 14 for instruction: %V8I32 = call <8 x i32> @llvm.uadd.sat.v8i32(<8 x i32> undef, <8 x i32> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 28 for instruction: %V16I32 = call <16 x i32> @llvm.uadd.sat.v16i32(<16 x i32> undef, <16 x i32> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V4I32 = call <4 x i32> @llvm.uadd.sat.v4i32(<4 x i32> undef, <4 x i32> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %V8I32 = call <8 x i32> @llvm.uadd.sat.v8i32(<8 x i32> undef, <8 x i32> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 24 for instruction: %V16I32 = call <16 x i32> @llvm.uadd.sat.v16i32(<16 x i32> undef, <16 x i32> undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %I16 = call i16 @llvm.uadd.sat.i16(i16 undef, i16 undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V8I16 = call <8 x i16> @llvm.uadd.sat.v8i16(<8 x i16> undef, <8 x i16> undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V16I16 = call <16 x i16> @llvm.uadd.sat.v16i16(<16 x i16> undef, <16 x i16> undef)
@@ -272,13 +272,13 @@ declare <64 x i8> @llvm.usub.sat.v64i8(<64 x i8>, <64 x i8>)
define i32 @sub(i32 %arg) {
; SSSE3-LABEL: 'sub'
; SSSE3-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %I64 = call i64 @llvm.usub.sat.i64(i64 undef, i64 undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 11 for instruction: %V2I64 = call <2 x i64> @llvm.usub.sat.v2i64(<2 x i64> undef, <2 x i64> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 22 for instruction: %V4I64 = call <4 x i64> @llvm.usub.sat.v4i64(<4 x i64> undef, <4 x i64> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 44 for instruction: %V8I64 = call <8 x i64> @llvm.usub.sat.v8i64(<8 x i64> undef, <8 x i64> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 10 for instruction: %V2I64 = call <2 x i64> @llvm.usub.sat.v2i64(<2 x i64> undef, <2 x i64> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 20 for instruction: %V4I64 = call <4 x i64> @llvm.usub.sat.v4i64(<4 x i64> undef, <4 x i64> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 40 for instruction: %V8I64 = call <8 x i64> @llvm.usub.sat.v8i64(<8 x i64> undef, <8 x i64> undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %I32 = call i32 @llvm.usub.sat.i32(i32 undef, i32 undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V4I32 = call <4 x i32> @llvm.usub.sat.v4i32(<4 x i32> undef, <4 x i32> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 14 for instruction: %V8I32 = call <8 x i32> @llvm.usub.sat.v8i32(<8 x i32> undef, <8 x i32> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 28 for instruction: %V16I32 = call <16 x i32> @llvm.usub.sat.v16i32(<16 x i32> undef, <16 x i32> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V4I32 = call <4 x i32> @llvm.usub.sat.v4i32(<4 x i32> undef, <4 x i32> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %V8I32 = call <8 x i32> @llvm.usub.sat.v8i32(<8 x i32> undef, <8 x i32> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 24 for instruction: %V16I32 = call <16 x i32> @llvm.usub.sat.v16i32(<16 x i32> undef, <16 x i32> undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %I16 = call i16 @llvm.usub.sat.i16(i16 undef, i16 undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V8I16 = call <8 x i16> @llvm.usub.sat.v8i16(<8 x i16> undef, <8 x i16> undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V16I16 = call <16 x i16> @llvm.usub.sat.v16i16(<16 x i16> undef, <16 x i16> undef)
diff --git a/llvm/test/Analysis/CostModel/X86/fptoi_sat.ll b/llvm/test/Analysis/CostModel/X86/fptoi_sat.ll
index df1625adbe9c1..ad6d63fbf0663 100644
--- a/llvm/test/Analysis/CostModel/X86/fptoi_sat.ll
+++ b/llvm/test/Analysis/CostModel/X86/fptoi_sat.ll
@@ -32,85 +32,85 @@ define void @casts() {
; SSE2-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %f64u32 = call i32 @llvm.fptoui.sat.i32.f64(double undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 14 for instruction: %f64s64 = call i64 @llvm.fptosi.sat.i64.f64(double undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 23 for instruction: %f64u64 = call i64 @llvm.fptoui.sat.i64.f64(double undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 20 for instruction: %v2f32s1 = call <2 x i1> @llvm.fptosi.sat.v2i1.v2f32(<2 x float> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 19 for instruction: %v2f32s1 = call <2 x i1> @llvm.fptosi.sat.v2i1.v2f32(<2 x float> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 15 for instruction: %v2f32u1 = call <2 x i1> @llvm.fptoui.sat.v2i1.v2f32(<2 x float> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 19 for instruction: %v2f32s8 = call <2 x i8> @llvm.fptosi.sat.v2i8.v2f32(<2 x float> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 18 for instruction: %v2f32s8 = call <2 x i8> @llvm.fptosi.sat.v2i8.v2f32(<2 x float> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 14 for instruction: %v2f32u8 = call <2 x i8> @llvm.fptoui.sat.v2i8.v2f32(<2 x float> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 18 for instruction: %v2f32s16 = call <2 x i16> @llvm.fptosi.sat.v2i16.v2f32(<2 x float> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 17 for instruction: %v2f32s16 = call <2 x i16> @llvm.fptosi.sat.v2i16.v2f32(<2 x float> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 13 for instruction: %v2f32u16 = call <2 x i16> @llvm.fptoui.sat.v2i16.v2f32(<2 x float> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 17 for instruction: %v2f32s32 = call <2 x i32> @llvm.fptosi.sat.v2i32.v2f32(<2 x float> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %v2f32s32 = call <2 x i32> @llvm.fptosi.sat.v2i32.v2f32(<2 x float> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %v2f32u32 = call <2 x i32> @llvm.fptoui.sat.v2i32.v2f32(<2 x float> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 27 for instruction: %v2f32s64 = call <2 x i64> @llvm.fptosi.sat.v2i64.v2f32(<2 x float> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 26 for instruction: %v2f32s64 = call <2 x i64> @llvm.fptosi.sat.v2i64.v2f32(<2 x float> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 22 for instruction: %v2f32u64 = call <2 x i64> @llvm.fptoui.sat.v2i64.v2f32(<2 x float> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 20 for instruction: %v2f64s1 = call <2 x i1> @llvm.fptosi.sat.v2i1.v2f64(<2 x double> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 19 for instruction: %v2f64s1 = call <2 x i1> @llvm.fptosi.sat.v2i1.v2f64(<2 x double> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 15 for instruction: %v2f64u1 = call <2 x i1> @llvm.fptoui.sat.v2i1.v2f64(<2 x double> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 19 for instruction: %v2f64s8 = call <2 x i8> @llvm.fptosi.sat.v2i8.v2f64(<2 x double> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 18 for instruction: %v2f64s8 = call <2 x i8> @llvm.fptosi.sat.v2i8.v2f64(<2 x double> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 14 for instruction: %v2f64u8 = call <2 x i8> @llvm.fptoui.sat.v2i8.v2f64(<2 x double> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 18 for instruction: %v2f64s16 = call <2 x i16> @llvm.fptosi.sat.v2i16.v2f64(<2 x double> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 17 for instruction: %v2f64s16 = call <2 x i16> @llvm.fptosi.sat.v2i16.v2f64(<2 x double> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 13 for instruction: %v2f64u16 = call <2 x i16> @llvm.fptoui.sat.v2i16.v2f64(<2 x double> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 17 for instruction: %v2f64s32 = call <2 x i32> @llvm.fptosi.sat.v2i32.v2f64(<2 x double> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %v2f64s32 = call <2 x i32> @llvm.fptosi.sat.v2i32.v2f64(<2 x double> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %v2f64u32 = call <2 x i32> @llvm.fptoui.sat.v2i32.v2f64(<2 x double> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 27 for instruction: %v2f64s64 = call <2 x i64> @llvm.fptosi.sat.v2i64.v2f64(<2 x double> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 26 for instruction: %v2f64s64 = call <2 x i64> @llvm.fptosi.sat.v2i64.v2f64(<2 x double> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 44 for instruction: %v2f64u64 = call <2 x i64> @llvm.fptoui.sat.v2i64.v2f64(<2 x double> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 17 for instruction: %v4f32s1 = call <4 x i1> @llvm.fptosi.sat.v4i1.v4f32(<4 x float> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %v4f32s1 = call <4 x i1> @llvm.fptosi.sat.v4i1.v4f32(<4 x float> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %v4f32u1 = call <4 x i1> @llvm.fptoui.sat.v4i1.v4f32(<4 x float> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 19 for instruction: %v4f32s8 = call <4 x i8> @llvm.fptosi.sat.v4i8.v4f32(<4 x float> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 18 for instruction: %v4f32s8 = call <4 x i8> @llvm.fptosi.sat.v4i8.v4f32(<4 x float> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 14 for instruction: %v4f32u8 = call <4 x i8> @llvm.fptoui.sat.v4i8.v4f32(<4 x float> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 18 for instruction: %v4f32s16 = call <4 x i16> @llvm.fptosi.sat.v4i16.v4f32(<4 x float> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 17 for instruction: %v4f32s16 = call <4 x i16> @llvm.fptosi.sat.v4i16.v4f32(<4 x float> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 13 for instruction: %v4f32u16 = call <4 x i16> @llvm.fptoui.sat.v4i16.v4f32(<4 x float> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 17 for instruction: %v4f32s32 = call <4 x i32> @llvm.fptosi.sat.v4i32.v4f32(<4 x float> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %v4f32s32 = call <4 x i32> @llvm.fptosi.sat.v4i32.v4f32(<4 x float> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %v4f32u32 = call <4 x i32> @llvm.fptoui.sat.v4i32.v4f32(<4 x float> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 45 for instruction: %v4f32s64 = call <4 x i64> @llvm.fptosi.sat.v4i64.v4f32(<4 x float> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 43 for instruction: %v4f32s64 = call <4 x i64> @llvm.fptosi.sat.v4i64.v4f32(<4 x float> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 37 for instruction: %v4f32u64 = call <4 x i64> @llvm.fptoui.sat.v4i64.v4f32(<4 x float> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 31 for instruction: %v4f64s1 = call <4 x i1> @llvm.fptosi.sat.v4i1.v4f64(<4 x double> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 30 for instruction: %v4f64s1 = call <4 x i1> @llvm.fptosi.sat.v4i1.v4f64(<4 x double> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %v4f64u1 = call <4 x i1> @llvm.fptoui.sat.v4i1.v4f64(<4 x double> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 35 for instruction: %v4f64s8 = call <4 x i8> @llvm.fptosi.sat.v4i8.v4f64(<4 x double> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 34 for instruction: %v4f64s8 = call <4 x i8> @llvm.fptosi.sat.v4i8.v4f64(<4 x double> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 28 for instruction: %v4f64u8 = call <4 x i8> @llvm.fptoui.sat.v4i8.v4f64(<4 x double> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 33 for instruction: %v4f64s16 = call <4 x i16> @llvm.fptosi.sat.v4i16.v4f64(<4 x double> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %v4f64s16 = call <4 x i16> @llvm.fptosi.sat.v4i16.v4f64(<4 x double> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 26 for instruction: %v4f64u16 = call <4 x i16> @llvm.fptoui.sat.v4i16.v4f64(<4 x double> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 31 for instruction: %v4f64s32 = call <4 x i32> @llvm.fptosi.sat.v4i32.v4f64(<4 x double> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 30 for instruction: %v4f64s32 = call <4 x i32> @llvm.fptosi.sat.v4i32.v4f64(<4 x double> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %v4f64u32 = call <4 x i32> @llvm.fptoui.sat.v4i32.v4f64(<4 x double> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 54 for instruction: %v4f64s64 = call <4 x i64> @llvm.fptosi.sat.v4i64.v4f64(<4 x double> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 52 for instruction: %v4f64s64 = call <4 x i64> @llvm.fptosi.sat.v4i64.v4f64(<4 x double> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 88 for instruction: %v4f64u64 = call <4 x i64> @llvm.fptoui.sat.v4i64.v4f64(<4 x double> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 33 for instruction: %v8f32s1 = call <8 x i1> @llvm.fptosi.sat.v8i1.v8f32(<8 x float> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %v8f32s1 = call <8 x i1> @llvm.fptosi.sat.v8i1.v8f32(<8 x float> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 26 for instruction: %v8f32u1 = call <8 x i1> @llvm.fptoui.sat.v8i1.v8f32(<8 x float> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 35 for instruction: %v8f32s8 = call <8 x i8> @llvm.fptosi.sat.v8i8.v8f32(<8 x float> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 34 for instruction: %v8f32s8 = call <8 x i8> @llvm.fptosi.sat.v8i8.v8f32(<8 x float> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 28 for instruction: %v8f32u8 = call <8 x i8> @llvm.fptoui.sat.v8i8.v8f32(<8 x float> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 33 for instruction: %v8f32s16 = call <8 x i16> @llvm.fptosi.sat.v8i16.v8f32(<8 x float> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %v8f32s16 = call <8 x i16> @llvm.fptosi.sat.v8i16.v8f32(<8 x float> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 26 for instruction: %v8f32u16 = call <8 x i16> @llvm.fptoui.sat.v8i16.v8f32(<8 x float> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 34 for instruction: %v8f32s32 = call <8 x i32> @llvm.fptosi.sat.v8i32.v8f32(<8 x float> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %v8f32s32 = call <8 x i32> @llvm.fptosi.sat.v8i32.v8f32(<8 x float> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %v8f32u32 = call <8 x i32> @llvm.fptoui.sat.v8i32.v8f32(<8 x float> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 90 for instruction: %v8f32s64 = call <8 x i64> @llvm.fptosi.sat.v8i64.v8f32(<8 x float> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 86 for instruction: %v8f32s64 = call <8 x i64> @llvm.fptosi.sat.v8i64.v8f32(<8 x float> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 74 for instruction: %v8f32u64 = call <8 x i64> @llvm.fptoui.sat.v8i64.v8f32(<8 x float> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 63 for instruction: %v8f64s1 = call <8 x i1> @llvm.fptosi.sat.v8i1.v8f64(<8 x double> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 62 for instruction: %v8f64s1 = call <8 x i1> @llvm.fptosi.sat.v8i1.v8f64(<8 x double> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 52 for instruction: %v8f64u1 = call <8 x i1> @llvm.fptoui.sat.v8i1.v8f64(<8 x double> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 67 for instruction: %v8f64s8 = call <8 x i8> @llvm.fptosi.sat.v8i8.v8f64(<8 x double> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 66 for instruction: %v8f64s8 = call <8 x i8> @llvm.fptosi.sat.v8i8.v8f64(<8 x double> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 56 for instruction: %v8f64u8 = call <8 x i8> @llvm.fptoui.sat.v8i8.v8f64(<8 x double> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 63 for instruction: %v8f64s16 = call <8 x i16> @llvm.fptosi.sat.v8i16.v8f64(<8 x double> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 62 for instruction: %v8f64s16 = call <8 x i16> @llvm.fptosi.sat.v8i16.v8f64(<8 x double> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 52 for instruction: %v8f64u16 = call <8 x i16> @llvm.fptoui.sat.v8i16.v8f64(<8 x double> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 62 for instruction: %v8f64s32 = call <8 x i32> @llvm.fptosi.sat.v8i32.v8f64(<8 x double> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 60 for instruction: %v8f64s32 = call <8 x i32> @llvm.fptosi.sat.v8i32.v8f64(<8 x double> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 64 for instruction: %v8f64u32 = call <8 x i32> @llvm.fptoui.sat.v8i32.v8f64(<8 x double> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 108 for instruction: %v8f64s64 = call <8 x i64> @llvm.fptosi.sat.v8i64.v8f64(<8 x double> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 104 for instruction: %v8f64s64 = call <8 x i64> @llvm.fptosi.sat.v8i64.v8f64(<8 x double> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 176 for instruction: %v8f64u64 = call <8 x i64> @llvm.fptoui.sat.v8i64.v8f64(<8 x double> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 67 for instruction: %v16f32s1 = call <16 x i1> @llvm.fptosi.sat.v16i1.v16f32(<16 x float> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 66 for instruction: %v16f32s1 = call <16 x i1> @llvm.fptosi.sat.v16i1.v16f32(<16 x float> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 56 for instruction: %v16f32u1 = call <16 x i1> @llvm.fptoui.sat.v16i1.v16f32(<16 x float> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 67 for instruction: %v16f32s8 = call <16 x i8> @llvm.fptosi.sat.v16i8.v16f32(<16 x float> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 66 for instruction: %v16f32s8 = call <16 x i8> @llvm.fptosi.sat.v16i8.v16f32(<16 x float> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 56 for instruction: %v16f32u8 = call <16 x i8> @llvm.fptoui.sat.v16i8.v16f32(<16 x float> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 66 for instruction: %v16f32s16 = call <16 x i16> @llvm.fptosi.sat.v16i16.v16f32(<16 x float> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 64 for instruction: %v16f32s16 = call <16 x i16> @llvm.fptosi.sat.v16i16.v16f32(<16 x float> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 52 for instruction: %v16f32u16 = call <16 x i16> @llvm.fptoui.sat.v16i16.v16f32(<16 x float> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 68 for instruction: %v16f32s32 = call <16 x i32> @llvm.fptosi.sat.v16i32.v16f32(<16 x float> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 64 for instruction: %v16f32s32 = call <16 x i32> @llvm.fptosi.sat.v16i32.v16f32(<16 x float> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 64 for instruction: %v16f32u32 = call <16 x i32> @llvm.fptoui.sat.v16i32.v16f32(<16 x float> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 180 for instruction: %v16f32s64 = call <16 x i64> @llvm.fptosi.sat.v16i64.v16f32(<16 x float> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 172 for instruction: %v16f32s64 = call <16 x i64> @llvm.fptosi.sat.v16i64.v16f32(<16 x float> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 148 for instruction: %v16f32u64 = call <16 x i64> @llvm.fptoui.sat.v16i64.v16f32(<16 x float> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 131 for instruction: %v16f64s1 = call <16 x i1> @llvm.fptosi.sat.v16i1.v16f64(<16 x double> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 130 for instruction: %v16f64s1 = call <16 x i1> @llvm.fptosi.sat.v16i1.v16f64(<16 x double> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 112 for instruction: %v16f64u1 = call <16 x i1> @llvm.fptoui.sat.v16i1.v16f64(<16 x double> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 131 for instruction: %v16f64s8 = call <16 x i8> @llvm.fptosi.sat.v16i8.v16f64(<16 x double> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 130 for instruction: %v16f64s8 = call <16 x i8> @llvm.fptosi.sat.v16i8.v16f64(<16 x double> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 112 for instruction: %v16f64u8 = call <16 x i8> @llvm.fptoui.sat.v16i8.v16f64(<16 x double> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 126 for instruction: %v16f64s16 = call <16 x i16> @llvm.fptosi.sat.v16i16.v16f64(<16 x double> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 124 for instruction: %v16f64s16 = call <16 x i16> @llvm.fptosi.sat.v16i16.v16f64(<16 x double> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 104 for instruction: %v16f64u16 = call <16 x i16> @llvm.fptoui.sat.v16i16.v16f64(<16 x double> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 124 for instruction: %v16f64s32 = call <16 x i32> @llvm.fptosi.sat.v16i32.v16f64(<16 x double> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 120 for instruction: %v16f64s32 = call <16 x i32> @llvm.fptosi.sat.v16i32.v16f64(<16 x double> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 128 for instruction: %v16f64u32 = call <16 x i32> @llvm.fptoui.sat.v16i32.v16f64(<16 x double> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 216 for instruction: %v16f64s64 = call <16 x i64> @llvm.fptosi.sat.v16i64.v16f64(<16 x double> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 208 for instruction: %v16f64s64 = call <16 x i64> @llvm.fptosi.sat.v16i64.v16f64(<16 x double> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 352 for instruction: %v16f64u64 = call <16 x i64> @llvm.fptoui.sat.v16i64.v16f64(<16 x double> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
@@ -302,7 +302,7 @@ define void @casts() {
; 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 27 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 28 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 28 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)
@@ -312,7 +312,7 @@ define void @casts() {
; 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 53 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 54 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 56 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)
@@ -857,45 +857,45 @@ define void @fp16() {
; SSE2-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %f16u32 = call i32 @llvm.fptoui.sat.i32.f16(half undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 14 for instruction: %f16s64 = call i64 @llvm.fptosi.sat.i64.f16(half undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %f16u64 = call i64 @llvm.fptoui.sat.i64.f16(half undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 24 for instruction: %v2f16s1 = call <2 x i1> @llvm.fptosi.sat.v2i1.v2f16(<2 x half> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 23 for instruction: %v2f16s1 = call <2 x i1> @llvm.fptosi.sat.v2i1.v2f16(<2 x half> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 19 for instruction: %v2f16u1 = call <2 x i1> @llvm.fptoui.sat.v2i1.v2f16(<2 x half> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 33 for instruction: %v2f16s8 = call <2 x i8> @llvm.fptosi.sat.v2i8.v2f16(<2 x half> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %v2f16s8 = call <2 x i8> @llvm.fptosi.sat.v2i8.v2f16(<2 x half> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 28 for instruction: %v2f16u8 = call <2 x i8> @llvm.fptoui.sat.v2i8.v2f16(<2 x half> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 31 for instruction: %v2f16s16 = call <2 x i16> @llvm.fptosi.sat.v2i16.v2f16(<2 x half> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 30 for instruction: %v2f16s16 = call <2 x i16> @llvm.fptosi.sat.v2i16.v2f16(<2 x half> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 26 for instruction: %v2f16u16 = call <2 x i16> @llvm.fptoui.sat.v2i16.v2f16(<2 x half> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 30 for instruction: %v2f16s32 = call <2 x i32> @llvm.fptosi.sat.v2i32.v2f16(<2 x half> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 29 for instruction: %v2f16s32 = call <2 x i32> @llvm.fptosi.sat.v2i32.v2f16(<2 x half> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 25 for instruction: %v2f16u32 = call <2 x i32> @llvm.fptoui.sat.v2i32.v2f16(<2 x half> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 30 for instruction: %v2f16s64 = call <2 x i64> @llvm.fptosi.sat.v2i64.v2f16(<2 x half> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 29 for instruction: %v2f16s64 = call <2 x i64> @llvm.fptosi.sat.v2i64.v2f16(<2 x half> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 25 for instruction: %v2f16u64 = call <2 x i64> @llvm.fptoui.sat.v2i64.v2f16(<2 x half> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 46 for instruction: %v4f16s1 = call <4 x i1> @llvm.fptosi.sat.v4i1.v4f16(<4 x half> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 45 for instruction: %v4f16s1 = call <4 x i1> @llvm.fptosi.sat.v4i1.v4f16(<4 x half> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 39 for instruction: %v4f16u1 = call <4 x i1> @llvm.fptoui.sat.v4i1.v4f16(<4 x half> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 61 for instruction: %v4f16s8 = call <4 x i8> @llvm.fptosi.sat.v4i8.v4f16(<4 x half> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 60 for instruction: %v4f16s8 = call <4 x i8> @llvm.fptosi.sat.v4i8.v4f16(<4 x half> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 54 for instruction: %v4f16u8 = call <4 x i8> @llvm.fptoui.sat.v4i8.v4f16(<4 x half> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 61 for instruction: %v4f16s16 = call <4 x i16> @llvm.fptosi.sat.v4i16.v4f16(<4 x half> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 60 for instruction: %v4f16s16 = call <4 x i16> @llvm.fptosi.sat.v4i16.v4f16(<4 x half> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 54 for instruction: %v4f16u16 = call <4 x i16> @llvm.fptoui.sat.v4i16.v4f16(<4 x half> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 58 for instruction: %v4f16s32 = call <4 x i32> @llvm.fptosi.sat.v4i32.v4f16(<4 x half> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 57 for instruction: %v4f16s32 = call <4 x i32> @llvm.fptosi.sat.v4i32.v4f16(<4 x half> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 51 for instruction: %v4f16u32 = call <4 x i32> @llvm.fptoui.sat.v4i32.v4f16(<4 x half> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 60 for instruction: %v4f16s64 = call <4 x i64> @llvm.fptosi.sat.v4i64.v4f16(<4 x half> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 58 for instruction: %v4f16s64 = call <4 x i64> @llvm.fptosi.sat.v4i64.v4f16(<4 x half> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 50 for instruction: %v4f16u64 = call <4 x i64> @llvm.fptoui.sat.v4i64.v4f16(<4 x half> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 90 for instruction: %v8f16s1 = call <8 x i1> @llvm.fptosi.sat.v8i1.v8f16(<8 x half> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 89 for instruction: %v8f16s1 = call <8 x i1> @llvm.fptosi.sat.v8i1.v8f16(<8 x half> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 79 for instruction: %v8f16u1 = call <8 x i1> @llvm.fptoui.sat.v8i1.v8f16(<8 x half> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 119 for instruction: %v8f16s8 = call <8 x i8> @llvm.fptosi.sat.v8i8.v8f16(<8 x half> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 118 for instruction: %v8f16s8 = call <8 x i8> @llvm.fptosi.sat.v8i8.v8f16(<8 x half> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 108 for instruction: %v8f16u8 = call <8 x i8> @llvm.fptoui.sat.v8i8.v8f16(<8 x half> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 118 for instruction: %v8f16s16 = call <8 x i16> @llvm.fptosi.sat.v8i16.v8f16(<8 x half> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 117 for instruction: %v8f16s16 = call <8 x i16> @llvm.fptosi.sat.v8i16.v8f16(<8 x half> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 107 for instruction: %v8f16u16 = call <8 x i16> @llvm.fptoui.sat.v8i16.v8f16(<8 x half> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 116 for instruction: %v8f16s32 = call <8 x i32> @llvm.fptosi.sat.v8i32.v8f16(<8 x half> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 114 for instruction: %v8f16s32 = call <8 x i32> @llvm.fptosi.sat.v8i32.v8f16(<8 x half> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 102 for instruction: %v8f16u32 = call <8 x i32> @llvm.fptoui.sat.v8i32.v8f16(<8 x half> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 120 for instruction: %v8f16s64 = call <8 x i64> @llvm.fptosi.sat.v8i64.v8f16(<8 x half> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 116 for instruction: %v8f16s64 = call <8 x i64> @llvm.fptosi.sat.v8i64.v8f16(<8 x half> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 100 for instruction: %v8f16u64 = call <8 x i64> @llvm.fptoui.sat.v8i64.v8f16(<8 x half> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 178 for instruction: %v16f16s1 = call <16 x i1> @llvm.fptosi.sat.v16i1.v16f16(<16 x half> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 177 for instruction: %v16f16s1 = call <16 x i1> @llvm.fptosi.sat.v16i1.v16f16(<16 x half> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 159 for instruction: %v16f16u1 = call <16 x i1> @llvm.fptoui.sat.v16i1.v16f16(<16 x half> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 230 for instruction: %v16f16s8 = call <16 x i8> @llvm.fptosi.sat.v16i8.v16f16(<16 x half> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 229 for instruction: %v16f16s8 = call <16 x i8> @llvm.fptosi.sat.v16i8.v16f16(<16 x half> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 211 for instruction: %v16f16u8 = call <16 x i8> @llvm.fptoui.sat.v16i8.v16f16(<16 x half> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 236 for instruction: %v16f16s16 = call <16 x i16> @llvm.fptosi.sat.v16i16.v16f16(<16 x half> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 234 for instruction: %v16f16s16 = call <16 x i16> @llvm.fptosi.sat.v16i16.v16f16(<16 x half> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 214 for instruction: %v16f16u16 = call <16 x i16> @llvm.fptoui.sat.v16i16.v16f16(<16 x half> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 232 for instruction: %v16f16s32 = call <16 x i32> @llvm.fptosi.sat.v16i32.v16f16(<16 x half> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 228 for instruction: %v16f16s32 = call <16 x i32> @llvm.fptosi.sat.v16i32.v16f16(<16 x half> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 204 for instruction: %v16f16u32 = call <16 x i32> @llvm.fptoui.sat.v16i32.v16f16(<16 x half> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 240 for instruction: %v16f16s64 = call <16 x i64> @llvm.fptosi.sat.v16i64.v16f16(<16 x half> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 232 for instruction: %v16f16s64 = call <16 x i64> @llvm.fptosi.sat.v16i64.v16f16(<16 x half> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 200 for instruction: %v16f16u64 = call <16 x i64> @llvm.fptoui.sat.v16i64.v16f16(<16 x half> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
@@ -997,7 +997,7 @@ define void @fp16() {
; AVX1-NEXT: Cost Model: Found an estimated cost of 127 for instruction: %v16f16u1 = call <16 x i1> @llvm.fptoui.sat.v16i1.v16f16(<16 x half> undef)
; AVX1-NEXT: Cost Model: Found an estimated cost of 149 for instruction: %v16f16s8 = call <16 x i8> @llvm.fptosi.sat.v16i8.v16f16(<16 x half> undef)
; AVX1-NEXT: Cost Model: Found an estimated cost of 132 for instruction: %v16f16u8 = call <16 x i8> @llvm.fptoui.sat.v16i8.v16f16(<16 x half> undef)
-; AVX1-NEXT: Cost Model: Found an estimated cost of 151 for instruction: %v16f16s16 = call <16 x i16> @llvm.fptosi.sat.v16i16.v16f16(<16 x half> undef)
+; AVX1-NEXT: Cost Model: Found an estimated cost of 150 for instruction: %v16f16s16 = call <16 x i16> @llvm.fptosi.sat.v16i16.v16f16(<16 x half> undef)
; AVX1-NEXT: Cost Model: Found an estimated cost of 132 for instruction: %v16f16u16 = call <16 x i16> @llvm.fptoui.sat.v16i16.v16f16(<16 x half> undef)
; AVX1-NEXT: Cost Model: Found an estimated cost of 144 for instruction: %v16f16s32 = call <16 x i32> @llvm.fptosi.sat.v16i32.v16f16(<16 x half> undef)
; AVX1-NEXT: Cost Model: Found an estimated cost of 126 for instruction: %v16f16u32 = call <16 x i32> @llvm.fptoui.sat.v16i32.v16f16(<16 x half> undef)
diff --git a/llvm/test/Analysis/CostModel/X86/fshl.ll b/llvm/test/Analysis/CostModel/X86/fshl.ll
index ca0d549ee8557..0abfb22a0abeb 100644
--- a/llvm/test/Analysis/CostModel/X86/fshl.ll
+++ b/llvm/test/Analysis/CostModel/X86/fshl.ll
@@ -22,9 +22,9 @@ target triple = "x86_64-apple-macosx10.8.0"
define void @var_funnel_i64(i64 %a64, <2 x i64> %a128, <4 x i64> %a256, <8 x i64> %a512, i64 %b64, <2 x i64> %b128, <4 x i64> %b256, <8 x i64> %b512, i64 %c64, <2 x i64> %c128, <4 x i64> %c256, <8 x i64> %c512) {
; SSSE3-LABEL: 'var_funnel_i64'
; SSSE3-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I64 = call i64 @llvm.fshl.i64(i64 %a64, i64 %b64, i64 %c64)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 19 for instruction: %V2I64 = call <2 x i64> @llvm.fshl.v2i64(<2 x i64> %a128, <2 x i64> %b128, <2 x i64> %c128)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 38 for instruction: %V4I64 = call <4 x i64> @llvm.fshl.v4i64(<4 x i64> %a256, <4 x i64> %b256, <4 x i64> %c256)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 76 for instruction: %V8I64 = call <8 x i64> @llvm.fshl.v8i64(<8 x i64> %a512, <8 x i64> %b512, <8 x i64> %c512)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 18 for instruction: %V2I64 = call <2 x i64> @llvm.fshl.v2i64(<2 x i64> %a128, <2 x i64> %b128, <2 x i64> %c128)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 36 for instruction: %V4I64 = call <4 x i64> @llvm.fshl.v4i64(<4 x i64> %a256, <4 x i64> %b256, <4 x i64> %c256)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 72 for instruction: %V8I64 = call <8 x i64> @llvm.fshl.v8i64(<8 x i64> %a512, <8 x i64> %b512, <8 x i64> %c512)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; SSE42-LABEL: 'var_funnel_i64'
@@ -86,9 +86,9 @@ define void @var_funnel_i64(i64 %a64, <2 x i64> %a128, <4 x i64> %a256, <8 x i64
define void @var_funnel_i32(i32 %a32, <4 x i32> %a128, <8 x i32> %a256, <16 x i32> %a512, i32 %b32, <4 x i32> %b128, <8 x i32> %b256, <16 x i32> %b512, i32 %c32, <4 x i32> %c128, <8 x i32> %c256, <16 x i32> %c512) {
; SSSE3-LABEL: 'var_funnel_i32'
; SSSE3-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I32 = call i32 @llvm.fshl.i32(i32 %a32, i32 %b32, i32 %c32)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 35 for instruction: %V2I32 = call <4 x i32> @llvm.fshl.v4i32(<4 x i32> %a128, <4 x i32> %b128, <4 x i32> %c128)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 70 for instruction: %V4I32 = call <8 x i32> @llvm.fshl.v8i32(<8 x i32> %a256, <8 x i32> %b256, <8 x i32> %c256)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 140 for instruction: %V8I32 = call <16 x i32> @llvm.fshl.v16i32(<16 x i32> %a512, <16 x i32> %b512, <16 x i32> %c512)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 34 for instruction: %V2I32 = call <4 x i32> @llvm.fshl.v4i32(<4 x i32> %a128, <4 x i32> %b128, <4 x i32> %c128)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 68 for instruction: %V4I32 = call <8 x i32> @llvm.fshl.v8i32(<8 x i32> %a256, <8 x i32> %b256, <8 x i32> %c256)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 136 for instruction: %V8I32 = call <16 x i32> @llvm.fshl.v16i32(<16 x i32> %a512, <16 x i32> %b512, <16 x i32> %c512)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; SSE42-LABEL: 'var_funnel_i32'
@@ -150,9 +150,9 @@ define void @var_funnel_i32(i32 %a32, <4 x i32> %a128, <8 x i32> %a256, <16 x i3
define void @var_funnel_i16(i16 %a16, <8 x i16> %a128, <16 x i16> %a256, <32 x i16> %a512, i16 %b16, <8 x i16> %b128, <16 x i16> %b256, <32 x i16> %b512, i16 %c16, <8 x i16> %c128, <16 x i16> %c256, <32 x i16> %c512) {
; SSSE3-LABEL: 'var_funnel_i16'
; SSSE3-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I16 = call i16 @llvm.fshl.i16(i16 %a16, i16 %b16, i16 %c16)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 48 for instruction: %V8I16 = call <8 x i16> @llvm.fshl.v8i16(<8 x i16> %a128, <8 x i16> %b128, <8 x i16> %c128)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 96 for instruction: %V16I16 = call <16 x i16> @llvm.fshl.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> %c256)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 192 for instruction: %V32I16 = call <32 x i16> @llvm.fshl.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> %c512)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 47 for instruction: %V8I16 = call <8 x i16> @llvm.fshl.v8i16(<8 x i16> %a128, <8 x i16> %b128, <8 x i16> %c128)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 94 for instruction: %V16I16 = call <16 x i16> @llvm.fshl.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> %c256)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 188 for instruction: %V32I16 = call <32 x i16> @llvm.fshl.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> %c512)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; SSE42-LABEL: 'var_funnel_i16'
@@ -165,8 +165,8 @@ define void @var_funnel_i16(i16 %a16, <8 x i16> %a128, <16 x i16> %a256, <32 x i
; AVX1-LABEL: 'var_funnel_i16'
; AVX1-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I16 = call i16 @llvm.fshl.i16(i16 %a16, i16 %b16, i16 %c16)
; AVX1-NEXT: Cost Model: Found an estimated cost of 24 for instruction: %V8I16 = call <8 x i16> @llvm.fshl.v8i16(<8 x i16> %a128, <8 x i16> %b128, <8 x i16> %c128)
-; AVX1-NEXT: Cost Model: Found an estimated cost of 54 for instruction: %V16I16 = call <16 x i16> @llvm.fshl.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> %c256)
-; AVX1-NEXT: Cost Model: Found an estimated cost of 108 for instruction: %V32I16 = call <32 x i16> @llvm.fshl.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> %c512)
+; AVX1-NEXT: Cost Model: Found an estimated cost of 53 for instruction: %V16I16 = call <16 x i16> @llvm.fshl.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> %c256)
+; AVX1-NEXT: Cost Model: Found an estimated cost of 106 for instruction: %V32I16 = call <32 x i16> @llvm.fshl.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> %c512)
; AVX1-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; AVX2-LABEL: 'var_funnel_i16'
@@ -214,8 +214,8 @@ define void @var_funnel_i16(i16 %a16, <8 x i16> %a128, <16 x i16> %a256, <32 x i
; XOP-LABEL: 'var_funnel_i16'
; XOP-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I16 = call i16 @llvm.fshl.i16(i16 %a16, i16 %b16, i16 %c16)
; XOP-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V8I16 = call <8 x i16> @llvm.fshl.v8i16(<8 x i16> %a128, <8 x i16> %b128, <8 x i16> %c128)
-; XOP-NEXT: Cost Model: Found an estimated cost of 23 for instruction: %V16I16 = call <16 x i16> @llvm.fshl.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> %c256)
-; XOP-NEXT: Cost Model: Found an estimated cost of 46 for instruction: %V32I16 = call <32 x i16> @llvm.fshl.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> %c512)
+; XOP-NEXT: Cost Model: Found an estimated cost of 22 for instruction: %V16I16 = call <16 x i16> @llvm.fshl.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> %c256)
+; XOP-NEXT: Cost Model: Found an estimated cost of 44 for instruction: %V32I16 = call <32 x i16> @llvm.fshl.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> %c512)
; XOP-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
%I16 = call i16 @llvm.fshl.i16(i16 %a16, i16 %b16, i16 %c16)
@@ -228,9 +228,9 @@ define void @var_funnel_i16(i16 %a16, <8 x i16> %a128, <16 x i16> %a256, <32 x i
define void @var_funnel_i8(i8 %a8, <16 x i8> %a128, <32 x i8> %a256, <64 x i8> %a512, i8 %b8, <16 x i8> %b128, <32 x i8> %b256, <64 x i8> %b512, i8 %c8, <16 x i8> %c128, <32 x i8> %c256, <64 x i8> %c512) {
; SSSE3-LABEL: 'var_funnel_i8'
; SSSE3-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I8 = call i8 @llvm.fshl.i8(i8 %a8, i8 %b8, i8 %c8)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 34 for instruction: %V16I8 = call <16 x i8> @llvm.fshl.v16i8(<16 x i8> %a128, <16 x i8> %b128, <16 x i8> %c128)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 68 for instruction: %V32I8 = call <32 x i8> @llvm.fshl.v32i8(<32 x i8> %a256, <32 x i8> %b256, <32 x i8> %c256)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 136 for instruction: %V64I8 = call <64 x i8> @llvm.fshl.v64i8(<64 x i8> %a512, <64 x i8> %b512, <64 x i8> %c512)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 33 for instruction: %V16I8 = call <16 x i8> @llvm.fshl.v16i8(<16 x i8> %a128, <16 x i8> %b128, <16 x i8> %c128)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 66 for instruction: %V32I8 = call <32 x i8> @llvm.fshl.v32i8(<32 x i8> %a256, <32 x i8> %b256, <32 x i8> %c256)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 132 for instruction: %V64I8 = call <64 x i8> @llvm.fshl.v64i8(<64 x i8> %a512, <64 x i8> %b512, <64 x i8> %c512)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; SSE42-LABEL: 'var_funnel_i8'
@@ -243,8 +243,8 @@ define void @var_funnel_i8(i8 %a8, <16 x i8> %a128, <32 x i8> %a256, <64 x i8> %
; AVX1-LABEL: 'var_funnel_i8'
; AVX1-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I8 = call i8 @llvm.fshl.i8(i8 %a8, i8 %b8, i8 %c8)
; AVX1-NEXT: Cost Model: Found an estimated cost of 26 for instruction: %V16I8 = call <16 x i8> @llvm.fshl.v16i8(<16 x i8> %a128, <16 x i8> %b128, <16 x i8> %c128)
-; AVX1-NEXT: Cost Model: Found an estimated cost of 58 for instruction: %V32I8 = call <32 x i8> @llvm.fshl.v32i8(<32 x i8> %a256, <32 x i8> %b256, <32 x i8> %c256)
-; AVX1-NEXT: Cost Model: Found an estimated cost of 116 for instruction: %V64I8 = call <64 x i8> @llvm.fshl.v64i8(<64 x i8> %a512, <64 x i8> %b512, <64 x i8> %c512)
+; AVX1-NEXT: Cost Model: Found an estimated cost of 57 for instruction: %V32I8 = call <32 x i8> @llvm.fshl.v32i8(<32 x i8> %a256, <32 x i8> %b256, <32 x i8> %c256)
+; AVX1-NEXT: Cost Model: Found an estimated cost of 114 for instruction: %V64I8 = call <64 x i8> @llvm.fshl.v64i8(<64 x i8> %a512, <64 x i8> %b512, <64 x i8> %c512)
; AVX1-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; AVX2-LABEL: 'var_funnel_i8'
@@ -292,8 +292,8 @@ define void @var_funnel_i8(i8 %a8, <16 x i8> %a128, <32 x i8> %a256, <64 x i8> %
; XOP-LABEL: 'var_funnel_i8'
; XOP-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I8 = call i8 @llvm.fshl.i8(i8 %a8, i8 %b8, i8 %c8)
; XOP-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V16I8 = call <16 x i8> @llvm.fshl.v16i8(<16 x i8> %a128, <16 x i8> %b128, <16 x i8> %c128)
-; XOP-NEXT: Cost Model: Found an estimated cost of 23 for instruction: %V32I8 = call <32 x i8> @llvm.fshl.v32i8(<32 x i8> %a256, <32 x i8> %b256, <32 x i8> %c256)
-; XOP-NEXT: Cost Model: Found an estimated cost of 46 for instruction: %V64I8 = call <64 x i8> @llvm.fshl.v64i8(<64 x i8> %a512, <64 x i8> %b512, <64 x i8> %c512)
+; XOP-NEXT: Cost Model: Found an estimated cost of 22 for instruction: %V32I8 = call <32 x i8> @llvm.fshl.v32i8(<32 x i8> %a256, <32 x i8> %b256, <32 x i8> %c256)
+; XOP-NEXT: Cost Model: Found an estimated cost of 44 for instruction: %V64I8 = call <64 x i8> @llvm.fshl.v64i8(<64 x i8> %a512, <64 x i8> %b512, <64 x i8> %c512)
; XOP-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
%I8 = call i8 @llvm.fshl.i8(i8 %a8, i8 %b8, i8 %c8)
@@ -312,9 +312,9 @@ define void @splatvar_funnel_i64(i64 %a64, <2 x i64> %a128, <4 x i64> %a256, <8
; SSSE3-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %u128 = shufflevector <2 x i64> %c128, <2 x i64> undef, <2 x i32> zeroinitializer
; SSSE3-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %u256 = shufflevector <4 x i64> %c256, <4 x i64> undef, <4 x i32> zeroinitializer
; SSSE3-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %u512 = shufflevector <8 x i64> %c512, <8 x i64> undef, <8 x i32> zeroinitializer
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 13 for instruction: %V2I64 = call <2 x i64> @llvm.fshl.v2i64(<2 x i64> %a128, <2 x i64> %b128, <2 x i64> %u128)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 26 for instruction: %V4I64 = call <4 x i64> @llvm.fshl.v4i64(<4 x i64> %a256, <4 x i64> %b256, <4 x i64> %u256)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 52 for instruction: %V8I64 = call <8 x i64> @llvm.fshl.v8i64(<8 x i64> %a512, <8 x i64> %b512, <8 x i64> %u512)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %V2I64 = call <2 x i64> @llvm.fshl.v2i64(<2 x i64> %a128, <2 x i64> %b128, <2 x i64> %u128)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 24 for instruction: %V4I64 = call <4 x i64> @llvm.fshl.v4i64(<4 x i64> %a256, <4 x i64> %b256, <4 x i64> %u256)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 48 for instruction: %V8I64 = call <8 x i64> @llvm.fshl.v8i64(<8 x i64> %a512, <8 x i64> %b512, <8 x i64> %u512)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; SSE42-LABEL: 'splatvar_funnel_i64'
@@ -394,9 +394,9 @@ define void @splatvar_funnel_i32(i32 %a32, <4 x i32> %a128, <8 x i32> %a256, <16
; SSSE3-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %u128 = shufflevector <4 x i32> %c128, <4 x i32> undef, <4 x i32> zeroinitializer
; SSSE3-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %u256 = shufflevector <8 x i32> %c256, <8 x i32> undef, <8 x i32> zeroinitializer
; SSSE3-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %u512 = shufflevector <16 x i32> %c512, <16 x i32> undef, <16 x i32> zeroinitializer
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V2I32 = call <4 x i32> @llvm.fshl.v4i32(<4 x i32> %a128, <4 x i32> %b128, <4 x i32> %u128)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 18 for instruction: %V4I32 = call <8 x i32> @llvm.fshl.v8i32(<8 x i32> %a256, <8 x i32> %b256, <8 x i32> %u256)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 36 for instruction: %V8I32 = call <16 x i32> @llvm.fshl.v16i32(<16 x i32> %a512, <16 x i32> %b512, <16 x i32> %u512)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V2I32 = call <4 x i32> @llvm.fshl.v4i32(<4 x i32> %a128, <4 x i32> %b128, <4 x i32> %u128)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V4I32 = call <8 x i32> @llvm.fshl.v8i32(<8 x i32> %a256, <8 x i32> %b256, <8 x i32> %u256)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V8I32 = call <16 x i32> @llvm.fshl.v16i32(<16 x i32> %a512, <16 x i32> %b512, <16 x i32> %u512)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; SSE42-LABEL: 'splatvar_funnel_i32'
@@ -476,9 +476,9 @@ define void @splatvar_funnel_i16(i16 %a16, <8 x i16> %a128, <16 x i16> %a256, <3
; SSSE3-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %u128 = shufflevector <8 x i16> %c128, <8 x i16> undef, <8 x i32> zeroinitializer
; SSSE3-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %u256 = shufflevector <16 x i16> %c256, <16 x i16> undef, <16 x i32> zeroinitializer
; SSSE3-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %u512 = shufflevector <32 x i16> %c512, <32 x i16> undef, <32 x i32> zeroinitializer
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V8I16 = call <8 x i16> @llvm.fshl.v8i16(<8 x i16> %a128, <8 x i16> %b128, <8 x i16> %u128)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 18 for instruction: %V16I16 = call <16 x i16> @llvm.fshl.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> %u256)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 36 for instruction: %V32I16 = call <32 x i16> @llvm.fshl.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> %u512)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V8I16 = call <8 x i16> @llvm.fshl.v8i16(<8 x i16> %a128, <8 x i16> %b128, <8 x i16> %u128)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V16I16 = call <16 x i16> @llvm.fshl.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> %u256)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V32I16 = call <32 x i16> @llvm.fshl.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> %u512)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; SSE42-LABEL: 'splatvar_funnel_i16'
@@ -495,8 +495,8 @@ define void @splatvar_funnel_i16(i16 %a16, <8 x i16> %a128, <16 x i16> %a256, <3
; AVX1-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %u256 = shufflevector <16 x i16> %c256, <16 x i16> undef, <16 x i32> zeroinitializer
; AVX1-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %u512 = shufflevector <32 x i16> %c512, <32 x i16> undef, <32 x i32> zeroinitializer
; AVX1-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V8I16 = call <8 x i16> @llvm.fshl.v8i16(<8 x i16> %a128, <8 x i16> %b128, <8 x i16> %u128)
-; AVX1-NEXT: Cost Model: Found an estimated cost of 21 for instruction: %V16I16 = call <16 x i16> @llvm.fshl.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> %u256)
-; AVX1-NEXT: Cost Model: Found an estimated cost of 42 for instruction: %V32I16 = call <32 x i16> @llvm.fshl.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> %u512)
+; AVX1-NEXT: Cost Model: Found an estimated cost of 20 for instruction: %V16I16 = call <16 x i16> @llvm.fshl.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> %u256)
+; AVX1-NEXT: Cost Model: Found an estimated cost of 40 for instruction: %V32I16 = call <32 x i16> @llvm.fshl.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> %u512)
; AVX1-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; AVX2-LABEL: 'splatvar_funnel_i16'
@@ -558,8 +558,8 @@ define void @splatvar_funnel_i16(i16 %a16, <8 x i16> %a128, <16 x i16> %a256, <3
; XOP-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %u256 = shufflevector <16 x i16> %c256, <16 x i16> undef, <16 x i32> zeroinitializer
; XOP-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %u512 = shufflevector <32 x i16> %c512, <32 x i16> undef, <32 x i32> zeroinitializer
; XOP-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V8I16 = call <8 x i16> @llvm.fshl.v8i16(<8 x i16> %a128, <8 x i16> %b128, <8 x i16> %u128)
-; XOP-NEXT: Cost Model: Found an estimated cost of 23 for instruction: %V16I16 = call <16 x i16> @llvm.fshl.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> %u256)
-; XOP-NEXT: Cost Model: Found an estimated cost of 46 for instruction: %V32I16 = call <32 x i16> @llvm.fshl.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> %u512)
+; XOP-NEXT: Cost Model: Found an estimated cost of 22 for instruction: %V16I16 = call <16 x i16> @llvm.fshl.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> %u256)
+; XOP-NEXT: Cost Model: Found an estimated cost of 44 for instruction: %V32I16 = call <32 x i16> @llvm.fshl.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> %u512)
; XOP-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
%u128 = shufflevector <8 x i16> %c128, <8 x i16> undef, <8 x i32> zeroinitializer
@@ -576,9 +576,9 @@ define void @splatvar_funnel_i8(i8 %a8, <16 x i8> %a128, <32 x i8> %a256, <64 x
; SSSE3-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %u128 = shufflevector <16 x i8> %c128, <16 x i8> undef, <16 x i32> zeroinitializer
; SSSE3-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %u256 = shufflevector <32 x i8> %c256, <32 x i8> undef, <32 x i32> zeroinitializer
; SSSE3-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %u512 = shufflevector <64 x i8> %c512, <64 x i8> undef, <64 x i32> zeroinitializer
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 34 for instruction: %V16I8 = call <16 x i8> @llvm.fshl.v16i8(<16 x i8> %a128, <16 x i8> %b128, <16 x i8> %u128)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 68 for instruction: %V32I8 = call <32 x i8> @llvm.fshl.v32i8(<32 x i8> %a256, <32 x i8> %b256, <32 x i8> %u256)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 136 for instruction: %V64I8 = call <64 x i8> @llvm.fshl.v64i8(<64 x i8> %a512, <64 x i8> %b512, <64 x i8> %u512)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 33 for instruction: %V16I8 = call <16 x i8> @llvm.fshl.v16i8(<16 x i8> %a128, <16 x i8> %b128, <16 x i8> %u128)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 66 for instruction: %V32I8 = call <32 x i8> @llvm.fshl.v32i8(<32 x i8> %a256, <32 x i8> %b256, <32 x i8> %u256)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 132 for instruction: %V64I8 = call <64 x i8> @llvm.fshl.v64i8(<64 x i8> %a512, <64 x i8> %b512, <64 x i8> %u512)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; SSE42-LABEL: 'splatvar_funnel_i8'
@@ -595,8 +595,8 @@ define void @splatvar_funnel_i8(i8 %a8, <16 x i8> %a128, <32 x i8> %a256, <64 x
; AVX1-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %u256 = shufflevector <32 x i8> %c256, <32 x i8> undef, <32 x i32> zeroinitializer
; AVX1-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %u512 = shufflevector <64 x i8> %c512, <64 x i8> undef, <64 x i32> zeroinitializer
; AVX1-NEXT: Cost Model: Found an estimated cost of 26 for instruction: %V16I8 = call <16 x i8> @llvm.fshl.v16i8(<16 x i8> %a128, <16 x i8> %b128, <16 x i8> %u128)
-; AVX1-NEXT: Cost Model: Found an estimated cost of 58 for instruction: %V32I8 = call <32 x i8> @llvm.fshl.v32i8(<32 x i8> %a256, <32 x i8> %b256, <32 x i8> %u256)
-; AVX1-NEXT: Cost Model: Found an estimated cost of 116 for instruction: %V64I8 = call <64 x i8> @llvm.fshl.v64i8(<64 x i8> %a512, <64 x i8> %b512, <64 x i8> %u512)
+; AVX1-NEXT: Cost Model: Found an estimated cost of 57 for instruction: %V32I8 = call <32 x i8> @llvm.fshl.v32i8(<32 x i8> %a256, <32 x i8> %b256, <32 x i8> %u256)
+; AVX1-NEXT: Cost Model: Found an estimated cost of 114 for instruction: %V64I8 = call <64 x i8> @llvm.fshl.v64i8(<64 x i8> %a512, <64 x i8> %b512, <64 x i8> %u512)
; AVX1-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; AVX2-LABEL: 'splatvar_funnel_i8'
@@ -658,8 +658,8 @@ define void @splatvar_funnel_i8(i8 %a8, <16 x i8> %a128, <32 x i8> %a256, <64 x
; XOP-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %u256 = shufflevector <32 x i8> %c256, <32 x i8> undef, <32 x i32> zeroinitializer
; XOP-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %u512 = shufflevector <64 x i8> %c512, <64 x i8> undef, <64 x i32> zeroinitializer
; XOP-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V16I8 = call <16 x i8> @llvm.fshl.v16i8(<16 x i8> %a128, <16 x i8> %b128, <16 x i8> %u128)
-; XOP-NEXT: Cost Model: Found an estimated cost of 23 for instruction: %V32I8 = call <32 x i8> @llvm.fshl.v32i8(<32 x i8> %a256, <32 x i8> %b256, <32 x i8> %u256)
-; XOP-NEXT: Cost Model: Found an estimated cost of 46 for instruction: %V64I8 = call <64 x i8> @llvm.fshl.v64i8(<64 x i8> %a512, <64 x i8> %b512, <64 x i8> %u512)
+; XOP-NEXT: Cost Model: Found an estimated cost of 22 for instruction: %V32I8 = call <32 x i8> @llvm.fshl.v32i8(<32 x i8> %a256, <32 x i8> %b256, <32 x i8> %u256)
+; XOP-NEXT: Cost Model: Found an estimated cost of 44 for instruction: %V64I8 = call <64 x i8> @llvm.fshl.v64i8(<64 x i8> %a512, <64 x i8> %b512, <64 x i8> %u512)
; XOP-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
%u128 = shufflevector <16 x i8> %c128, <16 x i8> undef, <16 x i32> zeroinitializer
@@ -678,9 +678,9 @@ define void @splatvar_funnel_i8(i8 %a8, <16 x i8> %a128, <32 x i8> %a256, <64 x
define void @constant_funnel_i64(i64 %a64, <2 x i64> %a128, <4 x i64> %a256, <8 x i64> %a512, i64 %b64, <2 x i64> %b128, <4 x i64> %b256, <8 x i64> %b512) {
; SSSE3-LABEL: 'constant_funnel_i64'
; SSSE3-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I64 = call i64 @llvm.fshl.i64(i64 %a64, i64 %b64, i64 7)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 18 for instruction: %V2I64 = call <2 x i64> @llvm.fshl.v2i64(<2 x i64> %a128, <2 x i64> %b128, <2 x i64> <i64 1, i64 7>)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 36 for instruction: %V4I64 = call <4 x i64> @llvm.fshl.v4i64(<4 x i64> %a256, <4 x i64> %b256, <4 x i64> <i64 1, i64 7, i64 15, i64 31>)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 72 for instruction: %V8I64 = call <8 x i64> @llvm.fshl.v8i64(<8 x i64> %a512, <8 x i64> %b512, <8 x i64> <i64 1, i64 7, i64 15, i64 31, i64 1, i64 7, i64 15, i64 31>)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 17 for instruction: %V2I64 = call <2 x i64> @llvm.fshl.v2i64(<2 x i64> %a128, <2 x i64> %b128, <2 x i64> <i64 1, i64 7>)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 34 for instruction: %V4I64 = call <4 x i64> @llvm.fshl.v4i64(<4 x i64> %a256, <4 x i64> %b256, <4 x i64> <i64 1, i64 7, i64 15, i64 31>)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 68 for instruction: %V8I64 = call <8 x i64> @llvm.fshl.v8i64(<8 x i64> %a512, <8 x i64> %b512, <8 x i64> <i64 1, i64 7, i64 15, i64 31, i64 1, i64 7, i64 15, i64 31>)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; SSE42-LABEL: 'constant_funnel_i64'
@@ -742,9 +742,9 @@ define void @constant_funnel_i64(i64 %a64, <2 x i64> %a128, <4 x i64> %a256, <8
define void @constant_funnel_i32(i32 %a32, <4 x i32> %a128, <8 x i32> %a256, <16 x i32> %a512, i32 %b32, <4 x i32> %b128, <8 x i32> %b256, <16 x i32> %b512) {
; SSSE3-LABEL: 'constant_funnel_i32'
; SSSE3-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I32 = call i32 @llvm.fshl.i32(i32 %a32, i32 %b32, i32 7)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 24 for instruction: %V2I32 = call <4 x i32> @llvm.fshl.v4i32(<4 x i32> %a128, <4 x i32> %b128, <4 x i32> <i32 4, i32 5, i32 6, i32 7>)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 48 for instruction: %V4I32 = call <8 x i32> @llvm.fshl.v8i32(<8 x i32> %a256, <8 x i32> %b256, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 96 for instruction: %V8I32 = call <16 x i32> @llvm.fshl.v16i32(<16 x i32> %a512, <16 x i32> %b512, <16 x i32> <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 23 for instruction: %V2I32 = call <4 x i32> @llvm.fshl.v4i32(<4 x i32> %a128, <4 x i32> %b128, <4 x i32> <i32 4, i32 5, i32 6, i32 7>)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 46 for instruction: %V4I32 = call <8 x i32> @llvm.fshl.v8i32(<8 x i32> %a256, <8 x i32> %b256, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 92 for instruction: %V8I32 = call <16 x i32> @llvm.fshl.v16i32(<16 x i32> %a512, <16 x i32> %b512, <16 x i32> <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; SSE42-LABEL: 'constant_funnel_i32'
@@ -806,9 +806,9 @@ define void @constant_funnel_i32(i32 %a32, <4 x i32> %a128, <8 x i32> %a256, <16
define void @constant_funnel_i16(i16 %a16, <8 x i16> %a128, <16 x i16> %a256, <32 x i16> %a512, i16 %b16, <8 x i16> %b128, <16 x i16> %b256, <32 x i16> %b512) {
; SSSE3-LABEL: 'constant_funnel_i16'
; SSSE3-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I16 = call i16 @llvm.fshl.i16(i16 %a16, i16 %b16, i16 7)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 23 for instruction: %V8I16 = call <8 x i16> @llvm.fshl.v8i16(<8 x i16> %a128, <8 x i16> %b128, <8 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 46 for instruction: %V16I16 = call <16 x i16> @llvm.fshl.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 92 for instruction: %V32I16 = call <32 x i16> @llvm.fshl.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 22 for instruction: %V8I16 = call <8 x i16> @llvm.fshl.v8i16(<8 x i16> %a128, <8 x i16> %b128, <8 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 44 for instruction: %V16I16 = call <16 x i16> @llvm.fshl.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 88 for instruction: %V32I16 = call <32 x i16> @llvm.fshl.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; SSE42-LABEL: 'constant_funnel_i16'
@@ -821,8 +821,8 @@ define void @constant_funnel_i16(i16 %a16, <8 x i16> %a128, <16 x i16> %a256, <3
; AVX1-LABEL: 'constant_funnel_i16'
; AVX1-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I16 = call i16 @llvm.fshl.i16(i16 %a16, i16 %b16, i16 7)
; AVX1-NEXT: Cost Model: Found an estimated cost of 18 for instruction: %V8I16 = call <8 x i16> @llvm.fshl.v8i16(<8 x i16> %a128, <8 x i16> %b128, <8 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
-; AVX1-NEXT: Cost Model: Found an estimated cost of 44 for instruction: %V16I16 = call <16 x i16> @llvm.fshl.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
-; AVX1-NEXT: Cost Model: Found an estimated cost of 88 for instruction: %V32I16 = call <32 x i16> @llvm.fshl.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
+; AVX1-NEXT: Cost Model: Found an estimated cost of 43 for instruction: %V16I16 = call <16 x i16> @llvm.fshl.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
+; AVX1-NEXT: Cost Model: Found an estimated cost of 86 for instruction: %V32I16 = call <32 x i16> @llvm.fshl.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
; AVX1-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; AVX2-LABEL: 'constant_funnel_i16'
@@ -870,8 +870,8 @@ define void @constant_funnel_i16(i16 %a16, <8 x i16> %a128, <16 x i16> %a256, <3
; XOP-LABEL: 'constant_funnel_i16'
; XOP-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I16 = call i16 @llvm.fshl.i16(i16 %a16, i16 %b16, i16 7)
; XOP-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V8I16 = call <8 x i16> @llvm.fshl.v8i16(<8 x i16> %a128, <8 x i16> %b128, <8 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
-; XOP-NEXT: Cost Model: Found an estimated cost of 20 for instruction: %V16I16 = call <16 x i16> @llvm.fshl.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
-; XOP-NEXT: Cost Model: Found an estimated cost of 40 for instruction: %V32I16 = call <32 x i16> @llvm.fshl.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
+; XOP-NEXT: Cost Model: Found an estimated cost of 19 for instruction: %V16I16 = call <16 x i16> @llvm.fshl.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
+; XOP-NEXT: Cost Model: Found an estimated cost of 38 for instruction: %V32I16 = call <32 x i16> @llvm.fshl.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
; XOP-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
%I16 = call i16 @llvm.fshl.i16(i16 %a16, i16 %b16, i16 7)
@@ -884,9 +884,9 @@ define void @constant_funnel_i16(i16 %a16, <8 x i16> %a128, <16 x i16> %a256, <3
define void @constant_funnel_i8(i8 %a8, <16 x i8> %a128, <32 x i8> %a256, <64 x i8> %a512, i8 %b8, <16 x i8> %b128, <32 x i8> %b256, <64 x i8> %b512) {
; SSSE3-LABEL: 'constant_funnel_i8'
; SSSE3-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I8 = call i8 @llvm.fshl.i8(i8 %a8, i8 %b8, i8 7)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 33 for instruction: %V16I8 = call <16 x i8> @llvm.fshl.v16i8(<16 x i8> %a128, <16 x i8> %b128, <16 x i8> <i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0>)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 66 for instruction: %V32I8 = call <32 x i8> @llvm.fshl.v32i8(<32 x i8> %a256, <32 x i8> %b256, <32 x i8> <i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0>)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 132 for instruction: %V64I8 = call <64 x i8> @llvm.fshl.v64i8(<64 x i8> %a512, <64 x i8> %b512, <64 x i8> <i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0>)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V16I8 = call <16 x i8> @llvm.fshl.v16i8(<16 x i8> %a128, <16 x i8> %b128, <16 x i8> <i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0>)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 64 for instruction: %V32I8 = call <32 x i8> @llvm.fshl.v32i8(<32 x i8> %a256, <32 x i8> %b256, <32 x i8> <i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0>)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 128 for instruction: %V64I8 = call <64 x i8> @llvm.fshl.v64i8(<64 x i8> %a512, <64 x i8> %b512, <64 x i8> <i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0>)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; SSE42-LABEL: 'constant_funnel_i8'
@@ -899,8 +899,8 @@ define void @constant_funnel_i8(i8 %a8, <16 x i8> %a128, <32 x i8> %a256, <64 x
; AVX1-LABEL: 'constant_funnel_i8'
; AVX1-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I8 = call i8 @llvm.fshl.i8(i8 %a8, i8 %b8, i8 7)
; AVX1-NEXT: Cost Model: Found an estimated cost of 25 for instruction: %V16I8 = call <16 x i8> @llvm.fshl.v16i8(<16 x i8> %a128, <16 x i8> %b128, <16 x i8> <i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0>)
-; AVX1-NEXT: Cost Model: Found an estimated cost of 57 for instruction: %V32I8 = call <32 x i8> @llvm.fshl.v32i8(<32 x i8> %a256, <32 x i8> %b256, <32 x i8> <i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0>)
-; AVX1-NEXT: Cost Model: Found an estimated cost of 114 for instruction: %V64I8 = call <64 x i8> @llvm.fshl.v64i8(<64 x i8> %a512, <64 x i8> %b512, <64 x i8> <i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0>)
+; AVX1-NEXT: Cost Model: Found an estimated cost of 56 for instruction: %V32I8 = call <32 x i8> @llvm.fshl.v32i8(<32 x i8> %a256, <32 x i8> %b256, <32 x i8> <i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0>)
+; AVX1-NEXT: Cost Model: Found an estimated cost of 112 for instruction: %V64I8 = call <64 x i8> @llvm.fshl.v64i8(<64 x i8> %a512, <64 x i8> %b512, <64 x i8> <i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0>)
; AVX1-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; AVX2-LABEL: 'constant_funnel_i8'
@@ -948,8 +948,8 @@ define void @constant_funnel_i8(i8 %a8, <16 x i8> %a128, <32 x i8> %a256, <64 x
; XOP-LABEL: 'constant_funnel_i8'
; XOP-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I8 = call i8 @llvm.fshl.i8(i8 %a8, i8 %b8, i8 7)
; XOP-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V16I8 = call <16 x i8> @llvm.fshl.v16i8(<16 x i8> %a128, <16 x i8> %b128, <16 x i8> <i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0>)
-; XOP-NEXT: Cost Model: Found an estimated cost of 20 for instruction: %V32I8 = call <32 x i8> @llvm.fshl.v32i8(<32 x i8> %a256, <32 x i8> %b256, <32 x i8> <i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0>)
-; XOP-NEXT: Cost Model: Found an estimated cost of 40 for instruction: %V64I8 = call <64 x i8> @llvm.fshl.v64i8(<64 x i8> %a512, <64 x i8> %b512, <64 x i8> <i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0>)
+; XOP-NEXT: Cost Model: Found an estimated cost of 19 for instruction: %V32I8 = call <32 x i8> @llvm.fshl.v32i8(<32 x i8> %a256, <32 x i8> %b256, <32 x i8> <i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0>)
+; XOP-NEXT: Cost Model: Found an estimated cost of 38 for instruction: %V64I8 = call <64 x i8> @llvm.fshl.v64i8(<64 x i8> %a512, <64 x i8> %b512, <64 x i8> <i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0>)
; XOP-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
%I8 = call i8 @llvm.fshl.i8(i8 %a8, i8 %b8, i8 7)
@@ -966,9 +966,9 @@ define void @constant_funnel_i8(i8 %a8, <16 x i8> %a128, <32 x i8> %a256, <64 x
define void @splatconstant_funnel_i64(i64 %a64, <2 x i64> %a128, <4 x i64> %a256, <8 x i64> %a512, i64 %b64, <2 x i64> %b128, <4 x i64> %b256, <8 x i64> %b512) {
; SSSE3-LABEL: 'splatconstant_funnel_i64'
; SSSE3-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I64 = call i64 @llvm.fshl.i64(i64 %a64, i64 %b64, i64 7)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %V2I64 = call <2 x i64> @llvm.fshl.v2i64(<2 x i64> %a128, <2 x i64> %b128, <2 x i64> <i64 7, i64 7>)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 24 for instruction: %V4I64 = call <4 x i64> @llvm.fshl.v4i64(<4 x i64> %a256, <4 x i64> %b256, <4 x i64> <i64 7, i64 7, i64 7, i64 7>)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 48 for instruction: %V8I64 = call <8 x i64> @llvm.fshl.v8i64(<8 x i64> %a512, <8 x i64> %b512, <8 x i64> <i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7>)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 11 for instruction: %V2I64 = call <2 x i64> @llvm.fshl.v2i64(<2 x i64> %a128, <2 x i64> %b128, <2 x i64> <i64 7, i64 7>)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 22 for instruction: %V4I64 = call <4 x i64> @llvm.fshl.v4i64(<4 x i64> %a256, <4 x i64> %b256, <4 x i64> <i64 7, i64 7, i64 7, i64 7>)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 44 for instruction: %V8I64 = call <8 x i64> @llvm.fshl.v8i64(<8 x i64> %a512, <8 x i64> %b512, <8 x i64> <i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7>)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; SSE42-LABEL: 'splatconstant_funnel_i64'
@@ -1030,9 +1030,9 @@ define void @splatconstant_funnel_i64(i64 %a64, <2 x i64> %a128, <4 x i64> %a256
define void @splatconstant_funnel_i32(i32 %a32, <4 x i32> %a128, <8 x i32> %a256, <16 x i32> %a512, i32 %b32, <4 x i32> %b128, <8 x i32> %b256, <16 x i32> %b512) {
; SSSE3-LABEL: 'splatconstant_funnel_i32'
; SSSE3-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I32 = call i32 @llvm.fshl.i32(i32 %a32, i32 %b32, i32 5)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V2I32 = call <4 x i32> @llvm.fshl.v4i32(<4 x i32> %a128, <4 x i32> %b128, <4 x i32> <i32 5, i32 5, i32 5, i32 5>)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V4I32 = call <8 x i32> @llvm.fshl.v8i32(<8 x i32> %a256, <8 x i32> %b256, <8 x i32> <i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5>)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V8I32 = call <16 x i32> @llvm.fshl.v16i32(<16 x i32> %a512, <16 x i32> %b512, <16 x i32> <i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5>)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V2I32 = call <4 x i32> @llvm.fshl.v4i32(<4 x i32> %a128, <4 x i32> %b128, <4 x i32> <i32 5, i32 5, i32 5, i32 5>)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 14 for instruction: %V4I32 = call <8 x i32> @llvm.fshl.v8i32(<8 x i32> %a256, <8 x i32> %b256, <8 x i32> <i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5>)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 28 for instruction: %V8I32 = call <16 x i32> @llvm.fshl.v16i32(<16 x i32> %a512, <16 x i32> %b512, <16 x i32> <i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5>)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; SSE42-LABEL: 'splatconstant_funnel_i32'
@@ -1094,9 +1094,9 @@ define void @splatconstant_funnel_i32(i32 %a32, <4 x i32> %a128, <8 x i32> %a256
define void @splatconstant_funnel_i16(i16 %a16, <8 x i16> %a128, <16 x i16> %a256, <32 x i16> %a512, i16 %b16, <8 x i16> %b128, <16 x i16> %b256, <32 x i16> %b512) {
; SSSE3-LABEL: 'splatconstant_funnel_i16'
; SSSE3-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I16 = call i16 @llvm.fshl.i16(i16 %a16, i16 %b16, i16 3)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V8I16 = call <8 x i16> @llvm.fshl.v8i16(<8 x i16> %a128, <8 x i16> %b128, <8 x i16> <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V16I16 = call <16 x i16> @llvm.fshl.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V32I16 = call <32 x i16> @llvm.fshl.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V8I16 = call <8 x i16> @llvm.fshl.v8i16(<8 x i16> %a128, <8 x i16> %b128, <8 x i16> <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 14 for instruction: %V16I16 = call <16 x i16> @llvm.fshl.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 28 for instruction: %V32I16 = call <32 x i16> @llvm.fshl.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; SSE42-LABEL: 'splatconstant_funnel_i16'
@@ -1109,8 +1109,8 @@ define void @splatconstant_funnel_i16(i16 %a16, <8 x i16> %a128, <16 x i16> %a25
; AVX1-LABEL: 'splatconstant_funnel_i16'
; AVX1-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I16 = call i16 @llvm.fshl.i16(i16 %a16, i16 %b16, i16 3)
; AVX1-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V8I16 = call <8 x i16> @llvm.fshl.v8i16(<8 x i16> %a128, <8 x i16> %b128, <8 x i16> <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>)
-; AVX1-NEXT: Cost Model: Found an estimated cost of 20 for instruction: %V16I16 = call <16 x i16> @llvm.fshl.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>)
-; AVX1-NEXT: Cost Model: Found an estimated cost of 40 for instruction: %V32I16 = call <32 x i16> @llvm.fshl.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>)
+; AVX1-NEXT: Cost Model: Found an estimated cost of 19 for instruction: %V16I16 = call <16 x i16> @llvm.fshl.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>)
+; AVX1-NEXT: Cost Model: Found an estimated cost of 38 for instruction: %V32I16 = call <32 x i16> @llvm.fshl.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>)
; AVX1-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; AVX2-LABEL: 'splatconstant_funnel_i16'
@@ -1158,8 +1158,8 @@ define void @splatconstant_funnel_i16(i16 %a16, <8 x i16> %a128, <16 x i16> %a25
; XOP-LABEL: 'splatconstant_funnel_i16'
; XOP-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I16 = call i16 @llvm.fshl.i16(i16 %a16, i16 %b16, i16 3)
; XOP-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V8I16 = call <8 x i16> @llvm.fshl.v8i16(<8 x i16> %a128, <8 x i16> %b128, <8 x i16> <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>)
-; XOP-NEXT: Cost Model: Found an estimated cost of 20 for instruction: %V16I16 = call <16 x i16> @llvm.fshl.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>)
-; XOP-NEXT: Cost Model: Found an estimated cost of 40 for instruction: %V32I16 = call <32 x i16> @llvm.fshl.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>)
+; XOP-NEXT: Cost Model: Found an estimated cost of 19 for instruction: %V16I16 = call <16 x i16> @llvm.fshl.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>)
+; XOP-NEXT: Cost Model: Found an estimated cost of 38 for instruction: %V32I16 = call <32 x i16> @llvm.fshl.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>)
; XOP-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
%I16 = call i16 @llvm.fshl.i16(i16 %a16, i16 %b16, i16 3)
@@ -1172,9 +1172,9 @@ define void @splatconstant_funnel_i16(i16 %a16, <8 x i16> %a128, <16 x i16> %a25
define void @splatconstant_funnel_i8(i8 %a8, <16 x i8> %a128, <32 x i8> %a256, <64 x i8> %a512, i8 %b8, <16 x i8> %b128, <32 x i8> %b256, <64 x i8> %b512) {
; SSSE3-LABEL: 'splatconstant_funnel_i8'
; SSSE3-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I8 = call i8 @llvm.fshl.i8(i8 %a8, i8 %b8, i8 3)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 10 for instruction: %V16I8 = call <16 x i8> @llvm.fshl.v16i8(<16 x i8> %a128, <16 x i8> %b128, <16 x i8> <i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3>)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 20 for instruction: %V32I8 = call <32 x i8> @llvm.fshl.v32i8(<32 x i8> %a256, <32 x i8> %b256, <32 x i8> <i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3>)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 40 for instruction: %V64I8 = call <64 x i8> @llvm.fshl.v64i8(<64 x i8> %a512, <64 x i8> %b512, <64 x i8> <i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3>)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V16I8 = call <16 x i8> @llvm.fshl.v16i8(<16 x i8> %a128, <16 x i8> %b128, <16 x i8> <i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3>)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 18 for instruction: %V32I8 = call <32 x i8> @llvm.fshl.v32i8(<32 x i8> %a256, <32 x i8> %b256, <32 x i8> <i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3>)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 36 for instruction: %V64I8 = call <64 x i8> @llvm.fshl.v64i8(<64 x i8> %a512, <64 x i8> %b512, <64 x i8> <i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3>)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; SSE42-LABEL: 'splatconstant_funnel_i8'
@@ -1187,8 +1187,8 @@ define void @splatconstant_funnel_i8(i8 %a8, <16 x i8> %a128, <32 x i8> %a256, <
; AVX1-LABEL: 'splatconstant_funnel_i8'
; AVX1-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I8 = call i8 @llvm.fshl.i8(i8 %a8, i8 %b8, i8 3)
; AVX1-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V16I8 = call <16 x i8> @llvm.fshl.v16i8(<16 x i8> %a128, <16 x i8> %b128, <16 x i8> <i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3>)
-; AVX1-NEXT: Cost Model: Found an estimated cost of 24 for instruction: %V32I8 = call <32 x i8> @llvm.fshl.v32i8(<32 x i8> %a256, <32 x i8> %b256, <32 x i8> <i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3>)
-; AVX1-NEXT: Cost Model: Found an estimated cost of 48 for instruction: %V64I8 = call <64 x i8> @llvm.fshl.v64i8(<64 x i8> %a512, <64 x i8> %b512, <64 x i8> <i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3>)
+; AVX1-NEXT: Cost Model: Found an estimated cost of 23 for instruction: %V32I8 = call <32 x i8> @llvm.fshl.v32i8(<32 x i8> %a256, <32 x i8> %b256, <32 x i8> <i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3>)
+; AVX1-NEXT: Cost Model: Found an estimated cost of 46 for instruction: %V64I8 = call <64 x i8> @llvm.fshl.v64i8(<64 x i8> %a512, <64 x i8> %b512, <64 x i8> <i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3>)
; AVX1-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; AVX2-LABEL: 'splatconstant_funnel_i8'
@@ -1236,8 +1236,8 @@ define void @splatconstant_funnel_i8(i8 %a8, <16 x i8> %a128, <32 x i8> %a256, <
; XOP-LABEL: 'splatconstant_funnel_i8'
; XOP-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I8 = call i8 @llvm.fshl.i8(i8 %a8, i8 %b8, i8 3)
; XOP-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V16I8 = call <16 x i8> @llvm.fshl.v16i8(<16 x i8> %a128, <16 x i8> %b128, <16 x i8> <i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3>)
-; XOP-NEXT: Cost Model: Found an estimated cost of 20 for instruction: %V32I8 = call <32 x i8> @llvm.fshl.v32i8(<32 x i8> %a256, <32 x i8> %b256, <32 x i8> <i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3>)
-; XOP-NEXT: Cost Model: Found an estimated cost of 40 for instruction: %V64I8 = call <64 x i8> @llvm.fshl.v64i8(<64 x i8> %a512, <64 x i8> %b512, <64 x i8> <i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3>)
+; XOP-NEXT: Cost Model: Found an estimated cost of 19 for instruction: %V32I8 = call <32 x i8> @llvm.fshl.v32i8(<32 x i8> %a256, <32 x i8> %b256, <32 x i8> <i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3>)
+; XOP-NEXT: Cost Model: Found an estimated cost of 38 for instruction: %V64I8 = call <64 x i8> @llvm.fshl.v64i8(<64 x i8> %a512, <64 x i8> %b512, <64 x i8> <i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3>)
; XOP-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
%I8 = call i8 @llvm.fshl.i8(i8 %a8, i8 %b8, i8 3)
diff --git a/llvm/test/Analysis/CostModel/X86/fshr.ll b/llvm/test/Analysis/CostModel/X86/fshr.ll
index f5d7230e44dbc..78765bfbdd193 100644
--- a/llvm/test/Analysis/CostModel/X86/fshr.ll
+++ b/llvm/test/Analysis/CostModel/X86/fshr.ll
@@ -22,9 +22,9 @@ target triple = "x86_64-apple-macosx10.8.0"
define void @var_funnel_i64(i64 %a64, <2 x i64> %a128, <4 x i64> %a256, <8 x i64> %a512, i64 %b64, <2 x i64> %b128, <4 x i64> %b256, <8 x i64> %b512, i64 %c64, <2 x i64> %c128, <4 x i64> %c256, <8 x i64> %c512) {
; SSSE3-LABEL: 'var_funnel_i64'
; SSSE3-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I64 = call i64 @llvm.fshr.i64(i64 %a64, i64 %b64, i64 %c64)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 19 for instruction: %V2I64 = call <2 x i64> @llvm.fshr.v2i64(<2 x i64> %a128, <2 x i64> %b128, <2 x i64> %c128)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 38 for instruction: %V4I64 = call <4 x i64> @llvm.fshr.v4i64(<4 x i64> %a256, <4 x i64> %b256, <4 x i64> %c256)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 76 for instruction: %V8I64 = call <8 x i64> @llvm.fshr.v8i64(<8 x i64> %a512, <8 x i64> %b512, <8 x i64> %c512)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 18 for instruction: %V2I64 = call <2 x i64> @llvm.fshr.v2i64(<2 x i64> %a128, <2 x i64> %b128, <2 x i64> %c128)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 36 for instruction: %V4I64 = call <4 x i64> @llvm.fshr.v4i64(<4 x i64> %a256, <4 x i64> %b256, <4 x i64> %c256)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 72 for instruction: %V8I64 = call <8 x i64> @llvm.fshr.v8i64(<8 x i64> %a512, <8 x i64> %b512, <8 x i64> %c512)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; SSE42-LABEL: 'var_funnel_i64'
@@ -86,9 +86,9 @@ define void @var_funnel_i64(i64 %a64, <2 x i64> %a128, <4 x i64> %a256, <8 x i64
define void @var_funnel_i32(i32 %a32, <4 x i32> %a128, <8 x i32> %a256, <16 x i32> %a512, i32 %b32, <4 x i32> %b128, <8 x i32> %b256, <16 x i32> %b512, i32 %c32, <4 x i32> %c128, <8 x i32> %c256, <16 x i32> %c512) {
; SSSE3-LABEL: 'var_funnel_i32'
; SSSE3-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I32 = call i32 @llvm.fshr.i32(i32 %a32, i32 %b32, i32 %c32)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 35 for instruction: %V2I32 = call <4 x i32> @llvm.fshr.v4i32(<4 x i32> %a128, <4 x i32> %b128, <4 x i32> %c128)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 70 for instruction: %V4I32 = call <8 x i32> @llvm.fshr.v8i32(<8 x i32> %a256, <8 x i32> %b256, <8 x i32> %c256)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 140 for instruction: %V8I32 = call <16 x i32> @llvm.fshr.v16i32(<16 x i32> %a512, <16 x i32> %b512, <16 x i32> %c512)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 34 for instruction: %V2I32 = call <4 x i32> @llvm.fshr.v4i32(<4 x i32> %a128, <4 x i32> %b128, <4 x i32> %c128)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 68 for instruction: %V4I32 = call <8 x i32> @llvm.fshr.v8i32(<8 x i32> %a256, <8 x i32> %b256, <8 x i32> %c256)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 136 for instruction: %V8I32 = call <16 x i32> @llvm.fshr.v16i32(<16 x i32> %a512, <16 x i32> %b512, <16 x i32> %c512)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; SSE42-LABEL: 'var_funnel_i32'
@@ -150,9 +150,9 @@ define void @var_funnel_i32(i32 %a32, <4 x i32> %a128, <8 x i32> %a256, <16 x i3
define void @var_funnel_i16(i16 %a16, <8 x i16> %a128, <16 x i16> %a256, <32 x i16> %a512, i16 %b16, <8 x i16> %b128, <16 x i16> %b256, <32 x i16> %b512, i16 %c16, <8 x i16> %c128, <16 x i16> %c256, <32 x i16> %c512) {
; SSSE3-LABEL: 'var_funnel_i16'
; SSSE3-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I16 = call i16 @llvm.fshr.i16(i16 %a16, i16 %b16, i16 %c16)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 48 for instruction: %V8I16 = call <8 x i16> @llvm.fshr.v8i16(<8 x i16> %a128, <8 x i16> %b128, <8 x i16> %c128)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 96 for instruction: %V16I16 = call <16 x i16> @llvm.fshr.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> %c256)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 192 for instruction: %V32I16 = call <32 x i16> @llvm.fshr.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> %c512)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 47 for instruction: %V8I16 = call <8 x i16> @llvm.fshr.v8i16(<8 x i16> %a128, <8 x i16> %b128, <8 x i16> %c128)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 94 for instruction: %V16I16 = call <16 x i16> @llvm.fshr.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> %c256)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 188 for instruction: %V32I16 = call <32 x i16> @llvm.fshr.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> %c512)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; SSE42-LABEL: 'var_funnel_i16'
@@ -165,8 +165,8 @@ define void @var_funnel_i16(i16 %a16, <8 x i16> %a128, <16 x i16> %a256, <32 x i
; AVX1-LABEL: 'var_funnel_i16'
; AVX1-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I16 = call i16 @llvm.fshr.i16(i16 %a16, i16 %b16, i16 %c16)
; AVX1-NEXT: Cost Model: Found an estimated cost of 24 for instruction: %V8I16 = call <8 x i16> @llvm.fshr.v8i16(<8 x i16> %a128, <8 x i16> %b128, <8 x i16> %c128)
-; AVX1-NEXT: Cost Model: Found an estimated cost of 54 for instruction: %V16I16 = call <16 x i16> @llvm.fshr.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> %c256)
-; AVX1-NEXT: Cost Model: Found an estimated cost of 108 for instruction: %V32I16 = call <32 x i16> @llvm.fshr.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> %c512)
+; AVX1-NEXT: Cost Model: Found an estimated cost of 53 for instruction: %V16I16 = call <16 x i16> @llvm.fshr.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> %c256)
+; AVX1-NEXT: Cost Model: Found an estimated cost of 106 for instruction: %V32I16 = call <32 x i16> @llvm.fshr.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> %c512)
; AVX1-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; AVX2-LABEL: 'var_funnel_i16'
@@ -214,8 +214,8 @@ define void @var_funnel_i16(i16 %a16, <8 x i16> %a128, <16 x i16> %a256, <32 x i
; XOP-LABEL: 'var_funnel_i16'
; XOP-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I16 = call i16 @llvm.fshr.i16(i16 %a16, i16 %b16, i16 %c16)
; XOP-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V8I16 = call <8 x i16> @llvm.fshr.v8i16(<8 x i16> %a128, <8 x i16> %b128, <8 x i16> %c128)
-; XOP-NEXT: Cost Model: Found an estimated cost of 23 for instruction: %V16I16 = call <16 x i16> @llvm.fshr.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> %c256)
-; XOP-NEXT: Cost Model: Found an estimated cost of 46 for instruction: %V32I16 = call <32 x i16> @llvm.fshr.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> %c512)
+; XOP-NEXT: Cost Model: Found an estimated cost of 22 for instruction: %V16I16 = call <16 x i16> @llvm.fshr.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> %c256)
+; XOP-NEXT: Cost Model: Found an estimated cost of 44 for instruction: %V32I16 = call <32 x i16> @llvm.fshr.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> %c512)
; XOP-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
%I16 = call i16 @llvm.fshr.i16(i16 %a16, i16 %b16, i16 %c16)
@@ -228,9 +228,9 @@ define void @var_funnel_i16(i16 %a16, <8 x i16> %a128, <16 x i16> %a256, <32 x i
define void @var_funnel_i8(i8 %a8, <16 x i8> %a128, <32 x i8> %a256, <64 x i8> %a512, i8 %b8, <16 x i8> %b128, <32 x i8> %b256, <64 x i8> %b512, i8 %c8, <16 x i8> %c128, <32 x i8> %c256, <64 x i8> %c512) {
; SSSE3-LABEL: 'var_funnel_i8'
; SSSE3-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I8 = call i8 @llvm.fshr.i8(i8 %a8, i8 %b8, i8 %c8)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 34 for instruction: %V16I8 = call <16 x i8> @llvm.fshr.v16i8(<16 x i8> %a128, <16 x i8> %b128, <16 x i8> %c128)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 68 for instruction: %V32I8 = call <32 x i8> @llvm.fshr.v32i8(<32 x i8> %a256, <32 x i8> %b256, <32 x i8> %c256)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 136 for instruction: %V64I8 = call <64 x i8> @llvm.fshr.v64i8(<64 x i8> %a512, <64 x i8> %b512, <64 x i8> %c512)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 33 for instruction: %V16I8 = call <16 x i8> @llvm.fshr.v16i8(<16 x i8> %a128, <16 x i8> %b128, <16 x i8> %c128)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 66 for instruction: %V32I8 = call <32 x i8> @llvm.fshr.v32i8(<32 x i8> %a256, <32 x i8> %b256, <32 x i8> %c256)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 132 for instruction: %V64I8 = call <64 x i8> @llvm.fshr.v64i8(<64 x i8> %a512, <64 x i8> %b512, <64 x i8> %c512)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; SSE42-LABEL: 'var_funnel_i8'
@@ -243,8 +243,8 @@ define void @var_funnel_i8(i8 %a8, <16 x i8> %a128, <32 x i8> %a256, <64 x i8> %
; AVX1-LABEL: 'var_funnel_i8'
; AVX1-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I8 = call i8 @llvm.fshr.i8(i8 %a8, i8 %b8, i8 %c8)
; AVX1-NEXT: Cost Model: Found an estimated cost of 26 for instruction: %V16I8 = call <16 x i8> @llvm.fshr.v16i8(<16 x i8> %a128, <16 x i8> %b128, <16 x i8> %c128)
-; AVX1-NEXT: Cost Model: Found an estimated cost of 58 for instruction: %V32I8 = call <32 x i8> @llvm.fshr.v32i8(<32 x i8> %a256, <32 x i8> %b256, <32 x i8> %c256)
-; AVX1-NEXT: Cost Model: Found an estimated cost of 116 for instruction: %V64I8 = call <64 x i8> @llvm.fshr.v64i8(<64 x i8> %a512, <64 x i8> %b512, <64 x i8> %c512)
+; AVX1-NEXT: Cost Model: Found an estimated cost of 57 for instruction: %V32I8 = call <32 x i8> @llvm.fshr.v32i8(<32 x i8> %a256, <32 x i8> %b256, <32 x i8> %c256)
+; AVX1-NEXT: Cost Model: Found an estimated cost of 114 for instruction: %V64I8 = call <64 x i8> @llvm.fshr.v64i8(<64 x i8> %a512, <64 x i8> %b512, <64 x i8> %c512)
; AVX1-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; AVX2-LABEL: 'var_funnel_i8'
@@ -292,8 +292,8 @@ define void @var_funnel_i8(i8 %a8, <16 x i8> %a128, <32 x i8> %a256, <64 x i8> %
; XOP-LABEL: 'var_funnel_i8'
; XOP-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I8 = call i8 @llvm.fshr.i8(i8 %a8, i8 %b8, i8 %c8)
; XOP-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V16I8 = call <16 x i8> @llvm.fshr.v16i8(<16 x i8> %a128, <16 x i8> %b128, <16 x i8> %c128)
-; XOP-NEXT: Cost Model: Found an estimated cost of 23 for instruction: %V32I8 = call <32 x i8> @llvm.fshr.v32i8(<32 x i8> %a256, <32 x i8> %b256, <32 x i8> %c256)
-; XOP-NEXT: Cost Model: Found an estimated cost of 46 for instruction: %V64I8 = call <64 x i8> @llvm.fshr.v64i8(<64 x i8> %a512, <64 x i8> %b512, <64 x i8> %c512)
+; XOP-NEXT: Cost Model: Found an estimated cost of 22 for instruction: %V32I8 = call <32 x i8> @llvm.fshr.v32i8(<32 x i8> %a256, <32 x i8> %b256, <32 x i8> %c256)
+; XOP-NEXT: Cost Model: Found an estimated cost of 44 for instruction: %V64I8 = call <64 x i8> @llvm.fshr.v64i8(<64 x i8> %a512, <64 x i8> %b512, <64 x i8> %c512)
; XOP-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
%I8 = call i8 @llvm.fshr.i8(i8 %a8, i8 %b8, i8 %c8)
@@ -312,9 +312,9 @@ define void @splatvar_funnel_i64(i64 %a64, <2 x i64> %a128, <4 x i64> %a256, <8
; SSSE3-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %u128 = shufflevector <2 x i64> %c128, <2 x i64> undef, <2 x i32> zeroinitializer
; SSSE3-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %u256 = shufflevector <4 x i64> %c256, <4 x i64> undef, <4 x i32> zeroinitializer
; SSSE3-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %u512 = shufflevector <8 x i64> %c512, <8 x i64> undef, <8 x i32> zeroinitializer
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 13 for instruction: %V2I64 = call <2 x i64> @llvm.fshr.v2i64(<2 x i64> %a128, <2 x i64> %b128, <2 x i64> %u128)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 26 for instruction: %V4I64 = call <4 x i64> @llvm.fshr.v4i64(<4 x i64> %a256, <4 x i64> %b256, <4 x i64> %u256)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 52 for instruction: %V8I64 = call <8 x i64> @llvm.fshr.v8i64(<8 x i64> %a512, <8 x i64> %b512, <8 x i64> %u512)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %V2I64 = call <2 x i64> @llvm.fshr.v2i64(<2 x i64> %a128, <2 x i64> %b128, <2 x i64> %u128)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 24 for instruction: %V4I64 = call <4 x i64> @llvm.fshr.v4i64(<4 x i64> %a256, <4 x i64> %b256, <4 x i64> %u256)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 48 for instruction: %V8I64 = call <8 x i64> @llvm.fshr.v8i64(<8 x i64> %a512, <8 x i64> %b512, <8 x i64> %u512)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; SSE42-LABEL: 'splatvar_funnel_i64'
@@ -394,9 +394,9 @@ define void @splatvar_funnel_i32(i32 %a32, <4 x i32> %a128, <8 x i32> %a256, <16
; SSSE3-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %u128 = shufflevector <4 x i32> %c128, <4 x i32> undef, <4 x i32> zeroinitializer
; SSSE3-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %u256 = shufflevector <8 x i32> %c256, <8 x i32> undef, <8 x i32> zeroinitializer
; SSSE3-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %u512 = shufflevector <16 x i32> %c512, <16 x i32> undef, <16 x i32> zeroinitializer
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V2I32 = call <4 x i32> @llvm.fshr.v4i32(<4 x i32> %a128, <4 x i32> %b128, <4 x i32> %u128)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 18 for instruction: %V4I32 = call <8 x i32> @llvm.fshr.v8i32(<8 x i32> %a256, <8 x i32> %b256, <8 x i32> %u256)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 36 for instruction: %V8I32 = call <16 x i32> @llvm.fshr.v16i32(<16 x i32> %a512, <16 x i32> %b512, <16 x i32> %u512)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V2I32 = call <4 x i32> @llvm.fshr.v4i32(<4 x i32> %a128, <4 x i32> %b128, <4 x i32> %u128)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V4I32 = call <8 x i32> @llvm.fshr.v8i32(<8 x i32> %a256, <8 x i32> %b256, <8 x i32> %u256)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V8I32 = call <16 x i32> @llvm.fshr.v16i32(<16 x i32> %a512, <16 x i32> %b512, <16 x i32> %u512)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; SSE42-LABEL: 'splatvar_funnel_i32'
@@ -476,9 +476,9 @@ define void @splatvar_funnel_i16(i16 %a16, <8 x i16> %a128, <16 x i16> %a256, <3
; SSSE3-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %u128 = shufflevector <8 x i16> %c128, <8 x i16> undef, <8 x i32> zeroinitializer
; SSSE3-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %u256 = shufflevector <16 x i16> %c256, <16 x i16> undef, <16 x i32> zeroinitializer
; SSSE3-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %u512 = shufflevector <32 x i16> %c512, <32 x i16> undef, <32 x i32> zeroinitializer
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V8I16 = call <8 x i16> @llvm.fshr.v8i16(<8 x i16> %a128, <8 x i16> %b128, <8 x i16> %u128)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 18 for instruction: %V16I16 = call <16 x i16> @llvm.fshr.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> %u256)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 36 for instruction: %V32I16 = call <32 x i16> @llvm.fshr.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> %u512)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V8I16 = call <8 x i16> @llvm.fshr.v8i16(<8 x i16> %a128, <8 x i16> %b128, <8 x i16> %u128)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V16I16 = call <16 x i16> @llvm.fshr.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> %u256)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V32I16 = call <32 x i16> @llvm.fshr.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> %u512)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; SSE42-LABEL: 'splatvar_funnel_i16'
@@ -495,8 +495,8 @@ define void @splatvar_funnel_i16(i16 %a16, <8 x i16> %a128, <16 x i16> %a256, <3
; AVX1-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %u256 = shufflevector <16 x i16> %c256, <16 x i16> undef, <16 x i32> zeroinitializer
; AVX1-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %u512 = shufflevector <32 x i16> %c512, <32 x i16> undef, <32 x i32> zeroinitializer
; AVX1-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V8I16 = call <8 x i16> @llvm.fshr.v8i16(<8 x i16> %a128, <8 x i16> %b128, <8 x i16> %u128)
-; AVX1-NEXT: Cost Model: Found an estimated cost of 21 for instruction: %V16I16 = call <16 x i16> @llvm.fshr.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> %u256)
-; AVX1-NEXT: Cost Model: Found an estimated cost of 42 for instruction: %V32I16 = call <32 x i16> @llvm.fshr.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> %u512)
+; AVX1-NEXT: Cost Model: Found an estimated cost of 20 for instruction: %V16I16 = call <16 x i16> @llvm.fshr.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> %u256)
+; AVX1-NEXT: Cost Model: Found an estimated cost of 40 for instruction: %V32I16 = call <32 x i16> @llvm.fshr.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> %u512)
; AVX1-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; AVX2-LABEL: 'splatvar_funnel_i16'
@@ -558,8 +558,8 @@ define void @splatvar_funnel_i16(i16 %a16, <8 x i16> %a128, <16 x i16> %a256, <3
; XOP-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %u256 = shufflevector <16 x i16> %c256, <16 x i16> undef, <16 x i32> zeroinitializer
; XOP-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %u512 = shufflevector <32 x i16> %c512, <32 x i16> undef, <32 x i32> zeroinitializer
; XOP-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V8I16 = call <8 x i16> @llvm.fshr.v8i16(<8 x i16> %a128, <8 x i16> %b128, <8 x i16> %u128)
-; XOP-NEXT: Cost Model: Found an estimated cost of 23 for instruction: %V16I16 = call <16 x i16> @llvm.fshr.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> %u256)
-; XOP-NEXT: Cost Model: Found an estimated cost of 46 for instruction: %V32I16 = call <32 x i16> @llvm.fshr.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> %u512)
+; XOP-NEXT: Cost Model: Found an estimated cost of 22 for instruction: %V16I16 = call <16 x i16> @llvm.fshr.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> %u256)
+; XOP-NEXT: Cost Model: Found an estimated cost of 44 for instruction: %V32I16 = call <32 x i16> @llvm.fshr.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> %u512)
; XOP-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
%u128 = shufflevector <8 x i16> %c128, <8 x i16> undef, <8 x i32> zeroinitializer
@@ -576,9 +576,9 @@ define void @splatvar_funnel_i8(i8 %a8, <16 x i8> %a128, <32 x i8> %a256, <64 x
; SSSE3-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %u128 = shufflevector <16 x i8> %c128, <16 x i8> undef, <16 x i32> zeroinitializer
; SSSE3-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %u256 = shufflevector <32 x i8> %c256, <32 x i8> undef, <32 x i32> zeroinitializer
; SSSE3-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %u512 = shufflevector <64 x i8> %c512, <64 x i8> undef, <64 x i32> zeroinitializer
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 34 for instruction: %V16I8 = call <16 x i8> @llvm.fshr.v16i8(<16 x i8> %a128, <16 x i8> %b128, <16 x i8> %u128)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 68 for instruction: %V32I8 = call <32 x i8> @llvm.fshr.v32i8(<32 x i8> %a256, <32 x i8> %b256, <32 x i8> %u256)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 136 for instruction: %V64I8 = call <64 x i8> @llvm.fshr.v64i8(<64 x i8> %a512, <64 x i8> %b512, <64 x i8> %u512)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 33 for instruction: %V16I8 = call <16 x i8> @llvm.fshr.v16i8(<16 x i8> %a128, <16 x i8> %b128, <16 x i8> %u128)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 66 for instruction: %V32I8 = call <32 x i8> @llvm.fshr.v32i8(<32 x i8> %a256, <32 x i8> %b256, <32 x i8> %u256)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 132 for instruction: %V64I8 = call <64 x i8> @llvm.fshr.v64i8(<64 x i8> %a512, <64 x i8> %b512, <64 x i8> %u512)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; SSE42-LABEL: 'splatvar_funnel_i8'
@@ -595,8 +595,8 @@ define void @splatvar_funnel_i8(i8 %a8, <16 x i8> %a128, <32 x i8> %a256, <64 x
; AVX1-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %u256 = shufflevector <32 x i8> %c256, <32 x i8> undef, <32 x i32> zeroinitializer
; AVX1-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %u512 = shufflevector <64 x i8> %c512, <64 x i8> undef, <64 x i32> zeroinitializer
; AVX1-NEXT: Cost Model: Found an estimated cost of 26 for instruction: %V16I8 = call <16 x i8> @llvm.fshr.v16i8(<16 x i8> %a128, <16 x i8> %b128, <16 x i8> %u128)
-; AVX1-NEXT: Cost Model: Found an estimated cost of 58 for instruction: %V32I8 = call <32 x i8> @llvm.fshr.v32i8(<32 x i8> %a256, <32 x i8> %b256, <32 x i8> %u256)
-; AVX1-NEXT: Cost Model: Found an estimated cost of 116 for instruction: %V64I8 = call <64 x i8> @llvm.fshr.v64i8(<64 x i8> %a512, <64 x i8> %b512, <64 x i8> %u512)
+; AVX1-NEXT: Cost Model: Found an estimated cost of 57 for instruction: %V32I8 = call <32 x i8> @llvm.fshr.v32i8(<32 x i8> %a256, <32 x i8> %b256, <32 x i8> %u256)
+; AVX1-NEXT: Cost Model: Found an estimated cost of 114 for instruction: %V64I8 = call <64 x i8> @llvm.fshr.v64i8(<64 x i8> %a512, <64 x i8> %b512, <64 x i8> %u512)
; AVX1-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; AVX2-LABEL: 'splatvar_funnel_i8'
@@ -658,8 +658,8 @@ define void @splatvar_funnel_i8(i8 %a8, <16 x i8> %a128, <32 x i8> %a256, <64 x
; XOP-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %u256 = shufflevector <32 x i8> %c256, <32 x i8> undef, <32 x i32> zeroinitializer
; XOP-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %u512 = shufflevector <64 x i8> %c512, <64 x i8> undef, <64 x i32> zeroinitializer
; XOP-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V16I8 = call <16 x i8> @llvm.fshr.v16i8(<16 x i8> %a128, <16 x i8> %b128, <16 x i8> %u128)
-; XOP-NEXT: Cost Model: Found an estimated cost of 23 for instruction: %V32I8 = call <32 x i8> @llvm.fshr.v32i8(<32 x i8> %a256, <32 x i8> %b256, <32 x i8> %u256)
-; XOP-NEXT: Cost Model: Found an estimated cost of 46 for instruction: %V64I8 = call <64 x i8> @llvm.fshr.v64i8(<64 x i8> %a512, <64 x i8> %b512, <64 x i8> %u512)
+; XOP-NEXT: Cost Model: Found an estimated cost of 22 for instruction: %V32I8 = call <32 x i8> @llvm.fshr.v32i8(<32 x i8> %a256, <32 x i8> %b256, <32 x i8> %u256)
+; XOP-NEXT: Cost Model: Found an estimated cost of 44 for instruction: %V64I8 = call <64 x i8> @llvm.fshr.v64i8(<64 x i8> %a512, <64 x i8> %b512, <64 x i8> %u512)
; XOP-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
%u128 = shufflevector <16 x i8> %c128, <16 x i8> undef, <16 x i32> zeroinitializer
@@ -678,9 +678,9 @@ define void @splatvar_funnel_i8(i8 %a8, <16 x i8> %a128, <32 x i8> %a256, <64 x
define void @constant_funnel_i64(i64 %a64, <2 x i64> %a128, <4 x i64> %a256, <8 x i64> %a512, i64 %b64, <2 x i64> %b128, <4 x i64> %b256, <8 x i64> %b512) {
; SSSE3-LABEL: 'constant_funnel_i64'
; SSSE3-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I64 = call i64 @llvm.fshr.i64(i64 %a64, i64 %b64, i64 7)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 18 for instruction: %V2I64 = call <2 x i64> @llvm.fshr.v2i64(<2 x i64> %a128, <2 x i64> %b128, <2 x i64> <i64 1, i64 7>)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 36 for instruction: %V4I64 = call <4 x i64> @llvm.fshr.v4i64(<4 x i64> %a256, <4 x i64> %b256, <4 x i64> <i64 1, i64 7, i64 15, i64 31>)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 72 for instruction: %V8I64 = call <8 x i64> @llvm.fshr.v8i64(<8 x i64> %a512, <8 x i64> %b512, <8 x i64> <i64 1, i64 7, i64 15, i64 31, i64 1, i64 7, i64 15, i64 31>)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 17 for instruction: %V2I64 = call <2 x i64> @llvm.fshr.v2i64(<2 x i64> %a128, <2 x i64> %b128, <2 x i64> <i64 1, i64 7>)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 34 for instruction: %V4I64 = call <4 x i64> @llvm.fshr.v4i64(<4 x i64> %a256, <4 x i64> %b256, <4 x i64> <i64 1, i64 7, i64 15, i64 31>)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 68 for instruction: %V8I64 = call <8 x i64> @llvm.fshr.v8i64(<8 x i64> %a512, <8 x i64> %b512, <8 x i64> <i64 1, i64 7, i64 15, i64 31, i64 1, i64 7, i64 15, i64 31>)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; SSE42-LABEL: 'constant_funnel_i64'
@@ -742,9 +742,9 @@ define void @constant_funnel_i64(i64 %a64, <2 x i64> %a128, <4 x i64> %a256, <8
define void @constant_funnel_i32(i32 %a32, <4 x i32> %a128, <8 x i32> %a256, <16 x i32> %a512, i32 %b32, <4 x i32> %b128, <8 x i32> %b256, <16 x i32> %b512) {
; SSSE3-LABEL: 'constant_funnel_i32'
; SSSE3-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I32 = call i32 @llvm.fshr.i32(i32 %a32, i32 %b32, i32 7)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 24 for instruction: %V2I32 = call <4 x i32> @llvm.fshr.v4i32(<4 x i32> %a128, <4 x i32> %b128, <4 x i32> <i32 4, i32 5, i32 6, i32 7>)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 48 for instruction: %V4I32 = call <8 x i32> @llvm.fshr.v8i32(<8 x i32> %a256, <8 x i32> %b256, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 96 for instruction: %V8I32 = call <16 x i32> @llvm.fshr.v16i32(<16 x i32> %a512, <16 x i32> %b512, <16 x i32> <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 23 for instruction: %V2I32 = call <4 x i32> @llvm.fshr.v4i32(<4 x i32> %a128, <4 x i32> %b128, <4 x i32> <i32 4, i32 5, i32 6, i32 7>)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 46 for instruction: %V4I32 = call <8 x i32> @llvm.fshr.v8i32(<8 x i32> %a256, <8 x i32> %b256, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 92 for instruction: %V8I32 = call <16 x i32> @llvm.fshr.v16i32(<16 x i32> %a512, <16 x i32> %b512, <16 x i32> <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; SSE42-LABEL: 'constant_funnel_i32'
@@ -806,9 +806,9 @@ define void @constant_funnel_i32(i32 %a32, <4 x i32> %a128, <8 x i32> %a256, <16
define void @constant_funnel_i16(i16 %a16, <8 x i16> %a128, <16 x i16> %a256, <32 x i16> %a512, i16 %b16, <8 x i16> %b128, <16 x i16> %b256, <32 x i16> %b512) {
; SSSE3-LABEL: 'constant_funnel_i16'
; SSSE3-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I16 = call i16 @llvm.fshr.i16(i16 %a16, i16 %b16, i16 7)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 23 for instruction: %V8I16 = call <8 x i16> @llvm.fshr.v8i16(<8 x i16> %a128, <8 x i16> %b128, <8 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 46 for instruction: %V16I16 = call <16 x i16> @llvm.fshr.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 92 for instruction: %V32I16 = call <32 x i16> @llvm.fshr.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 22 for instruction: %V8I16 = call <8 x i16> @llvm.fshr.v8i16(<8 x i16> %a128, <8 x i16> %b128, <8 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 44 for instruction: %V16I16 = call <16 x i16> @llvm.fshr.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 88 for instruction: %V32I16 = call <32 x i16> @llvm.fshr.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; SSE42-LABEL: 'constant_funnel_i16'
@@ -821,8 +821,8 @@ define void @constant_funnel_i16(i16 %a16, <8 x i16> %a128, <16 x i16> %a256, <3
; AVX1-LABEL: 'constant_funnel_i16'
; AVX1-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I16 = call i16 @llvm.fshr.i16(i16 %a16, i16 %b16, i16 7)
; AVX1-NEXT: Cost Model: Found an estimated cost of 18 for instruction: %V8I16 = call <8 x i16> @llvm.fshr.v8i16(<8 x i16> %a128, <8 x i16> %b128, <8 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
-; AVX1-NEXT: Cost Model: Found an estimated cost of 44 for instruction: %V16I16 = call <16 x i16> @llvm.fshr.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
-; AVX1-NEXT: Cost Model: Found an estimated cost of 88 for instruction: %V32I16 = call <32 x i16> @llvm.fshr.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
+; AVX1-NEXT: Cost Model: Found an estimated cost of 43 for instruction: %V16I16 = call <16 x i16> @llvm.fshr.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
+; AVX1-NEXT: Cost Model: Found an estimated cost of 86 for instruction: %V32I16 = call <32 x i16> @llvm.fshr.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
; AVX1-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; AVX2-LABEL: 'constant_funnel_i16'
@@ -870,8 +870,8 @@ define void @constant_funnel_i16(i16 %a16, <8 x i16> %a128, <16 x i16> %a256, <3
; XOP-LABEL: 'constant_funnel_i16'
; XOP-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I16 = call i16 @llvm.fshr.i16(i16 %a16, i16 %b16, i16 7)
; XOP-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V8I16 = call <8 x i16> @llvm.fshr.v8i16(<8 x i16> %a128, <8 x i16> %b128, <8 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
-; XOP-NEXT: Cost Model: Found an estimated cost of 20 for instruction: %V16I16 = call <16 x i16> @llvm.fshr.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
-; XOP-NEXT: Cost Model: Found an estimated cost of 40 for instruction: %V32I16 = call <32 x i16> @llvm.fshr.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
+; XOP-NEXT: Cost Model: Found an estimated cost of 19 for instruction: %V16I16 = call <16 x i16> @llvm.fshr.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
+; XOP-NEXT: Cost Model: Found an estimated cost of 38 for instruction: %V32I16 = call <32 x i16> @llvm.fshr.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>)
; XOP-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
%I16 = call i16 @llvm.fshr.i16(i16 %a16, i16 %b16, i16 7)
@@ -884,9 +884,9 @@ define void @constant_funnel_i16(i16 %a16, <8 x i16> %a128, <16 x i16> %a256, <3
define void @constant_funnel_i8(i8 %a8, <16 x i8> %a128, <32 x i8> %a256, <64 x i8> %a512, i8 %b8, <16 x i8> %b128, <32 x i8> %b256, <64 x i8> %b512) {
; SSSE3-LABEL: 'constant_funnel_i8'
; SSSE3-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I8 = call i8 @llvm.fshr.i8(i8 %a8, i8 %b8, i8 7)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 33 for instruction: %V16I8 = call <16 x i8> @llvm.fshr.v16i8(<16 x i8> %a128, <16 x i8> %b128, <16 x i8> <i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0>)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 66 for instruction: %V32I8 = call <32 x i8> @llvm.fshr.v32i8(<32 x i8> %a256, <32 x i8> %b256, <32 x i8> <i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0>)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 132 for instruction: %V64I8 = call <64 x i8> @llvm.fshr.v64i8(<64 x i8> %a512, <64 x i8> %b512, <64 x i8> <i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0>)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V16I8 = call <16 x i8> @llvm.fshr.v16i8(<16 x i8> %a128, <16 x i8> %b128, <16 x i8> <i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0>)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 64 for instruction: %V32I8 = call <32 x i8> @llvm.fshr.v32i8(<32 x i8> %a256, <32 x i8> %b256, <32 x i8> <i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0>)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 128 for instruction: %V64I8 = call <64 x i8> @llvm.fshr.v64i8(<64 x i8> %a512, <64 x i8> %b512, <64 x i8> <i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0>)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; SSE42-LABEL: 'constant_funnel_i8'
@@ -899,8 +899,8 @@ define void @constant_funnel_i8(i8 %a8, <16 x i8> %a128, <32 x i8> %a256, <64 x
; AVX1-LABEL: 'constant_funnel_i8'
; AVX1-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I8 = call i8 @llvm.fshr.i8(i8 %a8, i8 %b8, i8 7)
; AVX1-NEXT: Cost Model: Found an estimated cost of 25 for instruction: %V16I8 = call <16 x i8> @llvm.fshr.v16i8(<16 x i8> %a128, <16 x i8> %b128, <16 x i8> <i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0>)
-; AVX1-NEXT: Cost Model: Found an estimated cost of 57 for instruction: %V32I8 = call <32 x i8> @llvm.fshr.v32i8(<32 x i8> %a256, <32 x i8> %b256, <32 x i8> <i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0>)
-; AVX1-NEXT: Cost Model: Found an estimated cost of 114 for instruction: %V64I8 = call <64 x i8> @llvm.fshr.v64i8(<64 x i8> %a512, <64 x i8> %b512, <64 x i8> <i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0>)
+; AVX1-NEXT: Cost Model: Found an estimated cost of 56 for instruction: %V32I8 = call <32 x i8> @llvm.fshr.v32i8(<32 x i8> %a256, <32 x i8> %b256, <32 x i8> <i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0>)
+; AVX1-NEXT: Cost Model: Found an estimated cost of 112 for instruction: %V64I8 = call <64 x i8> @llvm.fshr.v64i8(<64 x i8> %a512, <64 x i8> %b512, <64 x i8> <i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0>)
; AVX1-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; AVX2-LABEL: 'constant_funnel_i8'
@@ -948,8 +948,8 @@ define void @constant_funnel_i8(i8 %a8, <16 x i8> %a128, <32 x i8> %a256, <64 x
; XOP-LABEL: 'constant_funnel_i8'
; XOP-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I8 = call i8 @llvm.fshr.i8(i8 %a8, i8 %b8, i8 7)
; XOP-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V16I8 = call <16 x i8> @llvm.fshr.v16i8(<16 x i8> %a128, <16 x i8> %b128, <16 x i8> <i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0>)
-; XOP-NEXT: Cost Model: Found an estimated cost of 20 for instruction: %V32I8 = call <32 x i8> @llvm.fshr.v32i8(<32 x i8> %a256, <32 x i8> %b256, <32 x i8> <i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0>)
-; XOP-NEXT: Cost Model: Found an estimated cost of 40 for instruction: %V64I8 = call <64 x i8> @llvm.fshr.v64i8(<64 x i8> %a512, <64 x i8> %b512, <64 x i8> <i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0>)
+; XOP-NEXT: Cost Model: Found an estimated cost of 19 for instruction: %V32I8 = call <32 x i8> @llvm.fshr.v32i8(<32 x i8> %a256, <32 x i8> %b256, <32 x i8> <i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0>)
+; XOP-NEXT: Cost Model: Found an estimated cost of 38 for instruction: %V64I8 = call <64 x i8> @llvm.fshr.v64i8(<64 x i8> %a512, <64 x i8> %b512, <64 x i8> <i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0>)
; XOP-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
%I8 = call i8 @llvm.fshr.i8(i8 %a8, i8 %b8, i8 7)
@@ -966,9 +966,9 @@ define void @constant_funnel_i8(i8 %a8, <16 x i8> %a128, <32 x i8> %a256, <64 x
define void @splatconstant_funnel_i64(i64 %a64, <2 x i64> %a128, <4 x i64> %a256, <8 x i64> %a512, i64 %b64, <2 x i64> %b128, <4 x i64> %b256, <8 x i64> %b512) {
; SSSE3-LABEL: 'splatconstant_funnel_i64'
; SSSE3-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I64 = call i64 @llvm.fshr.i64(i64 %a64, i64 %b64, i64 7)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %V2I64 = call <2 x i64> @llvm.fshr.v2i64(<2 x i64> %a128, <2 x i64> %b128, <2 x i64> <i64 7, i64 7>)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 24 for instruction: %V4I64 = call <4 x i64> @llvm.fshr.v4i64(<4 x i64> %a256, <4 x i64> %b256, <4 x i64> <i64 7, i64 7, i64 7, i64 7>)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 48 for instruction: %V8I64 = call <8 x i64> @llvm.fshr.v8i64(<8 x i64> %a512, <8 x i64> %b512, <8 x i64> <i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7>)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 11 for instruction: %V2I64 = call <2 x i64> @llvm.fshr.v2i64(<2 x i64> %a128, <2 x i64> %b128, <2 x i64> <i64 7, i64 7>)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 22 for instruction: %V4I64 = call <4 x i64> @llvm.fshr.v4i64(<4 x i64> %a256, <4 x i64> %b256, <4 x i64> <i64 7, i64 7, i64 7, i64 7>)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 44 for instruction: %V8I64 = call <8 x i64> @llvm.fshr.v8i64(<8 x i64> %a512, <8 x i64> %b512, <8 x i64> <i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7, i64 7>)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; SSE42-LABEL: 'splatconstant_funnel_i64'
@@ -1030,9 +1030,9 @@ define void @splatconstant_funnel_i64(i64 %a64, <2 x i64> %a128, <4 x i64> %a256
define void @splatconstant_funnel_i32(i32 %a32, <4 x i32> %a128, <8 x i32> %a256, <16 x i32> %a512, i32 %b32, <4 x i32> %b128, <8 x i32> %b256, <16 x i32> %b512) {
; SSSE3-LABEL: 'splatconstant_funnel_i32'
; SSSE3-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I32 = call i32 @llvm.fshr.i32(i32 %a32, i32 %b32, i32 5)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V2I32 = call <4 x i32> @llvm.fshr.v4i32(<4 x i32> %a128, <4 x i32> %b128, <4 x i32> <i32 5, i32 5, i32 5, i32 5>)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V4I32 = call <8 x i32> @llvm.fshr.v8i32(<8 x i32> %a256, <8 x i32> %b256, <8 x i32> <i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5>)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V8I32 = call <16 x i32> @llvm.fshr.v16i32(<16 x i32> %a512, <16 x i32> %b512, <16 x i32> <i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5>)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V2I32 = call <4 x i32> @llvm.fshr.v4i32(<4 x i32> %a128, <4 x i32> %b128, <4 x i32> <i32 5, i32 5, i32 5, i32 5>)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 14 for instruction: %V4I32 = call <8 x i32> @llvm.fshr.v8i32(<8 x i32> %a256, <8 x i32> %b256, <8 x i32> <i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5>)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 28 for instruction: %V8I32 = call <16 x i32> @llvm.fshr.v16i32(<16 x i32> %a512, <16 x i32> %b512, <16 x i32> <i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5>)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; SSE42-LABEL: 'splatconstant_funnel_i32'
@@ -1094,9 +1094,9 @@ define void @splatconstant_funnel_i32(i32 %a32, <4 x i32> %a128, <8 x i32> %a256
define void @splatconstant_funnel_i16(i16 %a16, <8 x i16> %a128, <16 x i16> %a256, <32 x i16> %a512, i16 %b16, <8 x i16> %b128, <16 x i16> %b256, <32 x i16> %b512) {
; SSSE3-LABEL: 'splatconstant_funnel_i16'
; SSSE3-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I16 = call i16 @llvm.fshr.i16(i16 %a16, i16 %b16, i16 3)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V8I16 = call <8 x i16> @llvm.fshr.v8i16(<8 x i16> %a128, <8 x i16> %b128, <8 x i16> <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V16I16 = call <16 x i16> @llvm.fshr.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V32I16 = call <32 x i16> @llvm.fshr.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V8I16 = call <8 x i16> @llvm.fshr.v8i16(<8 x i16> %a128, <8 x i16> %b128, <8 x i16> <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 14 for instruction: %V16I16 = call <16 x i16> @llvm.fshr.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 28 for instruction: %V32I16 = call <32 x i16> @llvm.fshr.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; SSE42-LABEL: 'splatconstant_funnel_i16'
@@ -1109,8 +1109,8 @@ define void @splatconstant_funnel_i16(i16 %a16, <8 x i16> %a128, <16 x i16> %a25
; AVX1-LABEL: 'splatconstant_funnel_i16'
; AVX1-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I16 = call i16 @llvm.fshr.i16(i16 %a16, i16 %b16, i16 3)
; AVX1-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V8I16 = call <8 x i16> @llvm.fshr.v8i16(<8 x i16> %a128, <8 x i16> %b128, <8 x i16> <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>)
-; AVX1-NEXT: Cost Model: Found an estimated cost of 20 for instruction: %V16I16 = call <16 x i16> @llvm.fshr.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>)
-; AVX1-NEXT: Cost Model: Found an estimated cost of 40 for instruction: %V32I16 = call <32 x i16> @llvm.fshr.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>)
+; AVX1-NEXT: Cost Model: Found an estimated cost of 19 for instruction: %V16I16 = call <16 x i16> @llvm.fshr.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>)
+; AVX1-NEXT: Cost Model: Found an estimated cost of 38 for instruction: %V32I16 = call <32 x i16> @llvm.fshr.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>)
; AVX1-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; AVX2-LABEL: 'splatconstant_funnel_i16'
@@ -1158,8 +1158,8 @@ define void @splatconstant_funnel_i16(i16 %a16, <8 x i16> %a128, <16 x i16> %a25
; XOP-LABEL: 'splatconstant_funnel_i16'
; XOP-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I16 = call i16 @llvm.fshr.i16(i16 %a16, i16 %b16, i16 3)
; XOP-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V8I16 = call <8 x i16> @llvm.fshr.v8i16(<8 x i16> %a128, <8 x i16> %b128, <8 x i16> <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>)
-; XOP-NEXT: Cost Model: Found an estimated cost of 20 for instruction: %V16I16 = call <16 x i16> @llvm.fshr.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>)
-; XOP-NEXT: Cost Model: Found an estimated cost of 40 for instruction: %V32I16 = call <32 x i16> @llvm.fshr.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>)
+; XOP-NEXT: Cost Model: Found an estimated cost of 19 for instruction: %V16I16 = call <16 x i16> @llvm.fshr.v16i16(<16 x i16> %a256, <16 x i16> %b256, <16 x i16> <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>)
+; XOP-NEXT: Cost Model: Found an estimated cost of 38 for instruction: %V32I16 = call <32 x i16> @llvm.fshr.v32i16(<32 x i16> %a512, <32 x i16> %b512, <32 x i16> <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>)
; XOP-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
%I16 = call i16 @llvm.fshr.i16(i16 %a16, i16 %b16, i16 3)
@@ -1172,9 +1172,9 @@ define void @splatconstant_funnel_i16(i16 %a16, <8 x i16> %a128, <16 x i16> %a25
define void @splatconstant_funnel_i8(i8 %a8, <16 x i8> %a128, <32 x i8> %a256, <64 x i8> %a512, i8 %b8, <16 x i8> %b128, <32 x i8> %b256, <64 x i8> %b512) {
; SSSE3-LABEL: 'splatconstant_funnel_i8'
; SSSE3-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I8 = call i8 @llvm.fshr.i8(i8 %a8, i8 %b8, i8 3)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 10 for instruction: %V16I8 = call <16 x i8> @llvm.fshr.v16i8(<16 x i8> %a128, <16 x i8> %b128, <16 x i8> <i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3>)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 20 for instruction: %V32I8 = call <32 x i8> @llvm.fshr.v32i8(<32 x i8> %a256, <32 x i8> %b256, <32 x i8> <i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3>)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 40 for instruction: %V64I8 = call <64 x i8> @llvm.fshr.v64i8(<64 x i8> %a512, <64 x i8> %b512, <64 x i8> <i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3>)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V16I8 = call <16 x i8> @llvm.fshr.v16i8(<16 x i8> %a128, <16 x i8> %b128, <16 x i8> <i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3>)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 18 for instruction: %V32I8 = call <32 x i8> @llvm.fshr.v32i8(<32 x i8> %a256, <32 x i8> %b256, <32 x i8> <i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3>)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 36 for instruction: %V64I8 = call <64 x i8> @llvm.fshr.v64i8(<64 x i8> %a512, <64 x i8> %b512, <64 x i8> <i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3>)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; SSE42-LABEL: 'splatconstant_funnel_i8'
@@ -1187,8 +1187,8 @@ define void @splatconstant_funnel_i8(i8 %a8, <16 x i8> %a128, <32 x i8> %a256, <
; AVX1-LABEL: 'splatconstant_funnel_i8'
; AVX1-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I8 = call i8 @llvm.fshr.i8(i8 %a8, i8 %b8, i8 3)
; AVX1-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V16I8 = call <16 x i8> @llvm.fshr.v16i8(<16 x i8> %a128, <16 x i8> %b128, <16 x i8> <i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3>)
-; AVX1-NEXT: Cost Model: Found an estimated cost of 24 for instruction: %V32I8 = call <32 x i8> @llvm.fshr.v32i8(<32 x i8> %a256, <32 x i8> %b256, <32 x i8> <i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3>)
-; AVX1-NEXT: Cost Model: Found an estimated cost of 48 for instruction: %V64I8 = call <64 x i8> @llvm.fshr.v64i8(<64 x i8> %a512, <64 x i8> %b512, <64 x i8> <i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3>)
+; AVX1-NEXT: Cost Model: Found an estimated cost of 23 for instruction: %V32I8 = call <32 x i8> @llvm.fshr.v32i8(<32 x i8> %a256, <32 x i8> %b256, <32 x i8> <i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3>)
+; AVX1-NEXT: Cost Model: Found an estimated cost of 46 for instruction: %V64I8 = call <64 x i8> @llvm.fshr.v64i8(<64 x i8> %a512, <64 x i8> %b512, <64 x i8> <i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3>)
; AVX1-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; AVX2-LABEL: 'splatconstant_funnel_i8'
@@ -1236,8 +1236,8 @@ define void @splatconstant_funnel_i8(i8 %a8, <16 x i8> %a128, <32 x i8> %a256, <
; XOP-LABEL: 'splatconstant_funnel_i8'
; XOP-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %I8 = call i8 @llvm.fshr.i8(i8 %a8, i8 %b8, i8 3)
; XOP-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V16I8 = call <16 x i8> @llvm.fshr.v16i8(<16 x i8> %a128, <16 x i8> %b128, <16 x i8> <i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3>)
-; XOP-NEXT: Cost Model: Found an estimated cost of 20 for instruction: %V32I8 = call <32 x i8> @llvm.fshr.v32i8(<32 x i8> %a256, <32 x i8> %b256, <32 x i8> <i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3>)
-; XOP-NEXT: Cost Model: Found an estimated cost of 40 for instruction: %V64I8 = call <64 x i8> @llvm.fshr.v64i8(<64 x i8> %a512, <64 x i8> %b512, <64 x i8> <i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3>)
+; XOP-NEXT: Cost Model: Found an estimated cost of 19 for instruction: %V32I8 = call <32 x i8> @llvm.fshr.v32i8(<32 x i8> %a256, <32 x i8> %b256, <32 x i8> <i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3>)
+; XOP-NEXT: Cost Model: Found an estimated cost of 38 for instruction: %V64I8 = call <64 x i8> @llvm.fshr.v64i8(<64 x i8> %a512, <64 x i8> %b512, <64 x i8> <i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3>)
; XOP-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
%I8 = call i8 @llvm.fshr.i8(i8 %a8, i8 %b8, i8 3)
diff --git a/llvm/test/Analysis/CostModel/X86/intrinsic-cost-kinds.ll b/llvm/test/Analysis/CostModel/X86/intrinsic-cost-kinds.ll
index b1994678a114d..1859309552356 100644
--- a/llvm/test/Analysis/CostModel/X86/intrinsic-cost-kinds.ll
+++ b/llvm/test/Analysis/CostModel/X86/intrinsic-cost-kinds.ll
@@ -74,7 +74,7 @@ define void @umul(i32 %a, i32 %b, <16 x i32> %va, <16 x i32> %vb) {
define void @smax(i32 %a, i32 %b, <16 x i32> %va, <16 x i32> %vb) {
; THRU-LABEL: 'smax'
; THRU-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %s = call i32 @llvm.smax.i32(i32 %a, i32 %b)
-; THRU-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %v = call <16 x i32> @llvm.smax.v16i32(<16 x i32> %va, <16 x i32> %vb)
+; THRU-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %v = call <16 x i32> @llvm.smax.v16i32(<16 x i32> %va, <16 x i32> %vb)
; THRU-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; LATE-LABEL: 'smax'
@@ -256,7 +256,7 @@ define void @ctlz(i32 %a, <16 x i32> %va) {
define void @fshl(i32 %a, i32 %b, i32 %c, <16 x i32> %va, <16 x i32> %vb, <16 x i32> %vc) {
; THRU-LABEL: 'fshl'
; THRU-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %s = call i32 @llvm.fshl.i32(i32 %a, i32 %b, i32 %c)
-; THRU-NEXT: Cost Model: Found an estimated cost of 140 for instruction: %v = call <16 x i32> @llvm.fshl.v16i32(<16 x i32> %va, <16 x i32> %vb, <16 x i32> %vc)
+; THRU-NEXT: Cost Model: Found an estimated cost of 136 for instruction: %v = call <16 x i32> @llvm.fshl.v16i32(<16 x i32> %va, <16 x i32> %vb, <16 x i32> %vc)
; THRU-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; LATE-LABEL: 'fshl'
diff --git a/llvm/test/Analysis/CostModel/X86/reduce-smax.ll b/llvm/test/Analysis/CostModel/X86/reduce-smax.ll
index a7e5808efec3f..67e213fcb76cc 100644
--- a/llvm/test/Analysis/CostModel/X86/reduce-smax.ll
+++ b/llvm/test/Analysis/CostModel/X86/reduce-smax.ll
@@ -12,18 +12,18 @@
define i32 @reduce_i64(i32 %arg) {
; SSE2-LABEL: 'reduce_i64'
; SSE2-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %V1 = call i64 @llvm.vector.reduce.smax.v1i64(<1 x i64> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 13 for instruction: %V2 = call i64 @llvm.vector.reduce.smax.v2i64(<2 x i64> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 24 for instruction: %V4 = call i64 @llvm.vector.reduce.smax.v4i64(<4 x i64> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 46 for instruction: %V8 = call i64 @llvm.vector.reduce.smax.v8i64(<8 x i64> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 90 for instruction: %V16 = call i64 @llvm.vector.reduce.smax.v16i64(<16 x i64> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %V2 = call i64 @llvm.vector.reduce.smax.v2i64(<2 x i64> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 22 for instruction: %V4 = call i64 @llvm.vector.reduce.smax.v4i64(<4 x i64> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 42 for instruction: %V8 = call i64 @llvm.vector.reduce.smax.v8i64(<8 x i64> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 82 for instruction: %V16 = call i64 @llvm.vector.reduce.smax.v16i64(<16 x i64> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; SSSE3-LABEL: 'reduce_i64'
; SSSE3-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %V1 = call i64 @llvm.vector.reduce.smax.v1i64(<1 x i64> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 13 for instruction: %V2 = call i64 @llvm.vector.reduce.smax.v2i64(<2 x i64> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 24 for instruction: %V4 = call i64 @llvm.vector.reduce.smax.v4i64(<4 x i64> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 46 for instruction: %V8 = call i64 @llvm.vector.reduce.smax.v8i64(<8 x i64> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 90 for instruction: %V16 = call i64 @llvm.vector.reduce.smax.v16i64(<16 x i64> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %V2 = call i64 @llvm.vector.reduce.smax.v2i64(<2 x i64> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 22 for instruction: %V4 = call i64 @llvm.vector.reduce.smax.v4i64(<4 x i64> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 42 for instruction: %V8 = call i64 @llvm.vector.reduce.smax.v8i64(<8 x i64> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 82 for instruction: %V16 = call i64 @llvm.vector.reduce.smax.v16i64(<16 x i64> undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; SSE41-LABEL: 'reduce_i64'
@@ -76,19 +76,19 @@ define i32 @reduce_i64(i32 %arg) {
define i32 @reduce_i32(i32 %arg) {
; SSE2-LABEL: 'reduce_i32'
-; SSE2-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V2 = call i32 @llvm.vector.reduce.smax.v2i32(<2 x i32> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 17 for instruction: %V4 = call i32 @llvm.vector.reduce.smax.v4i32(<4 x i32> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 24 for instruction: %V8 = call i32 @llvm.vector.reduce.smax.v8i32(<8 x i32> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 38 for instruction: %V16 = call i32 @llvm.vector.reduce.smax.v16i32(<16 x i32> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 66 for instruction: %V32 = call i32 @llvm.vector.reduce.smax.v32i32(<32 x i32> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V2 = call i32 @llvm.vector.reduce.smax.v2i32(<2 x i32> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 15 for instruction: %V4 = call i32 @llvm.vector.reduce.smax.v4i32(<4 x i32> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 21 for instruction: %V8 = call i32 @llvm.vector.reduce.smax.v8i32(<8 x i32> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 33 for instruction: %V16 = call i32 @llvm.vector.reduce.smax.v16i32(<16 x i32> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 57 for instruction: %V32 = call i32 @llvm.vector.reduce.smax.v32i32(<32 x i32> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; SSSE3-LABEL: 'reduce_i32'
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V2 = call i32 @llvm.vector.reduce.smax.v2i32(<2 x i32> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 17 for instruction: %V4 = call i32 @llvm.vector.reduce.smax.v4i32(<4 x i32> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 24 for instruction: %V8 = call i32 @llvm.vector.reduce.smax.v8i32(<8 x i32> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 38 for instruction: %V16 = call i32 @llvm.vector.reduce.smax.v16i32(<16 x i32> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 66 for instruction: %V32 = call i32 @llvm.vector.reduce.smax.v32i32(<32 x i32> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V2 = call i32 @llvm.vector.reduce.smax.v2i32(<2 x i32> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 15 for instruction: %V4 = call i32 @llvm.vector.reduce.smax.v4i32(<4 x i32> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 21 for instruction: %V8 = call i32 @llvm.vector.reduce.smax.v8i32(<8 x i32> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 33 for instruction: %V16 = call i32 @llvm.vector.reduce.smax.v16i32(<16 x i32> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 57 for instruction: %V32 = call i32 @llvm.vector.reduce.smax.v32i32(<32 x i32> undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; SSE4-LABEL: 'reduce_i32'
@@ -215,23 +215,23 @@ define i32 @reduce_i16(i32 %arg) {
define i32 @reduce_i8(i32 %arg) {
; SSE2-LABEL: 'reduce_i8'
-; SSE2-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V2 = call i8 @llvm.vector.reduce.smax.v2i8(<2 x i8> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 17 for instruction: %V4 = call i8 @llvm.vector.reduce.smax.v4i8(<4 x i8> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 25 for instruction: %V8 = call i8 @llvm.vector.reduce.smax.v8i8(<8 x i8> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 33 for instruction: %V16 = call i8 @llvm.vector.reduce.smax.v16i8(<16 x i8> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 40 for instruction: %V32 = call i8 @llvm.vector.reduce.smax.v32i8(<32 x i8> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 54 for instruction: %V64 = call i8 @llvm.vector.reduce.smax.v64i8(<64 x i8> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 82 for instruction: %V128 = call i8 @llvm.vector.reduce.smax.v128i8(<128 x i8> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V2 = call i8 @llvm.vector.reduce.smax.v2i8(<2 x i8> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 15 for instruction: %V4 = call i8 @llvm.vector.reduce.smax.v4i8(<4 x i8> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 22 for instruction: %V8 = call i8 @llvm.vector.reduce.smax.v8i8(<8 x i8> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 29 for instruction: %V16 = call i8 @llvm.vector.reduce.smax.v16i8(<16 x i8> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 35 for instruction: %V32 = call i8 @llvm.vector.reduce.smax.v32i8(<32 x i8> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 47 for instruction: %V64 = call i8 @llvm.vector.reduce.smax.v64i8(<64 x i8> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 71 for instruction: %V128 = call i8 @llvm.vector.reduce.smax.v128i8(<128 x i8> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; SSSE3-LABEL: 'reduce_i8'
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V2 = call i8 @llvm.vector.reduce.smax.v2i8(<2 x i8> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 17 for instruction: %V4 = call i8 @llvm.vector.reduce.smax.v4i8(<4 x i8> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 25 for instruction: %V8 = call i8 @llvm.vector.reduce.smax.v8i8(<8 x i8> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 33 for instruction: %V16 = call i8 @llvm.vector.reduce.smax.v16i8(<16 x i8> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 40 for instruction: %V32 = call i8 @llvm.vector.reduce.smax.v32i8(<32 x i8> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 54 for instruction: %V64 = call i8 @llvm.vector.reduce.smax.v64i8(<64 x i8> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 82 for instruction: %V128 = call i8 @llvm.vector.reduce.smax.v128i8(<128 x i8> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V2 = call i8 @llvm.vector.reduce.smax.v2i8(<2 x i8> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 15 for instruction: %V4 = call i8 @llvm.vector.reduce.smax.v4i8(<4 x i8> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 22 for instruction: %V8 = call i8 @llvm.vector.reduce.smax.v8i8(<8 x i8> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 29 for instruction: %V16 = call i8 @llvm.vector.reduce.smax.v16i8(<16 x i8> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 35 for instruction: %V32 = call i8 @llvm.vector.reduce.smax.v32i8(<32 x i8> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 47 for instruction: %V64 = call i8 @llvm.vector.reduce.smax.v64i8(<64 x i8> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 71 for instruction: %V128 = call i8 @llvm.vector.reduce.smax.v128i8(<128 x i8> undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; SSE4-LABEL: 'reduce_i8'
diff --git a/llvm/test/Analysis/CostModel/X86/reduce-smin.ll b/llvm/test/Analysis/CostModel/X86/reduce-smin.ll
index 49aee00bb9f9c..db7ddfc5678ef 100644
--- a/llvm/test/Analysis/CostModel/X86/reduce-smin.ll
+++ b/llvm/test/Analysis/CostModel/X86/reduce-smin.ll
@@ -12,18 +12,18 @@
define i32 @reduce_i64(i32 %arg) {
; SSE2-LABEL: 'reduce_i64'
; SSE2-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %V1 = call i64 @llvm.vector.reduce.smin.v1i64(<1 x i64> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 13 for instruction: %V2 = call i64 @llvm.vector.reduce.smin.v2i64(<2 x i64> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 24 for instruction: %V4 = call i64 @llvm.vector.reduce.smin.v4i64(<4 x i64> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 46 for instruction: %V8 = call i64 @llvm.vector.reduce.smin.v8i64(<8 x i64> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 90 for instruction: %V16 = call i64 @llvm.vector.reduce.smin.v16i64(<16 x i64> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %V2 = call i64 @llvm.vector.reduce.smin.v2i64(<2 x i64> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 22 for instruction: %V4 = call i64 @llvm.vector.reduce.smin.v4i64(<4 x i64> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 42 for instruction: %V8 = call i64 @llvm.vector.reduce.smin.v8i64(<8 x i64> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 82 for instruction: %V16 = call i64 @llvm.vector.reduce.smin.v16i64(<16 x i64> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; SSSE3-LABEL: 'reduce_i64'
; SSSE3-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %V1 = call i64 @llvm.vector.reduce.smin.v1i64(<1 x i64> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 13 for instruction: %V2 = call i64 @llvm.vector.reduce.smin.v2i64(<2 x i64> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 24 for instruction: %V4 = call i64 @llvm.vector.reduce.smin.v4i64(<4 x i64> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 46 for instruction: %V8 = call i64 @llvm.vector.reduce.smin.v8i64(<8 x i64> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 90 for instruction: %V16 = call i64 @llvm.vector.reduce.smin.v16i64(<16 x i64> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %V2 = call i64 @llvm.vector.reduce.smin.v2i64(<2 x i64> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 22 for instruction: %V4 = call i64 @llvm.vector.reduce.smin.v4i64(<4 x i64> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 42 for instruction: %V8 = call i64 @llvm.vector.reduce.smin.v8i64(<8 x i64> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 82 for instruction: %V16 = call i64 @llvm.vector.reduce.smin.v16i64(<16 x i64> undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; SSE41-LABEL: 'reduce_i64'
@@ -76,19 +76,19 @@ define i32 @reduce_i64(i32 %arg) {
define i32 @reduce_i32(i32 %arg) {
; SSE2-LABEL: 'reduce_i32'
-; SSE2-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V2 = call i32 @llvm.vector.reduce.smin.v2i32(<2 x i32> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 17 for instruction: %V4 = call i32 @llvm.vector.reduce.smin.v4i32(<4 x i32> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 24 for instruction: %V8 = call i32 @llvm.vector.reduce.smin.v8i32(<8 x i32> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 38 for instruction: %V16 = call i32 @llvm.vector.reduce.smin.v16i32(<16 x i32> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 66 for instruction: %V32 = call i32 @llvm.vector.reduce.smin.v32i32(<32 x i32> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V2 = call i32 @llvm.vector.reduce.smin.v2i32(<2 x i32> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 15 for instruction: %V4 = call i32 @llvm.vector.reduce.smin.v4i32(<4 x i32> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 21 for instruction: %V8 = call i32 @llvm.vector.reduce.smin.v8i32(<8 x i32> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 33 for instruction: %V16 = call i32 @llvm.vector.reduce.smin.v16i32(<16 x i32> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 57 for instruction: %V32 = call i32 @llvm.vector.reduce.smin.v32i32(<32 x i32> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; SSSE3-LABEL: 'reduce_i32'
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V2 = call i32 @llvm.vector.reduce.smin.v2i32(<2 x i32> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 17 for instruction: %V4 = call i32 @llvm.vector.reduce.smin.v4i32(<4 x i32> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 24 for instruction: %V8 = call i32 @llvm.vector.reduce.smin.v8i32(<8 x i32> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 38 for instruction: %V16 = call i32 @llvm.vector.reduce.smin.v16i32(<16 x i32> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 66 for instruction: %V32 = call i32 @llvm.vector.reduce.smin.v32i32(<32 x i32> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V2 = call i32 @llvm.vector.reduce.smin.v2i32(<2 x i32> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 15 for instruction: %V4 = call i32 @llvm.vector.reduce.smin.v4i32(<4 x i32> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 21 for instruction: %V8 = call i32 @llvm.vector.reduce.smin.v8i32(<8 x i32> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 33 for instruction: %V16 = call i32 @llvm.vector.reduce.smin.v16i32(<16 x i32> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 57 for instruction: %V32 = call i32 @llvm.vector.reduce.smin.v32i32(<32 x i32> undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; SSE4-LABEL: 'reduce_i32'
@@ -215,23 +215,23 @@ define i32 @reduce_i16(i32 %arg) {
define i32 @reduce_i8(i32 %arg) {
; SSE2-LABEL: 'reduce_i8'
-; SSE2-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V2 = call i8 @llvm.vector.reduce.smin.v2i8(<2 x i8> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 17 for instruction: %V4 = call i8 @llvm.vector.reduce.smin.v4i8(<4 x i8> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 25 for instruction: %V8 = call i8 @llvm.vector.reduce.smin.v8i8(<8 x i8> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 33 for instruction: %V16 = call i8 @llvm.vector.reduce.smin.v16i8(<16 x i8> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 40 for instruction: %V32 = call i8 @llvm.vector.reduce.smin.v32i8(<32 x i8> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 54 for instruction: %V64 = call i8 @llvm.vector.reduce.smin.v64i8(<64 x i8> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 82 for instruction: %V128 = call i8 @llvm.vector.reduce.smin.v128i8(<128 x i8> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V2 = call i8 @llvm.vector.reduce.smin.v2i8(<2 x i8> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 15 for instruction: %V4 = call i8 @llvm.vector.reduce.smin.v4i8(<4 x i8> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 22 for instruction: %V8 = call i8 @llvm.vector.reduce.smin.v8i8(<8 x i8> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 29 for instruction: %V16 = call i8 @llvm.vector.reduce.smin.v16i8(<16 x i8> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 35 for instruction: %V32 = call i8 @llvm.vector.reduce.smin.v32i8(<32 x i8> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 47 for instruction: %V64 = call i8 @llvm.vector.reduce.smin.v64i8(<64 x i8> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 71 for instruction: %V128 = call i8 @llvm.vector.reduce.smin.v128i8(<128 x i8> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; SSSE3-LABEL: 'reduce_i8'
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V2 = call i8 @llvm.vector.reduce.smin.v2i8(<2 x i8> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 17 for instruction: %V4 = call i8 @llvm.vector.reduce.smin.v4i8(<4 x i8> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 25 for instruction: %V8 = call i8 @llvm.vector.reduce.smin.v8i8(<8 x i8> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 33 for instruction: %V16 = call i8 @llvm.vector.reduce.smin.v16i8(<16 x i8> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 40 for instruction: %V32 = call i8 @llvm.vector.reduce.smin.v32i8(<32 x i8> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 54 for instruction: %V64 = call i8 @llvm.vector.reduce.smin.v64i8(<64 x i8> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 82 for instruction: %V128 = call i8 @llvm.vector.reduce.smin.v128i8(<128 x i8> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V2 = call i8 @llvm.vector.reduce.smin.v2i8(<2 x i8> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 15 for instruction: %V4 = call i8 @llvm.vector.reduce.smin.v4i8(<4 x i8> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 22 for instruction: %V8 = call i8 @llvm.vector.reduce.smin.v8i8(<8 x i8> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 29 for instruction: %V16 = call i8 @llvm.vector.reduce.smin.v16i8(<16 x i8> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 35 for instruction: %V32 = call i8 @llvm.vector.reduce.smin.v32i8(<32 x i8> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 47 for instruction: %V64 = call i8 @llvm.vector.reduce.smin.v64i8(<64 x i8> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 71 for instruction: %V128 = call i8 @llvm.vector.reduce.smin.v128i8(<128 x i8> undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; SSE4-LABEL: 'reduce_i8'
diff --git a/llvm/test/Analysis/CostModel/X86/reduce-umax.ll b/llvm/test/Analysis/CostModel/X86/reduce-umax.ll
index 2979197e05e16..8e26f87b9640c 100644
--- a/llvm/test/Analysis/CostModel/X86/reduce-umax.ll
+++ b/llvm/test/Analysis/CostModel/X86/reduce-umax.ll
@@ -12,18 +12,18 @@
define i32 @reduce_i64(i32 %arg) {
; SSE2-LABEL: 'reduce_i64'
; SSE2-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %V1 = call i64 @llvm.vector.reduce.umax.v1i64(<1 x i64> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 13 for instruction: %V2 = call i64 @llvm.vector.reduce.umax.v2i64(<2 x i64> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 24 for instruction: %V4 = call i64 @llvm.vector.reduce.umax.v4i64(<4 x i64> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 46 for instruction: %V8 = call i64 @llvm.vector.reduce.umax.v8i64(<8 x i64> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 90 for instruction: %V16 = call i64 @llvm.vector.reduce.umax.v16i64(<16 x i64> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %V2 = call i64 @llvm.vector.reduce.umax.v2i64(<2 x i64> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 22 for instruction: %V4 = call i64 @llvm.vector.reduce.umax.v4i64(<4 x i64> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 42 for instruction: %V8 = call i64 @llvm.vector.reduce.umax.v8i64(<8 x i64> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 82 for instruction: %V16 = call i64 @llvm.vector.reduce.umax.v16i64(<16 x i64> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; SSSE3-LABEL: 'reduce_i64'
; SSSE3-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %V1 = call i64 @llvm.vector.reduce.umax.v1i64(<1 x i64> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 13 for instruction: %V2 = call i64 @llvm.vector.reduce.umax.v2i64(<2 x i64> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 24 for instruction: %V4 = call i64 @llvm.vector.reduce.umax.v4i64(<4 x i64> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 46 for instruction: %V8 = call i64 @llvm.vector.reduce.umax.v8i64(<8 x i64> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 90 for instruction: %V16 = call i64 @llvm.vector.reduce.umax.v16i64(<16 x i64> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %V2 = call i64 @llvm.vector.reduce.umax.v2i64(<2 x i64> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 22 for instruction: %V4 = call i64 @llvm.vector.reduce.umax.v4i64(<4 x i64> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 42 for instruction: %V8 = call i64 @llvm.vector.reduce.umax.v8i64(<8 x i64> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 82 for instruction: %V16 = call i64 @llvm.vector.reduce.umax.v16i64(<16 x i64> undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; SSE41-LABEL: 'reduce_i64'
@@ -76,19 +76,19 @@ define i32 @reduce_i64(i32 %arg) {
define i32 @reduce_i32(i32 %arg) {
; SSE2-LABEL: 'reduce_i32'
-; SSE2-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V2 = call i32 @llvm.vector.reduce.umax.v2i32(<2 x i32> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 17 for instruction: %V4 = call i32 @llvm.vector.reduce.umax.v4i32(<4 x i32> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 24 for instruction: %V8 = call i32 @llvm.vector.reduce.umax.v8i32(<8 x i32> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 38 for instruction: %V16 = call i32 @llvm.vector.reduce.umax.v16i32(<16 x i32> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 66 for instruction: %V32 = call i32 @llvm.vector.reduce.umax.v32i32(<32 x i32> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V2 = call i32 @llvm.vector.reduce.umax.v2i32(<2 x i32> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 15 for instruction: %V4 = call i32 @llvm.vector.reduce.umax.v4i32(<4 x i32> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 21 for instruction: %V8 = call i32 @llvm.vector.reduce.umax.v8i32(<8 x i32> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 33 for instruction: %V16 = call i32 @llvm.vector.reduce.umax.v16i32(<16 x i32> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 57 for instruction: %V32 = call i32 @llvm.vector.reduce.umax.v32i32(<32 x i32> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; SSSE3-LABEL: 'reduce_i32'
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V2 = call i32 @llvm.vector.reduce.umax.v2i32(<2 x i32> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 17 for instruction: %V4 = call i32 @llvm.vector.reduce.umax.v4i32(<4 x i32> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 24 for instruction: %V8 = call i32 @llvm.vector.reduce.umax.v8i32(<8 x i32> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 38 for instruction: %V16 = call i32 @llvm.vector.reduce.umax.v16i32(<16 x i32> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 66 for instruction: %V32 = call i32 @llvm.vector.reduce.umax.v32i32(<32 x i32> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V2 = call i32 @llvm.vector.reduce.umax.v2i32(<2 x i32> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 15 for instruction: %V4 = call i32 @llvm.vector.reduce.umax.v4i32(<4 x i32> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 21 for instruction: %V8 = call i32 @llvm.vector.reduce.umax.v8i32(<8 x i32> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 33 for instruction: %V16 = call i32 @llvm.vector.reduce.umax.v16i32(<16 x i32> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 57 for instruction: %V32 = call i32 @llvm.vector.reduce.umax.v32i32(<32 x i32> undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; SSE4-LABEL: 'reduce_i32'
@@ -136,18 +136,18 @@ define i32 @reduce_i16(i32 %arg) {
; SSE2-NEXT: Cost Model: Found an estimated cost of 5 for instruction: %V2 = call i16 @llvm.vector.reduce.umax.v2i16(<2 x i16> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V4 = call i16 @llvm.vector.reduce.umax.v4i16(<4 x i16> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V8 = call i16 @llvm.vector.reduce.umax.v8i16(<8 x i16> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V16 = call i16 @llvm.vector.reduce.umax.v16i16(<16 x i16> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 30 for instruction: %V32 = call i16 @llvm.vector.reduce.umax.v32i16(<32 x i16> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 58 for instruction: %V64 = call i16 @llvm.vector.reduce.umax.v64i16(<64 x i16> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 15 for instruction: %V16 = call i16 @llvm.vector.reduce.umax.v16i16(<16 x i16> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 27 for instruction: %V32 = call i16 @llvm.vector.reduce.umax.v32i16(<32 x i16> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 51 for instruction: %V64 = call i16 @llvm.vector.reduce.umax.v64i16(<64 x i16> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; SSSE3-LABEL: 'reduce_i16'
; SSSE3-NEXT: Cost Model: Found an estimated cost of 5 for instruction: %V2 = call i16 @llvm.vector.reduce.umax.v2i16(<2 x i16> undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V4 = call i16 @llvm.vector.reduce.umax.v4i16(<4 x i16> undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V8 = call i16 @llvm.vector.reduce.umax.v8i16(<8 x i16> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V16 = call i16 @llvm.vector.reduce.umax.v16i16(<16 x i16> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 30 for instruction: %V32 = call i16 @llvm.vector.reduce.umax.v32i16(<32 x i16> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 58 for instruction: %V64 = call i16 @llvm.vector.reduce.umax.v64i16(<64 x i16> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 15 for instruction: %V16 = call i16 @llvm.vector.reduce.umax.v16i16(<16 x i16> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 27 for instruction: %V32 = call i16 @llvm.vector.reduce.umax.v32i16(<32 x i16> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 51 for instruction: %V64 = call i16 @llvm.vector.reduce.umax.v64i16(<64 x i16> undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; SSE4-LABEL: 'reduce_i16'
diff --git a/llvm/test/Analysis/CostModel/X86/reduce-umin.ll b/llvm/test/Analysis/CostModel/X86/reduce-umin.ll
index 356a99eb0f243..93f30b0d72bfd 100644
--- a/llvm/test/Analysis/CostModel/X86/reduce-umin.ll
+++ b/llvm/test/Analysis/CostModel/X86/reduce-umin.ll
@@ -12,18 +12,18 @@
define i32 @reduce_i64(i32 %arg) {
; SSE2-LABEL: 'reduce_i64'
; SSE2-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %V1 = call i64 @llvm.vector.reduce.umin.v1i64(<1 x i64> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 13 for instruction: %V2 = call i64 @llvm.vector.reduce.umin.v2i64(<2 x i64> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 24 for instruction: %V4 = call i64 @llvm.vector.reduce.umin.v4i64(<4 x i64> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 46 for instruction: %V8 = call i64 @llvm.vector.reduce.umin.v8i64(<8 x i64> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 90 for instruction: %V16 = call i64 @llvm.vector.reduce.umin.v16i64(<16 x i64> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %V2 = call i64 @llvm.vector.reduce.umin.v2i64(<2 x i64> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 22 for instruction: %V4 = call i64 @llvm.vector.reduce.umin.v4i64(<4 x i64> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 42 for instruction: %V8 = call i64 @llvm.vector.reduce.umin.v8i64(<8 x i64> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 82 for instruction: %V16 = call i64 @llvm.vector.reduce.umin.v16i64(<16 x i64> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; SSSE3-LABEL: 'reduce_i64'
; SSSE3-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %V1 = call i64 @llvm.vector.reduce.umin.v1i64(<1 x i64> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 13 for instruction: %V2 = call i64 @llvm.vector.reduce.umin.v2i64(<2 x i64> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 24 for instruction: %V4 = call i64 @llvm.vector.reduce.umin.v4i64(<4 x i64> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 46 for instruction: %V8 = call i64 @llvm.vector.reduce.umin.v8i64(<8 x i64> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 90 for instruction: %V16 = call i64 @llvm.vector.reduce.umin.v16i64(<16 x i64> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %V2 = call i64 @llvm.vector.reduce.umin.v2i64(<2 x i64> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 22 for instruction: %V4 = call i64 @llvm.vector.reduce.umin.v4i64(<4 x i64> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 42 for instruction: %V8 = call i64 @llvm.vector.reduce.umin.v8i64(<8 x i64> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 82 for instruction: %V16 = call i64 @llvm.vector.reduce.umin.v16i64(<16 x i64> undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; SSE41-LABEL: 'reduce_i64'
@@ -76,19 +76,19 @@ define i32 @reduce_i64(i32 %arg) {
define i32 @reduce_i32(i32 %arg) {
; SSE2-LABEL: 'reduce_i32'
-; SSE2-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V2 = call i32 @llvm.vector.reduce.umin.v2i32(<2 x i32> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 17 for instruction: %V4 = call i32 @llvm.vector.reduce.umin.v4i32(<4 x i32> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 24 for instruction: %V8 = call i32 @llvm.vector.reduce.umin.v8i32(<8 x i32> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 38 for instruction: %V16 = call i32 @llvm.vector.reduce.umin.v16i32(<16 x i32> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 66 for instruction: %V32 = call i32 @llvm.vector.reduce.umin.v32i32(<32 x i32> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V2 = call i32 @llvm.vector.reduce.umin.v2i32(<2 x i32> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 15 for instruction: %V4 = call i32 @llvm.vector.reduce.umin.v4i32(<4 x i32> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 21 for instruction: %V8 = call i32 @llvm.vector.reduce.umin.v8i32(<8 x i32> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 33 for instruction: %V16 = call i32 @llvm.vector.reduce.umin.v16i32(<16 x i32> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 57 for instruction: %V32 = call i32 @llvm.vector.reduce.umin.v32i32(<32 x i32> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; SSSE3-LABEL: 'reduce_i32'
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V2 = call i32 @llvm.vector.reduce.umin.v2i32(<2 x i32> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 17 for instruction: %V4 = call i32 @llvm.vector.reduce.umin.v4i32(<4 x i32> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 24 for instruction: %V8 = call i32 @llvm.vector.reduce.umin.v8i32(<8 x i32> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 38 for instruction: %V16 = call i32 @llvm.vector.reduce.umin.v16i32(<16 x i32> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 66 for instruction: %V32 = call i32 @llvm.vector.reduce.umin.v32i32(<32 x i32> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V2 = call i32 @llvm.vector.reduce.umin.v2i32(<2 x i32> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 15 for instruction: %V4 = call i32 @llvm.vector.reduce.umin.v4i32(<4 x i32> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 21 for instruction: %V8 = call i32 @llvm.vector.reduce.umin.v8i32(<8 x i32> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 33 for instruction: %V16 = call i32 @llvm.vector.reduce.umin.v16i32(<16 x i32> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 57 for instruction: %V32 = call i32 @llvm.vector.reduce.umin.v32i32(<32 x i32> undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; SSE4-LABEL: 'reduce_i32'
@@ -136,18 +136,18 @@ define i32 @reduce_i16(i32 %arg) {
; SSE2-NEXT: Cost Model: Found an estimated cost of 5 for instruction: %V2 = call i16 @llvm.vector.reduce.umin.v2i16(<2 x i16> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V4 = call i16 @llvm.vector.reduce.umin.v4i16(<4 x i16> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V8 = call i16 @llvm.vector.reduce.umin.v8i16(<8 x i16> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V16 = call i16 @llvm.vector.reduce.umin.v16i16(<16 x i16> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 30 for instruction: %V32 = call i16 @llvm.vector.reduce.umin.v32i16(<32 x i16> undef)
-; SSE2-NEXT: Cost Model: Found an estimated cost of 58 for instruction: %V64 = call i16 @llvm.vector.reduce.umin.v64i16(<64 x i16> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 15 for instruction: %V16 = call i16 @llvm.vector.reduce.umin.v16i16(<16 x i16> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 27 for instruction: %V32 = call i16 @llvm.vector.reduce.umin.v32i16(<32 x i16> undef)
+; SSE2-NEXT: Cost Model: Found an estimated cost of 51 for instruction: %V64 = call i16 @llvm.vector.reduce.umin.v64i16(<64 x i16> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; SSSE3-LABEL: 'reduce_i16'
; SSSE3-NEXT: Cost Model: Found an estimated cost of 5 for instruction: %V2 = call i16 @llvm.vector.reduce.umin.v2i16(<2 x i16> undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V4 = call i16 @llvm.vector.reduce.umin.v4i16(<4 x i16> undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V8 = call i16 @llvm.vector.reduce.umin.v8i16(<8 x i16> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V16 = call i16 @llvm.vector.reduce.umin.v16i16(<16 x i16> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 30 for instruction: %V32 = call i16 @llvm.vector.reduce.umin.v32i16(<32 x i16> undef)
-; SSSE3-NEXT: Cost Model: Found an estimated cost of 58 for instruction: %V64 = call i16 @llvm.vector.reduce.umin.v64i16(<64 x i16> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 15 for instruction: %V16 = call i16 @llvm.vector.reduce.umin.v16i16(<16 x i16> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 27 for instruction: %V32 = call i16 @llvm.vector.reduce.umin.v32i16(<32 x i16> undef)
+; SSSE3-NEXT: Cost Model: Found an estimated cost of 51 for instruction: %V64 = call i16 @llvm.vector.reduce.umin.v64i16(<64 x i16> undef)
; SSSE3-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; SSE4-LABEL: 'reduce_i16'
diff --git a/llvm/test/Analysis/CostModel/X86/vselect-cost.ll b/llvm/test/Analysis/CostModel/X86/vselect-cost.ll
index a55190e21c6fd..693d47bd5eeac 100644
--- a/llvm/test/Analysis/CostModel/X86/vselect-cost.ll
+++ b/llvm/test/Analysis/CostModel/X86/vselect-cost.ll
@@ -11,21 +11,21 @@
define i32 @test_select() {
; SSE2-LABEL: 'test_select'
; SSE2-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I64 = select i1 undef, i64 undef, i64 undef
-; SSE2-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V2I64 = select <2 x i1> undef, <2 x i64> undef, <2 x i64> undef
-; SSE2-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V4I64 = select <4 x i1> undef, <4 x i64> undef, <4 x i64> undef
-; SSE2-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %V8I64 = select <8 x i1> undef, <8 x i64> undef, <8 x i64> undef
+; SSE2-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2I64 = select <2 x i1> undef, <2 x i64> undef, <2 x i64> undef
+; SSE2-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V4I64 = select <4 x i1> undef, <4 x i64> undef, <4 x i64> undef
+; SSE2-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V8I64 = select <8 x i1> undef, <8 x i64> undef, <8 x i64> undef
; SSE2-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I32 = select i1 undef, i32 undef, i32 undef
-; SSE2-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V4I32 = select <4 x i1> undef, <4 x i32> undef, <4 x i32> undef
-; SSE2-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V8I32 = select <8 x i1> undef, <8 x i32> undef, <8 x i32> undef
-; SSE2-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %V16I32 = select <16 x i1> undef, <16 x i32> undef, <16 x i32> undef
+; SSE2-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V4I32 = select <4 x i1> undef, <4 x i32> undef, <4 x i32> undef
+; SSE2-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V8I32 = select <8 x i1> undef, <8 x i32> undef, <8 x i32> undef
+; SSE2-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V16I32 = select <16 x i1> undef, <16 x i32> undef, <16 x i32> undef
; SSE2-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I16 = select i1 undef, i16 undef, i16 undef
-; SSE2-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V8I16 = select <8 x i1> undef, <8 x i16> undef, <8 x i16> undef
-; SSE2-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V16I16 = select <16 x i1> undef, <16 x i16> undef, <16 x i16> undef
-; SSE2-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %V32I16 = select <32 x i1> undef, <32 x i16> undef, <32 x i16> undef
+; SSE2-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V8I16 = select <8 x i1> undef, <8 x i16> undef, <8 x i16> undef
+; SSE2-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V16I16 = select <16 x i1> undef, <16 x i16> undef, <16 x i16> undef
+; SSE2-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V32I16 = select <32 x i1> undef, <32 x i16> undef, <32 x i16> undef
; SSE2-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I8 = select i1 undef, i8 undef, i8 undef
-; SSE2-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V16I8 = select <16 x i1> undef, <16 x i8> undef, <16 x i8> undef
-; SSE2-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V32I8 = select <32 x i1> undef, <32 x i8> undef, <32 x i8> undef
-; SSE2-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %V64I8 = select <64 x i1> undef, <64 x i8> undef, <64 x i8> undef
+; SSE2-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V16I8 = select <16 x i1> undef, <16 x i8> undef, <16 x i8> undef
+; SSE2-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V32I8 = select <32 x i1> undef, <32 x i8> undef, <32 x i8> undef
+; SSE2-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V64I8 = select <64 x i1> undef, <64 x i8> undef, <64 x i8> undef
; SSE2-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; SSE41-LABEL: 'test_select'
@@ -58,12 +58,12 @@ define i32 @test_select() {
; AVX1-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V16I32 = select <16 x i1> undef, <16 x i32> undef, <16 x i32> undef
; AVX1-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I16 = select i1 undef, i16 undef, i16 undef
; AVX1-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V8I16 = select <8 x i1> undef, <8 x i16> undef, <8 x i16> undef
-; AVX1-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V16I16 = select <16 x i1> undef, <16 x i16> undef, <16 x i16> undef
-; AVX1-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V32I16 = select <32 x i1> undef, <32 x i16> undef, <32 x i16> undef
+; AVX1-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V16I16 = select <16 x i1> undef, <16 x i16> undef, <16 x i16> undef
+; AVX1-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V32I16 = select <32 x i1> undef, <32 x i16> undef, <32 x i16> undef
; AVX1-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I8 = select i1 undef, i8 undef, i8 undef
; AVX1-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V16I8 = select <16 x i1> undef, <16 x i8> undef, <16 x i8> undef
-; AVX1-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V32I8 = select <32 x i1> undef, <32 x i8> undef, <32 x i8> undef
-; AVX1-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V64I8 = select <64 x i1> undef, <64 x i8> undef, <64 x i8> undef
+; AVX1-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V32I8 = select <32 x i1> undef, <32 x i8> undef, <32 x i8> undef
+; AVX1-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V64I8 = select <64 x i1> undef, <64 x i8> undef, <64 x i8> undef
; AVX1-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; AVX2-LABEL: 'test_select'
@@ -151,7 +151,7 @@ define i32 @test_select() {
define <2 x i64> @test_2i64(<2 x i64> %a, <2 x i64> %b) {
; SSE2-LABEL: 'test_2i64'
-; SSE2-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %sel = select <2 x i1> <i1 true, i1 false>, <2 x i64> %a, <2 x i64> %b
+; SSE2-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %sel = select <2 x i1> <i1 true, i1 false>, <2 x i64> %a, <2 x i64> %b
; SSE2-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <2 x i64> %sel
;
; SSE41-LABEL: 'test_2i64'
@@ -168,7 +168,7 @@ define <2 x i64> @test_2i64(<2 x i64> %a, <2 x i64> %b) {
define <2 x double> @test_2double(<2 x double> %a, <2 x double> %b) {
; SSE2-LABEL: 'test_2double'
-; SSE2-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %sel = select <2 x i1> <i1 true, i1 false>, <2 x double> %a, <2 x double> %b
+; SSE2-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %sel = select <2 x i1> <i1 true, i1 false>, <2 x double> %a, <2 x double> %b
; SSE2-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <2 x double> %sel
;
; SSE41-LABEL: 'test_2double'
@@ -185,7 +185,7 @@ define <2 x double> @test_2double(<2 x double> %a, <2 x double> %b) {
define <4 x i32> @test_4i32(<4 x i32> %a, <4 x i32> %b) {
; SSE2-LABEL: 'test_4i32'
-; SSE2-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %sel = select <4 x i1> <i1 true, i1 false, i1 true, i1 false>, <4 x i32> %a, <4 x i32> %b
+; SSE2-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %sel = select <4 x i1> <i1 true, i1 false, i1 true, i1 false>, <4 x i32> %a, <4 x i32> %b
; SSE2-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <4 x i32> %sel
;
; SSE41-LABEL: 'test_4i32'
@@ -202,7 +202,7 @@ define <4 x i32> @test_4i32(<4 x i32> %a, <4 x i32> %b) {
define <4 x float> @test_4float(<4 x float> %a, <4 x float> %b) {
; SSE2-LABEL: 'test_4float'
-; SSE2-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %sel = select <4 x i1> <i1 true, i1 false, i1 true, i1 true>, <4 x float> %a, <4 x float> %b
+; SSE2-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %sel = select <4 x i1> <i1 true, i1 false, i1 true, i1 true>, <4 x float> %a, <4 x float> %b
; SSE2-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <4 x float> %sel
;
; SSE41-LABEL: 'test_4float'
@@ -219,7 +219,7 @@ define <4 x float> @test_4float(<4 x float> %a, <4 x float> %b) {
define <16 x i8> @test_16i8(<16 x i8> %a, <16 x i8> %b) {
; SSE2-LABEL: 'test_16i8'
-; SSE2-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %sel = select <16 x i1> <i1 true, i1 false, i1 true, i1 true, i1 true, i1 false, i1 true, i1 true, i1 true, i1 false, i1 true, i1 true, i1 true, i1 false, i1 true, i1 true>, <16 x i8> %a, <16 x i8> %b
+; SSE2-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %sel = select <16 x i1> <i1 true, i1 false, i1 true, i1 true, i1 true, i1 false, i1 true, i1 true, i1 true, i1 false, i1 true, i1 true, i1 true, i1 false, i1 true, i1 true>, <16 x i8> %a, <16 x i8> %b
; SSE2-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i8> %sel
;
; SSE41-LABEL: 'test_16i8'
@@ -239,7 +239,7 @@ define <16 x i8> @test_16i8(<16 x i8> %a, <16 x i8> %b) {
define <4 x i64> @test_4i64(<4 x i64> %a, <4 x i64> %b) {
; SSE2-LABEL: 'test_4i64'
-; SSE2-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %sel = select <4 x i1> <i1 true, i1 false, i1 false, i1 true>, <4 x i64> %a, <4 x i64> %b
+; SSE2-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %sel = select <4 x i1> <i1 true, i1 false, i1 false, i1 true>, <4 x i64> %a, <4 x i64> %b
; SSE2-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <4 x i64> %sel
;
; SSE41-LABEL: 'test_4i64'
@@ -256,7 +256,7 @@ define <4 x i64> @test_4i64(<4 x i64> %a, <4 x i64> %b) {
define <4 x double> @test_4double(<4 x double> %a, <4 x double> %b) {
; SSE2-LABEL: 'test_4double'
-; SSE2-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %sel = select <4 x i1> <i1 true, i1 false, i1 true, i1 false>, <4 x double> %a, <4 x double> %b
+; SSE2-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %sel = select <4 x i1> <i1 true, i1 false, i1 true, i1 false>, <4 x double> %a, <4 x double> %b
; SSE2-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <4 x double> %sel
;
; SSE41-LABEL: 'test_4double'
@@ -273,7 +273,7 @@ define <4 x double> @test_4double(<4 x double> %a, <4 x double> %b) {
define <8 x i32> @test_8i32(<8 x i32> %a, <8 x i32> %b) {
; SSE2-LABEL: 'test_8i32'
-; SSE2-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %sel = select <8 x i1> <i1 true, i1 false, i1 true, i1 true, i1 true, i1 false, i1 true, i1 false>, <8 x i32> %a, <8 x i32> %b
+; SSE2-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %sel = select <8 x i1> <i1 true, i1 false, i1 true, i1 true, i1 true, i1 false, i1 true, i1 false>, <8 x i32> %a, <8 x i32> %b
; SSE2-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <8 x i32> %sel
;
; SSE41-LABEL: 'test_8i32'
@@ -290,7 +290,7 @@ define <8 x i32> @test_8i32(<8 x i32> %a, <8 x i32> %b) {
define <8 x float> @test_8float(<8 x float> %a, <8 x float> %b) {
; SSE2-LABEL: 'test_8float'
-; SSE2-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %sel = select <8 x i1> <i1 true, i1 false, i1 false, i1 false, i1 true, i1 false, i1 false, i1 false>, <8 x float> %a, <8 x float> %b
+; SSE2-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %sel = select <8 x i1> <i1 true, i1 false, i1 false, i1 false, i1 true, i1 false, i1 false, i1 false>, <8 x float> %a, <8 x float> %b
; SSE2-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <8 x float> %sel
;
; SSE41-LABEL: 'test_8float'
@@ -307,7 +307,7 @@ define <8 x float> @test_8float(<8 x float> %a, <8 x float> %b) {
define <16 x i16> @test_16i16(<16 x i16> %a, <16 x i16> %b) {
; SSE2-LABEL: 'test_16i16'
-; SSE2-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %sel = select <16 x i1> <i1 true, i1 false, i1 false, i1 false, i1 true, i1 false, i1 false, i1 false, i1 true, i1 false, i1 false, i1 false, i1 true, i1 false, i1 false, i1 false>, <16 x i16> %a, <16 x i16> %b
+; SSE2-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %sel = select <16 x i1> <i1 true, i1 false, i1 false, i1 false, i1 true, i1 false, i1 false, i1 false, i1 true, i1 false, i1 false, i1 false, i1 true, i1 false, i1 false, i1 false>, <16 x i16> %a, <16 x i16> %b
; SSE2-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i16> %sel
;
; SSE41-LABEL: 'test_16i16'
@@ -315,7 +315,7 @@ define <16 x i16> @test_16i16(<16 x i16> %a, <16 x i16> %b) {
; SSE41-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i16> %sel
;
; AVX1-LABEL: 'test_16i16'
-; AVX1-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %sel = select <16 x i1> <i1 true, i1 false, i1 false, i1 false, i1 true, i1 false, i1 false, i1 false, i1 true, i1 false, i1 false, i1 false, i1 true, i1 false, i1 false, i1 false>, <16 x i16> %a, <16 x i16> %b
+; AVX1-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %sel = select <16 x i1> <i1 true, i1 false, i1 false, i1 false, i1 true, i1 false, i1 false, i1 false, i1 true, i1 false, i1 false, i1 false, i1 true, i1 false, i1 false, i1 false>, <16 x i16> %a, <16 x i16> %b
; AVX1-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <16 x i16> %sel
;
; AVX2-LABEL: 'test_16i16'
@@ -332,7 +332,7 @@ define <16 x i16> @test_16i16(<16 x i16> %a, <16 x i16> %b) {
define <32 x i8> @test_32i8(<32 x i8> %a, <32 x i8> %b) {
; SSE2-LABEL: 'test_32i8'
-; SSE2-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %sel = select <32 x i1> <i1 true, i1 false, i1 true, i1 true, i1 true, i1 false, i1 true, i1 true, i1 true, i1 false, i1 true, i1 true, i1 true, i1 false, i1 true, i1 true, i1 true, i1 false, i1 true, i1 true, i1 true, i1 false, i1 true, i1 true, i1 true, i1 false, i1 true, i1 true, i1 true, i1 false, i1 true, i1 true>, <32 x i8> %a, <32 x i8> %b
+; SSE2-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %sel = select <32 x i1> <i1 true, i1 false, i1 true, i1 true, i1 true, i1 false, i1 true, i1 true, i1 true, i1 false, i1 true, i1 true, i1 true, i1 false, i1 true, i1 true, i1 true, i1 false, i1 true, i1 true, i1 true, i1 false, i1 true, i1 true, i1 true, i1 false, i1 true, i1 true, i1 true, i1 false, i1 true, i1 true>, <32 x i8> %a, <32 x i8> %b
; SSE2-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <32 x i8> %sel
;
; SSE41-LABEL: 'test_32i8'
@@ -340,7 +340,7 @@ define <32 x i8> @test_32i8(<32 x i8> %a, <32 x i8> %b) {
; SSE41-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <32 x i8> %sel
;
; AVX1-LABEL: 'test_32i8'
-; AVX1-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %sel = select <32 x i1> <i1 true, i1 false, i1 true, i1 true, i1 true, i1 false, i1 true, i1 true, i1 true, i1 false, i1 true, i1 true, i1 true, i1 false, i1 true, i1 true, i1 true, i1 false, i1 true, i1 true, i1 true, i1 false, i1 true, i1 true, i1 true, i1 false, i1 true, i1 true, i1 true, i1 false, i1 true, i1 true>, <32 x i8> %a, <32 x i8> %b
+; AVX1-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %sel = select <32 x i1> <i1 true, i1 false, i1 true, i1 true, i1 true, i1 false, i1 true, i1 true, i1 true, i1 false, i1 true, i1 true, i1 true, i1 false, i1 true, i1 true, i1 true, i1 false, i1 true, i1 true, i1 true, i1 false, i1 true, i1 true, i1 true, i1 false, i1 true, i1 true, i1 true, i1 false, i1 true, i1 true>, <32 x i8> %a, <32 x i8> %b
; AVX1-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret <32 x i8> %sel
;
; AVX2-LABEL: 'test_32i8'
diff --git a/llvm/test/Transforms/LoopVectorize/X86/imprecise-through-phis.ll b/llvm/test/Transforms/LoopVectorize/X86/imprecise-through-phis.ll
index e909403df52f3..096286d09ca2f 100644
--- a/llvm/test/Transforms/LoopVectorize/X86/imprecise-through-phis.ll
+++ b/llvm/test/Transforms/LoopVectorize/X86/imprecise-through-phis.ll
@@ -67,11 +67,36 @@ done:
define double @sumIfVector(double* nocapture readonly %arr) {
; SSE-LABEL: @sumIfVector(
; SSE-NEXT: entry:
+; SSE-NEXT: br i1 false, label [[SCALAR_PH:%.*]], label [[VECTOR_PH:%.*]]
+; SSE: vector.ph:
+; SSE-NEXT: br label [[VECTOR_BODY:%.*]]
+; SSE: vector.body:
+; SSE-NEXT: [[INDEX:%.*]] = phi i32 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
+; SSE-NEXT: [[VEC_PHI:%.*]] = phi <2 x double> [ zeroinitializer, [[VECTOR_PH]] ], [ [[PREDPHI:%.*]], [[VECTOR_BODY]] ]
+; SSE-NEXT: [[TMP0:%.*]] = add i32 [[INDEX]], 0
+; SSE-NEXT: [[TMP1:%.*]] = getelementptr double, double* [[ARR:%.*]], i32 [[TMP0]]
+; SSE-NEXT: [[TMP2:%.*]] = getelementptr double, double* [[TMP1]], i32 0
+; SSE-NEXT: [[TMP3:%.*]] = bitcast double* [[TMP2]] to <2 x double>*
+; SSE-NEXT: [[WIDE_LOAD:%.*]] = load <2 x double>, <2 x double>* [[TMP3]], align 8
+; SSE-NEXT: [[TMP4:%.*]] = fcmp fast une <2 x double> [[WIDE_LOAD]], <double 4.200000e+01, double 4.200000e+01>
+; SSE-NEXT: [[TMP5:%.*]] = fadd fast <2 x double> [[VEC_PHI]], [[WIDE_LOAD]]
+; SSE-NEXT: [[TMP6:%.*]] = xor <2 x i1> [[TMP4]], <i1 true, i1 true>
+; SSE-NEXT: [[PREDPHI]] = select <2 x i1> [[TMP4]], <2 x double> [[TMP5]], <2 x double> [[VEC_PHI]]
+; SSE-NEXT: [[INDEX_NEXT]] = add nuw i32 [[INDEX]], 2
+; SSE-NEXT: [[TMP7:%.*]] = icmp eq i32 [[INDEX_NEXT]], 32
+; SSE-NEXT: br i1 [[TMP7]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
+; SSE: middle.block:
+; SSE-NEXT: [[TMP8:%.*]] = call fast double @llvm.vector.reduce.fadd.v2f64(double -0.000000e+00, <2 x double> [[PREDPHI]])
+; SSE-NEXT: [[CMP_N:%.*]] = icmp eq i32 32, 32
+; SSE-NEXT: br i1 [[CMP_N]], label [[DONE:%.*]], label [[SCALAR_PH]]
+; SSE: scalar.ph:
+; SSE-NEXT: [[BC_RESUME_VAL:%.*]] = phi i32 [ 32, [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY:%.*]] ]
+; SSE-NEXT: [[BC_MERGE_RDX:%.*]] = phi double [ 0.000000e+00, [[ENTRY]] ], [ [[TMP8]], [[MIDDLE_BLOCK]] ]
; SSE-NEXT: br label [[LOOP:%.*]]
; SSE: loop:
-; SSE-NEXT: [[I:%.*]] = phi i32 [ 0, [[ENTRY:%.*]] ], [ [[I_NEXT:%.*]], [[NEXT_ITER:%.*]] ]
-; SSE-NEXT: [[TOT:%.*]] = phi double [ 0.000000e+00, [[ENTRY]] ], [ [[TOT_NEXT:%.*]], [[NEXT_ITER]] ]
-; SSE-NEXT: [[ADDR:%.*]] = getelementptr double, double* [[ARR:%.*]], i32 [[I]]
+; SSE-NEXT: [[I:%.*]] = phi i32 [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], [ [[I_NEXT:%.*]], [[NEXT_ITER:%.*]] ]
+; SSE-NEXT: [[TOT:%.*]] = phi double [ [[BC_MERGE_RDX]], [[SCALAR_PH]] ], [ [[TOT_NEXT:%.*]], [[NEXT_ITER]] ]
+; SSE-NEXT: [[ADDR:%.*]] = getelementptr double, double* [[ARR]], i32 [[I]]
; SSE-NEXT: [[NEXTVAL:%.*]] = load double, double* [[ADDR]], align 8
; SSE-NEXT: [[TST:%.*]] = fcmp fast une double [[NEXTVAL]], 4.200000e+01
; SSE-NEXT: br i1 [[TST]], label [[DO_ADD:%.*]], label [[NO_ADD:%.*]]
@@ -84,9 +109,9 @@ define double @sumIfVector(double* nocapture readonly %arr) {
; SSE-NEXT: [[TOT_NEXT]] = phi double [ [[TOT]], [[NO_ADD]] ], [ [[TOT_NEW]], [[DO_ADD]] ]
; SSE-NEXT: [[I_NEXT]] = add i32 [[I]], 1
; SSE-NEXT: [[AGAIN:%.*]] = icmp ult i32 [[I_NEXT]], 32
-; SSE-NEXT: br i1 [[AGAIN]], label [[LOOP]], label [[DONE:%.*]]
+; SSE-NEXT: br i1 [[AGAIN]], label [[LOOP]], label [[DONE]], !llvm.loop [[LOOP2:![0-9]+]]
; SSE: done:
-; SSE-NEXT: [[TOT_NEXT_LCSSA:%.*]] = phi double [ [[TOT_NEXT]], [[NEXT_ITER]] ]
+; SSE-NEXT: [[TOT_NEXT_LCSSA:%.*]] = phi double [ [[TOT_NEXT]], [[NEXT_ITER]] ], [ [[TMP8]], [[MIDDLE_BLOCK]] ]
; SSE-NEXT: ret double [[TOT_NEXT_LCSSA]]
;
; AVX-LABEL: @sumIfVector(
@@ -108,7 +133,7 @@ define double @sumIfVector(double* nocapture readonly %arr) {
; AVX-NEXT: [[PREDPHI]] = select <4 x i1> [[TMP4]], <4 x double> [[TMP5]], <4 x double> [[VEC_PHI]]
; AVX-NEXT: [[INDEX_NEXT]] = add nuw i32 [[INDEX]], 4
; AVX-NEXT: [[TMP7:%.*]] = icmp eq i32 [[INDEX_NEXT]], 32
-; AVX-NEXT: br i1 [[TMP7]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], [[LOOP0:!llvm.loop !.*]]
+; AVX-NEXT: br i1 [[TMP7]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
; AVX: middle.block:
; AVX-NEXT: [[TMP8:%.*]] = call fast double @llvm.vector.reduce.fadd.v4f64(double -0.000000e+00, <4 x double> [[PREDPHI]])
; AVX-NEXT: [[CMP_N:%.*]] = icmp eq i32 32, 32
@@ -133,7 +158,7 @@ define double @sumIfVector(double* nocapture readonly %arr) {
; AVX-NEXT: [[TOT_NEXT]] = phi double [ [[TOT]], [[NO_ADD]] ], [ [[TOT_NEW]], [[DO_ADD]] ]
; AVX-NEXT: [[I_NEXT]] = add i32 [[I]], 1
; AVX-NEXT: [[AGAIN:%.*]] = icmp ult i32 [[I_NEXT]], 32
-; AVX-NEXT: br i1 [[AGAIN]], label [[LOOP]], label [[DONE]], [[LOOP2:!llvm.loop !.*]]
+; AVX-NEXT: br i1 [[AGAIN]], label [[LOOP]], label [[DONE]], !llvm.loop [[LOOP2:![0-9]+]]
; AVX: done:
; AVX-NEXT: [[TOT_NEXT_LCSSA:%.*]] = phi double [ [[TOT_NEXT]], [[NEXT_ITER]] ], [ [[TMP8]], [[MIDDLE_BLOCK]] ]
; AVX-NEXT: ret double [[TOT_NEXT_LCSSA]]
diff --git a/llvm/test/Transforms/SLPVectorizer/X86/crash_cmpop.ll b/llvm/test/Transforms/SLPVectorizer/X86/crash_cmpop.ll
index 7b914b43180ef..2e84c104adbb9 100644
--- a/llvm/test/Transforms/SLPVectorizer/X86/crash_cmpop.ll
+++ b/llvm/test/Transforms/SLPVectorizer/X86/crash_cmpop.ll
@@ -12,38 +12,35 @@ define void @testfunc(float* nocapture %dest, float* nocapture readonly %src) {
; SSE: for.body:
; SSE-NEXT: [[INDVARS_IV:%.*]] = phi i64 [ 0, [[ENTRY:%.*]] ], [ [[INDVARS_IV_NEXT:%.*]], [[FOR_BODY]] ]
; SSE-NEXT: [[ACC1_056:%.*]] = phi float [ 0.000000e+00, [[ENTRY]] ], [ [[ADD13:%.*]], [[FOR_BODY]] ]
-; SSE-NEXT: [[S1_055:%.*]] = phi float [ 0.000000e+00, [[ENTRY]] ], [ [[COND_I40:%.*]], [[FOR_BODY]] ]
-; SSE-NEXT: [[S0_054:%.*]] = phi float [ 0.000000e+00, [[ENTRY]] ], [ [[COND_I44:%.*]], [[FOR_BODY]] ]
+; SSE-NEXT: [[TMP0:%.*]] = phi <2 x float> [ zeroinitializer, [[ENTRY]] ], [ [[TMP23:%.*]], [[FOR_BODY]] ]
; SSE-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds float, float* [[SRC:%.*]], i64 [[INDVARS_IV]]
-; SSE-NEXT: [[TMP0:%.*]] = load float, float* [[ARRAYIDX]], align 4
+; SSE-NEXT: [[TMP1:%.*]] = load float, float* [[ARRAYIDX]], align 4
; SSE-NEXT: [[INDVARS_IV_NEXT]] = add nuw nsw i64 [[INDVARS_IV]], 1
; SSE-NEXT: [[ARRAYIDX2:%.*]] = getelementptr inbounds float, float* [[DEST:%.*]], i64 [[INDVARS_IV]]
; SSE-NEXT: store float [[ACC1_056]], float* [[ARRAYIDX2]], align 4
-; SSE-NEXT: [[ADD:%.*]] = fadd float [[S0_054]], [[TMP0]]
-; SSE-NEXT: [[ADD3:%.*]] = fadd float [[S1_055]], [[TMP0]]
-; SSE-NEXT: [[MUL:%.*]] = fmul float [[S0_054]], 0.000000e+00
-; SSE-NEXT: [[ADD4:%.*]] = fadd float [[MUL]], [[ADD3]]
-; SSE-NEXT: [[MUL5:%.*]] = fmul float [[S1_055]], 0.000000e+00
-; SSE-NEXT: [[ADD6:%.*]] = fadd float [[MUL5]], [[ADD]]
-; SSE-NEXT: [[CMP_I:%.*]] = fcmp olt float [[ADD6]], 1.000000e+00
-; SSE-NEXT: [[COND_I:%.*]] = select i1 [[CMP_I]], float [[ADD6]], float 1.000000e+00
-; SSE-NEXT: [[CMP_I51:%.*]] = fcmp olt float [[COND_I]], -1.000000e+00
-; SSE-NEXT: [[CMP_I49:%.*]] = fcmp olt float [[ADD4]], 1.000000e+00
-; SSE-NEXT: [[COND_I50:%.*]] = select i1 [[CMP_I49]], float [[ADD4]], float 1.000000e+00
-; SSE-NEXT: [[CMP_I47:%.*]] = fcmp olt float [[COND_I50]], -1.000000e+00
-; SSE-NEXT: [[COND_I_OP:%.*]] = fmul float [[COND_I]], 0.000000e+00
-; SSE-NEXT: [[MUL10:%.*]] = select i1 [[CMP_I51]], float -0.000000e+00, float [[COND_I_OP]]
-; SSE-NEXT: [[COND_I50_OP:%.*]] = fmul float [[COND_I50]], 0.000000e+00
-; SSE-NEXT: [[MUL11:%.*]] = select i1 [[CMP_I47]], float -0.000000e+00, float [[COND_I50_OP]]
-; SSE-NEXT: [[ADD13]] = fadd float [[MUL10]], [[MUL11]]
-; SSE-NEXT: [[CMP_I45:%.*]] = fcmp olt float [[ADD13]], 1.000000e+00
-; SSE-NEXT: [[COND_I46:%.*]] = select i1 [[CMP_I45]], float [[ADD13]], float 1.000000e+00
-; SSE-NEXT: [[CMP_I43:%.*]] = fcmp olt float [[COND_I46]], -1.000000e+00
-; SSE-NEXT: [[COND_I44]] = select i1 [[CMP_I43]], float -1.000000e+00, float [[COND_I46]]
-; SSE-NEXT: [[CMP_I41:%.*]] = fcmp olt float [[MUL11]], 1.000000e+00
-; SSE-NEXT: [[COND_I42:%.*]] = select i1 [[CMP_I41]], float [[MUL11]], float 1.000000e+00
-; SSE-NEXT: [[CMP_I39:%.*]] = fcmp olt float [[COND_I42]], -1.000000e+00
-; SSE-NEXT: [[COND_I40]] = select i1 [[CMP_I39]], float -1.000000e+00, float [[COND_I42]]
+; SSE-NEXT: [[TMP2:%.*]] = fmul <2 x float> [[TMP0]], zeroinitializer
+; SSE-NEXT: [[TMP3:%.*]] = extractelement <2 x float> [[TMP0]], i32 1
+; SSE-NEXT: [[TMP4:%.*]] = insertelement <2 x float> poison, float [[TMP3]], i32 0
+; SSE-NEXT: [[TMP5:%.*]] = extractelement <2 x float> [[TMP0]], i32 0
+; SSE-NEXT: [[TMP6:%.*]] = insertelement <2 x float> [[TMP4]], float [[TMP5]], i32 1
+; SSE-NEXT: [[TMP7:%.*]] = insertelement <2 x float> poison, float [[TMP1]], i32 0
+; SSE-NEXT: [[TMP8:%.*]] = insertelement <2 x float> [[TMP7]], float [[TMP1]], i32 1
+; SSE-NEXT: [[TMP9:%.*]] = fadd <2 x float> [[TMP6]], [[TMP8]]
+; SSE-NEXT: [[TMP10:%.*]] = fadd <2 x float> [[TMP2]], [[TMP9]]
+; SSE-NEXT: [[TMP11:%.*]] = fcmp olt <2 x float> [[TMP10]], <float 1.000000e+00, float 1.000000e+00>
+; SSE-NEXT: [[TMP12:%.*]] = select <2 x i1> [[TMP11]], <2 x float> [[TMP10]], <2 x float> <float 1.000000e+00, float 1.000000e+00>
+; SSE-NEXT: [[TMP13:%.*]] = fcmp olt <2 x float> [[TMP12]], <float -1.000000e+00, float -1.000000e+00>
+; SSE-NEXT: [[TMP14:%.*]] = fmul <2 x float> [[TMP12]], zeroinitializer
+; SSE-NEXT: [[TMP15:%.*]] = select <2 x i1> [[TMP13]], <2 x float> <float -0.000000e+00, float -0.000000e+00>, <2 x float> [[TMP14]]
+; SSE-NEXT: [[TMP16:%.*]] = extractelement <2 x float> [[TMP15]], i32 0
+; SSE-NEXT: [[TMP17:%.*]] = extractelement <2 x float> [[TMP15]], i32 1
+; SSE-NEXT: [[ADD13]] = fadd float [[TMP16]], [[TMP17]]
+; SSE-NEXT: [[TMP18:%.*]] = insertelement <2 x float> poison, float [[TMP17]], i32 0
+; SSE-NEXT: [[TMP19:%.*]] = insertelement <2 x float> [[TMP18]], float [[ADD13]], i32 1
+; SSE-NEXT: [[TMP20:%.*]] = fcmp olt <2 x float> [[TMP19]], <float 1.000000e+00, float 1.000000e+00>
+; SSE-NEXT: [[TMP21:%.*]] = select <2 x i1> [[TMP20]], <2 x float> [[TMP19]], <2 x float> <float 1.000000e+00, float 1.000000e+00>
+; SSE-NEXT: [[TMP22:%.*]] = fcmp olt <2 x float> [[TMP21]], <float -1.000000e+00, float -1.000000e+00>
+; SSE-NEXT: [[TMP23]] = select <2 x i1> [[TMP22]], <2 x float> <float -1.000000e+00, float -1.000000e+00>, <2 x float> [[TMP21]]
; SSE-NEXT: [[EXITCOND:%.*]] = icmp eq i64 [[INDVARS_IV_NEXT]], 32
; SSE-NEXT: br i1 [[EXITCOND]], label [[FOR_END:%.*]], label [[FOR_BODY]]
; SSE: for.end:
More information about the llvm-commits
mailing list