[llvm] b298308 - [CostModel][X86] fptosi/fptoui to i8/i16 are truncated from fptosi to i32
Simon Pilgrim via llvm-commits
llvm-commits at lists.llvm.org
Tue Jul 6 09:40:16 PDT 2021
Author: Simon Pilgrim
Date: 2021-07-06T17:28:03+01:00
New Revision: b298308ba2544a00c3e2ae0c1cc06fa058a4000c
URL: https://github.com/llvm/llvm-project/commit/b298308ba2544a00c3e2ae0c1cc06fa058a4000c
DIFF: https://github.com/llvm/llvm-project/commit/b298308ba2544a00c3e2ae0c1cc06fa058a4000c.diff
LOG: [CostModel][X86] fptosi/fptoui to i8/i16 are truncated from fptosi to i32
Provide a generic fallback that performs the fptosi to i32 types, then truncates to sub-i32 scalars.
These numbers can be tweaked for specific sse levels, but we should get the default handling in place first.
Added:
Modified:
llvm/lib/Target/X86/X86TargetTransformInfo.cpp
llvm/test/Analysis/CostModel/X86/fptosi.ll
llvm/test/Analysis/CostModel/X86/fptoui.ll
Removed:
################################################################################
diff --git a/llvm/lib/Target/X86/X86TargetTransformInfo.cpp b/llvm/lib/Target/X86/X86TargetTransformInfo.cpp
index cf767fc1c7e4..87b96b699928 100644
--- a/llvm/lib/Target/X86/X86TargetTransformInfo.cpp
+++ b/llvm/lib/Target/X86/X86TargetTransformInfo.cpp
@@ -2315,6 +2315,16 @@ InstructionCost X86TTIImpl::getCastInstrCost(unsigned Opcode, Type *Dst,
TTI::CastContextHint::None, CostKind);
}
+ // Fallback for fptosi/fptoui i8/i16 cases we need to truncate from fptosi
+ // i32.
+ if ((ISD == ISD::FP_TO_SINT || ISD == ISD::FP_TO_UINT) &&
+ 1 < Dst->getScalarSizeInBits() && Dst->getScalarSizeInBits() < 32) {
+ Type *TruncDst = Dst->getWithNewBitWidth(32);
+ return getCastInstrCost(Instruction::FPToSI, TruncDst, Src, CCH, CostKind) +
+ getCastInstrCost(Instruction::Trunc, Dst, TruncDst,
+ TTI::CastContextHint::None, CostKind);
+ }
+
return AdjustCost(
BaseT::getCastInstrCost(Opcode, Dst, Src, CCH, CostKind, I));
}
diff --git a/llvm/test/Analysis/CostModel/X86/fptosi.ll b/llvm/test/Analysis/CostModel/X86/fptosi.ll
index d92310cb79f8..4c9aa6cafc8e 100644
--- a/llvm/test/Analysis/CostModel/X86/fptosi.ll
+++ b/llvm/test/Analysis/CostModel/X86/fptosi.ll
@@ -97,19 +97,33 @@ define i32 @fptosi_double_i32(i32 %arg) {
}
define i32 @fptosi_double_i16(i32 %arg) {
-; SSE-LABEL: 'fptosi_double_i16'
-; SSE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I16 = fptosi double undef to i16
-; SSE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2I16 = fptosi <2 x double> undef to <2 x i16>
-; SSE-NEXT: Cost Model: Found an estimated cost of 5 for instruction: %V4I16 = fptosi <4 x double> undef to <4 x i16>
-; SSE-NEXT: Cost Model: Found an estimated cost of 11 for instruction: %V8I16 = fptosi <8 x double> undef to <8 x i16>
-; SSE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+; SSE2-LABEL: 'fptosi_double_i16'
+; SSE2-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I16 = fptosi double undef to i16
+; SSE2-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2I16 = fptosi <2 x double> undef to <2 x i16>
+; SSE2-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V4I16 = fptosi <4 x double> undef to <4 x i16>
+; SSE2-NEXT: Cost Model: Found an estimated cost of 11 for instruction: %V8I16 = fptosi <8 x double> undef to <8 x i16>
+; SSE2-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
-; AVX-LABEL: 'fptosi_double_i16'
-; AVX-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I16 = fptosi double undef to i16
-; AVX-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2I16 = fptosi <2 x double> undef to <2 x i16>
-; AVX-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V4I16 = fptosi <4 x double> undef to <4 x i16>
-; AVX-NEXT: Cost Model: Found an estimated cost of 5 for instruction: %V8I16 = fptosi <8 x double> undef to <8 x i16>
-; AVX-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+; SSE42-LABEL: 'fptosi_double_i16'
+; SSE42-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I16 = fptosi double undef to i16
+; SSE42-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2I16 = fptosi <2 x double> undef to <2 x i16>
+; SSE42-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V4I16 = fptosi <4 x double> undef to <4 x i16>
+; SSE42-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V8I16 = fptosi <8 x double> undef to <8 x i16>
+; SSE42-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+;
+; AVX1-LABEL: 'fptosi_double_i16'
+; AVX1-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I16 = fptosi double undef to i16
+; AVX1-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2I16 = fptosi <2 x double> undef to <2 x i16>
+; AVX1-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V4I16 = fptosi <4 x double> undef to <4 x i16>
+; AVX1-NEXT: Cost Model: Found an estimated cost of 10 for instruction: %V8I16 = fptosi <8 x double> undef to <8 x i16>
+; AVX1-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+;
+; AVX2-LABEL: 'fptosi_double_i16'
+; AVX2-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I16 = fptosi double undef to i16
+; AVX2-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2I16 = fptosi <2 x double> undef to <2 x i16>
+; AVX2-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V4I16 = fptosi <4 x double> undef to <4 x i16>
+; AVX2-NEXT: Cost Model: Found an estimated cost of 5 for instruction: %V8I16 = fptosi <8 x double> undef to <8 x i16>
+; AVX2-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; AVX512-LABEL: 'fptosi_double_i16'
; AVX512-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I16 = fptosi double undef to i16
@@ -117,6 +131,13 @@ define i32 @fptosi_double_i16(i32 %arg) {
; AVX512-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V4I16 = fptosi <4 x double> undef to <4 x i16>
; AVX512-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V8I16 = fptosi <8 x double> undef to <8 x i16>
; AVX512-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+;
+; SLM-LABEL: 'fptosi_double_i16'
+; SLM-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I16 = fptosi double undef to i16
+; SLM-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2I16 = fptosi <2 x double> undef to <2 x i16>
+; SLM-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V4I16 = fptosi <4 x double> undef to <4 x i16>
+; SLM-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V8I16 = fptosi <8 x double> undef to <8 x i16>
+; SLM-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
%I16 = fptosi double undef to i16
%V2I16 = fptosi <2 x double> undef to <2 x i16>
@@ -129,23 +150,30 @@ define i32 @fptosi_double_i8(i32 %arg) {
; SSE2-LABEL: 'fptosi_double_i8'
; SSE2-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I8 = fptosi double undef to i8
; SSE2-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2I8 = fptosi <2 x double> undef to <2 x i8>
-; SSE2-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V4I8 = fptosi <4 x double> undef to <4 x i8>
-; SSE2-NEXT: Cost Model: Found an estimated cost of 19 for instruction: %V8I8 = fptosi <8 x double> undef to <8 x i8>
+; SSE2-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V4I8 = fptosi <4 x double> undef to <4 x i8>
+; SSE2-NEXT: Cost Model: Found an estimated cost of 10 for instruction: %V8I8 = fptosi <8 x double> undef to <8 x i8>
; SSE2-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; SSE42-LABEL: 'fptosi_double_i8'
; SSE42-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I8 = fptosi double undef to i8
; SSE42-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V2I8 = fptosi <2 x double> undef to <2 x i8>
-; SSE42-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V4I8 = fptosi <4 x double> undef to <4 x i8>
-; SSE42-NEXT: Cost Model: Found an estimated cost of 15 for instruction: %V8I8 = fptosi <8 x double> undef to <8 x i8>
+; SSE42-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V4I8 = fptosi <4 x double> undef to <4 x i8>
+; SSE42-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V8I8 = fptosi <8 x double> undef to <8 x i8>
; SSE42-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
-; AVX-LABEL: 'fptosi_double_i8'
-; AVX-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I8 = fptosi double undef to i8
-; AVX-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V2I8 = fptosi <2 x double> undef to <2 x i8>
-; AVX-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V4I8 = fptosi <4 x double> undef to <4 x i8>
-; AVX-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V8I8 = fptosi <8 x double> undef to <8 x i8>
-; AVX-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+; AVX1-LABEL: 'fptosi_double_i8'
+; AVX1-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I8 = fptosi double undef to i8
+; AVX1-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V2I8 = fptosi <2 x double> undef to <2 x i8>
+; AVX1-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V4I8 = fptosi <4 x double> undef to <4 x i8>
+; AVX1-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V8I8 = fptosi <8 x double> undef to <8 x i8>
+; AVX1-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+;
+; AVX2-LABEL: 'fptosi_double_i8'
+; AVX2-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I8 = fptosi double undef to i8
+; AVX2-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V2I8 = fptosi <2 x double> undef to <2 x i8>
+; AVX2-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V4I8 = fptosi <4 x double> undef to <4 x i8>
+; AVX2-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V8I8 = fptosi <8 x double> undef to <8 x i8>
+; AVX2-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; AVX512-LABEL: 'fptosi_double_i8'
; AVX512-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I8 = fptosi double undef to i8
@@ -157,8 +185,8 @@ define i32 @fptosi_double_i8(i32 %arg) {
; SLM-LABEL: 'fptosi_double_i8'
; SLM-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I8 = fptosi double undef to i8
; SLM-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V2I8 = fptosi <2 x double> undef to <2 x i8>
-; SLM-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V4I8 = fptosi <4 x double> undef to <4 x i8>
-; SLM-NEXT: Cost Model: Found an estimated cost of 15 for instruction: %V8I8 = fptosi <8 x double> undef to <8 x i8>
+; SLM-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V4I8 = fptosi <4 x double> undef to <4 x i8>
+; SLM-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V8I8 = fptosi <8 x double> undef to <8 x i8>
; SLM-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
%I8 = fptosi double undef to i8
@@ -267,21 +295,37 @@ define i32 @fptosi_float_i32(i32 %arg) {
}
define i32 @fptosi_float_i16(i32 %arg) {
-; SSE-LABEL: 'fptosi_float_i16'
-; SSE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I16 = fptosi float undef to i16
-; SSE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2I16 = fptosi <2 x float> undef to <2 x i16>
-; SSE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V4I16 = fptosi <4 x float> undef to <4 x i16>
-; SSE-NEXT: Cost Model: Found an estimated cost of 5 for instruction: %V8I16 = fptosi <8 x float> undef to <8 x i16>
-; SSE-NEXT: Cost Model: Found an estimated cost of 10 for instruction: %V16I16 = fptosi <16 x float> undef to <16 x i16>
-; SSE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+; SSE2-LABEL: 'fptosi_float_i16'
+; SSE2-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I16 = fptosi float undef to i16
+; SSE2-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2I16 = fptosi <2 x float> undef to <2 x i16>
+; SSE2-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V4I16 = fptosi <4 x float> undef to <4 x i16>
+; SSE2-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V8I16 = fptosi <8 x float> undef to <8 x i16>
+; SSE2-NEXT: Cost Model: Found an estimated cost of 14 for instruction: %V16I16 = fptosi <16 x float> undef to <16 x i16>
+; SSE2-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
-; AVX-LABEL: 'fptosi_float_i16'
-; AVX-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I16 = fptosi float undef to i16
-; AVX-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2I16 = fptosi <2 x float> undef to <2 x i16>
-; AVX-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V4I16 = fptosi <4 x float> undef to <4 x i16>
-; AVX-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V8I16 = fptosi <8 x float> undef to <8 x i16>
-; AVX-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V16I16 = fptosi <16 x float> undef to <16 x i16>
-; AVX-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+; SSE42-LABEL: 'fptosi_float_i16'
+; SSE42-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I16 = fptosi float undef to i16
+; SSE42-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2I16 = fptosi <2 x float> undef to <2 x i16>
+; SSE42-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V4I16 = fptosi <4 x float> undef to <4 x i16>
+; SSE42-NEXT: Cost Model: Found an estimated cost of 5 for instruction: %V8I16 = fptosi <8 x float> undef to <8 x i16>
+; SSE42-NEXT: Cost Model: Found an estimated cost of 10 for instruction: %V16I16 = fptosi <16 x float> undef to <16 x i16>
+; SSE42-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+;
+; AVX1-LABEL: 'fptosi_float_i16'
+; AVX1-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I16 = fptosi float undef to i16
+; AVX1-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2I16 = fptosi <2 x float> undef to <2 x i16>
+; AVX1-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V4I16 = fptosi <4 x float> undef to <4 x i16>
+; AVX1-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V8I16 = fptosi <8 x float> undef to <8 x i16>
+; AVX1-NEXT: Cost Model: Found an estimated cost of 10 for instruction: %V16I16 = fptosi <16 x float> undef to <16 x i16>
+; AVX1-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+;
+; AVX2-LABEL: 'fptosi_float_i16'
+; AVX2-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I16 = fptosi float undef to i16
+; AVX2-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2I16 = fptosi <2 x float> undef to <2 x i16>
+; AVX2-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V4I16 = fptosi <4 x float> undef to <4 x i16>
+; AVX2-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V8I16 = fptosi <8 x float> undef to <8 x i16>
+; AVX2-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V16I16 = fptosi <16 x float> undef to <16 x i16>
+; AVX2-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; AVX512-LABEL: 'fptosi_float_i16'
; AVX512-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I16 = fptosi float undef to i16
@@ -290,6 +334,14 @@ define i32 @fptosi_float_i16(i32 %arg) {
; AVX512-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V8I16 = fptosi <8 x float> undef to <8 x i16>
; AVX512-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V16I16 = fptosi <16 x float> undef to <16 x i16>
; AVX512-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+;
+; SLM-LABEL: 'fptosi_float_i16'
+; SLM-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I16 = fptosi float undef to i16
+; SLM-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2I16 = fptosi <2 x float> undef to <2 x i16>
+; SLM-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V4I16 = fptosi <4 x float> undef to <4 x i16>
+; SLM-NEXT: Cost Model: Found an estimated cost of 5 for instruction: %V8I16 = fptosi <8 x float> undef to <8 x i16>
+; SLM-NEXT: Cost Model: Found an estimated cost of 10 for instruction: %V16I16 = fptosi <16 x float> undef to <16 x i16>
+; SLM-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
%I16 = fptosi float undef to i16
%V2I16 = fptosi <2 x float> undef to <2 x i16>
@@ -304,25 +356,33 @@ define i32 @fptosi_float_i8(i32 %arg) {
; SSE2-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I8 = fptosi float undef to i8
; SSE2-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2I8 = fptosi <2 x float> undef to <2 x i8>
; SSE2-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V4I8 = fptosi <4 x float> undef to <4 x i8>
-; SSE2-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V8I8 = fptosi <8 x float> undef to <8 x i8>
-; SSE2-NEXT: Cost Model: Found an estimated cost of 15 for instruction: %V16I8 = fptosi <16 x float> undef to <16 x i8>
+; SSE2-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V8I8 = fptosi <8 x float> undef to <8 x i8>
+; SSE2-NEXT: Cost Model: Found an estimated cost of 11 for instruction: %V16I8 = fptosi <16 x float> undef to <16 x i8>
; SSE2-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; SSE42-LABEL: 'fptosi_float_i8'
; SSE42-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I8 = fptosi float undef to i8
; SSE42-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V2I8 = fptosi <2 x float> undef to <2 x i8>
; SSE42-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V4I8 = fptosi <4 x float> undef to <4 x i8>
-; SSE42-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V8I8 = fptosi <8 x float> undef to <8 x i8>
-; SSE42-NEXT: Cost Model: Found an estimated cost of 15 for instruction: %V16I8 = fptosi <16 x float> undef to <16 x i8>
+; SSE42-NEXT: Cost Model: Found an estimated cost of 5 for instruction: %V8I8 = fptosi <8 x float> undef to <8 x i8>
+; SSE42-NEXT: Cost Model: Found an estimated cost of 11 for instruction: %V16I8 = fptosi <16 x float> undef to <16 x i8>
; SSE42-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
-; AVX-LABEL: 'fptosi_float_i8'
-; AVX-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I8 = fptosi float undef to i8
-; AVX-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V2I8 = fptosi <2 x float> undef to <2 x i8>
-; AVX-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V4I8 = fptosi <4 x float> undef to <4 x i8>
-; AVX-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V8I8 = fptosi <8 x float> undef to <8 x i8>
-; AVX-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V16I8 = fptosi <16 x float> undef to <16 x i8>
-; AVX-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+; AVX1-LABEL: 'fptosi_float_i8'
+; AVX1-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I8 = fptosi float undef to i8
+; AVX1-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V2I8 = fptosi <2 x float> undef to <2 x i8>
+; AVX1-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V4I8 = fptosi <4 x float> undef to <4 x i8>
+; AVX1-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V8I8 = fptosi <8 x float> undef to <8 x i8>
+; AVX1-NEXT: Cost Model: Found an estimated cost of 11 for instruction: %V16I8 = fptosi <16 x float> undef to <16 x i8>
+; AVX1-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+;
+; AVX2-LABEL: 'fptosi_float_i8'
+; AVX2-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I8 = fptosi float undef to i8
+; AVX2-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V2I8 = fptosi <2 x float> undef to <2 x i8>
+; AVX2-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V4I8 = fptosi <4 x float> undef to <4 x i8>
+; AVX2-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V8I8 = fptosi <8 x float> undef to <8 x i8>
+; AVX2-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V16I8 = fptosi <16 x float> undef to <16 x i8>
+; AVX2-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; AVX512-LABEL: 'fptosi_float_i8'
; AVX512-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I8 = fptosi float undef to i8
@@ -336,8 +396,8 @@ define i32 @fptosi_float_i8(i32 %arg) {
; SLM-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I8 = fptosi float undef to i8
; SLM-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V2I8 = fptosi <2 x float> undef to <2 x i8>
; SLM-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V4I8 = fptosi <4 x float> undef to <4 x i8>
-; SLM-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V8I8 = fptosi <8 x float> undef to <8 x i8>
-; SLM-NEXT: Cost Model: Found an estimated cost of 15 for instruction: %V16I8 = fptosi <16 x float> undef to <16 x i8>
+; SLM-NEXT: Cost Model: Found an estimated cost of 5 for instruction: %V8I8 = fptosi <8 x float> undef to <8 x i8>
+; SLM-NEXT: Cost Model: Found an estimated cost of 11 for instruction: %V16I8 = fptosi <16 x float> undef to <16 x i8>
; SLM-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
%I8 = fptosi float undef to i8
diff --git a/llvm/test/Analysis/CostModel/X86/fptoui.ll b/llvm/test/Analysis/CostModel/X86/fptoui.ll
index 5a9f6bf4eb79..dbe62a3e8031 100644
--- a/llvm/test/Analysis/CostModel/X86/fptoui.ll
+++ b/llvm/test/Analysis/CostModel/X86/fptoui.ll
@@ -111,19 +111,33 @@ define i32 @fptoui_double_i32(i32 %arg) {
}
define i32 @fptoui_double_i16(i32 %arg) {
-; SSE-LABEL: 'fptoui_double_i16'
-; SSE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I16 = fptoui double undef to i16
-; SSE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2I16 = fptoui <2 x double> undef to <2 x i16>
-; SSE-NEXT: Cost Model: Found an estimated cost of 5 for instruction: %V4I16 = fptoui <4 x double> undef to <4 x i16>
-; SSE-NEXT: Cost Model: Found an estimated cost of 11 for instruction: %V8I16 = fptoui <8 x double> undef to <8 x i16>
-; SSE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+; SSE2-LABEL: 'fptoui_double_i16'
+; SSE2-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I16 = fptoui double undef to i16
+; SSE2-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2I16 = fptoui <2 x double> undef to <2 x i16>
+; SSE2-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V4I16 = fptoui <4 x double> undef to <4 x i16>
+; SSE2-NEXT: Cost Model: Found an estimated cost of 11 for instruction: %V8I16 = fptoui <8 x double> undef to <8 x i16>
+; SSE2-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
-; AVX-LABEL: 'fptoui_double_i16'
-; AVX-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I16 = fptoui double undef to i16
-; AVX-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2I16 = fptoui <2 x double> undef to <2 x i16>
-; AVX-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V4I16 = fptoui <4 x double> undef to <4 x i16>
-; AVX-NEXT: Cost Model: Found an estimated cost of 5 for instruction: %V8I16 = fptoui <8 x double> undef to <8 x i16>
-; AVX-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+; SSE42-LABEL: 'fptoui_double_i16'
+; SSE42-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I16 = fptoui double undef to i16
+; SSE42-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2I16 = fptoui <2 x double> undef to <2 x i16>
+; SSE42-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V4I16 = fptoui <4 x double> undef to <4 x i16>
+; SSE42-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V8I16 = fptoui <8 x double> undef to <8 x i16>
+; SSE42-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+;
+; AVX1-LABEL: 'fptoui_double_i16'
+; AVX1-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I16 = fptoui double undef to i16
+; AVX1-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2I16 = fptoui <2 x double> undef to <2 x i16>
+; AVX1-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V4I16 = fptoui <4 x double> undef to <4 x i16>
+; AVX1-NEXT: Cost Model: Found an estimated cost of 10 for instruction: %V8I16 = fptoui <8 x double> undef to <8 x i16>
+; AVX1-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+;
+; AVX2-LABEL: 'fptoui_double_i16'
+; AVX2-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I16 = fptoui double undef to i16
+; AVX2-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2I16 = fptoui <2 x double> undef to <2 x i16>
+; AVX2-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V4I16 = fptoui <4 x double> undef to <4 x i16>
+; AVX2-NEXT: Cost Model: Found an estimated cost of 5 for instruction: %V8I16 = fptoui <8 x double> undef to <8 x i16>
+; AVX2-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; AVX512-LABEL: 'fptoui_double_i16'
; AVX512-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I16 = fptoui double undef to i16
@@ -131,6 +145,13 @@ define i32 @fptoui_double_i16(i32 %arg) {
; AVX512-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V4I16 = fptoui <4 x double> undef to <4 x i16>
; AVX512-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V8I16 = fptoui <8 x double> undef to <8 x i16>
; AVX512-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+;
+; SLM-LABEL: 'fptoui_double_i16'
+; SLM-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I16 = fptoui double undef to i16
+; SLM-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2I16 = fptoui <2 x double> undef to <2 x i16>
+; SLM-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V4I16 = fptoui <4 x double> undef to <4 x i16>
+; SLM-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V8I16 = fptoui <8 x double> undef to <8 x i16>
+; SLM-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
%I16 = fptoui double undef to i16
%V2I16 = fptoui <2 x double> undef to <2 x i16>
@@ -143,23 +164,30 @@ define i32 @fptoui_double_i8(i32 %arg) {
; SSE2-LABEL: 'fptoui_double_i8'
; SSE2-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I8 = fptoui double undef to i8
; SSE2-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2I8 = fptoui <2 x double> undef to <2 x i8>
-; SSE2-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V4I8 = fptoui <4 x double> undef to <4 x i8>
-; SSE2-NEXT: Cost Model: Found an estimated cost of 19 for instruction: %V8I8 = fptoui <8 x double> undef to <8 x i8>
+; SSE2-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V4I8 = fptoui <4 x double> undef to <4 x i8>
+; SSE2-NEXT: Cost Model: Found an estimated cost of 10 for instruction: %V8I8 = fptoui <8 x double> undef to <8 x i8>
; SSE2-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; SSE42-LABEL: 'fptoui_double_i8'
; SSE42-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I8 = fptoui double undef to i8
; SSE42-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V2I8 = fptoui <2 x double> undef to <2 x i8>
-; SSE42-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V4I8 = fptoui <4 x double> undef to <4 x i8>
-; SSE42-NEXT: Cost Model: Found an estimated cost of 15 for instruction: %V8I8 = fptoui <8 x double> undef to <8 x i8>
+; SSE42-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V4I8 = fptoui <4 x double> undef to <4 x i8>
+; SSE42-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V8I8 = fptoui <8 x double> undef to <8 x i8>
; SSE42-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
-; AVX-LABEL: 'fptoui_double_i8'
-; AVX-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I8 = fptoui double undef to i8
-; AVX-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V2I8 = fptoui <2 x double> undef to <2 x i8>
-; AVX-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V4I8 = fptoui <4 x double> undef to <4 x i8>
-; AVX-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V8I8 = fptoui <8 x double> undef to <8 x i8>
-; AVX-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+; AVX1-LABEL: 'fptoui_double_i8'
+; AVX1-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I8 = fptoui double undef to i8
+; AVX1-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V2I8 = fptoui <2 x double> undef to <2 x i8>
+; AVX1-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V4I8 = fptoui <4 x double> undef to <4 x i8>
+; AVX1-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V8I8 = fptoui <8 x double> undef to <8 x i8>
+; AVX1-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+;
+; AVX2-LABEL: 'fptoui_double_i8'
+; AVX2-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I8 = fptoui double undef to i8
+; AVX2-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V2I8 = fptoui <2 x double> undef to <2 x i8>
+; AVX2-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V4I8 = fptoui <4 x double> undef to <4 x i8>
+; AVX2-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V8I8 = fptoui <8 x double> undef to <8 x i8>
+; AVX2-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; AVX512-LABEL: 'fptoui_double_i8'
; AVX512-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I8 = fptoui double undef to i8
@@ -171,8 +199,8 @@ define i32 @fptoui_double_i8(i32 %arg) {
; SLM-LABEL: 'fptoui_double_i8'
; SLM-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I8 = fptoui double undef to i8
; SLM-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V2I8 = fptoui <2 x double> undef to <2 x i8>
-; SLM-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V4I8 = fptoui <4 x double> undef to <4 x i8>
-; SLM-NEXT: Cost Model: Found an estimated cost of 15 for instruction: %V8I8 = fptoui <8 x double> undef to <8 x i8>
+; SLM-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V4I8 = fptoui <4 x double> undef to <4 x i8>
+; SLM-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V8I8 = fptoui <8 x double> undef to <8 x i8>
; SLM-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
%I8 = fptoui double undef to i8
@@ -285,8 +313,8 @@ define i32 @fptoui_float_i16(i32 %arg) {
; SSE2-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I16 = fptoui float undef to i16
; SSE2-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2I16 = fptoui <2 x float> undef to <2 x i16>
; SSE2-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V4I16 = fptoui <4 x float> undef to <4 x i16>
-; SSE2-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V8I16 = fptoui <8 x float> undef to <8 x i16>
-; SSE2-NEXT: Cost Model: Found an estimated cost of 18 for instruction: %V16I16 = fptoui <16 x float> undef to <16 x i16>
+; SSE2-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V8I16 = fptoui <8 x float> undef to <8 x i16>
+; SSE2-NEXT: Cost Model: Found an estimated cost of 14 for instruction: %V16I16 = fptoui <16 x float> undef to <16 x i16>
; SSE2-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; SSE42-LABEL: 'fptoui_float_i16'
@@ -297,13 +325,21 @@ define i32 @fptoui_float_i16(i32 %arg) {
; SSE42-NEXT: Cost Model: Found an estimated cost of 10 for instruction: %V16I16 = fptoui <16 x float> undef to <16 x i16>
; SSE42-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
-; AVX-LABEL: 'fptoui_float_i16'
-; AVX-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I16 = fptoui float undef to i16
-; AVX-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2I16 = fptoui <2 x float> undef to <2 x i16>
-; AVX-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V4I16 = fptoui <4 x float> undef to <4 x i16>
-; AVX-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V8I16 = fptoui <8 x float> undef to <8 x i16>
-; AVX-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V16I16 = fptoui <16 x float> undef to <16 x i16>
-; AVX-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+; AVX1-LABEL: 'fptoui_float_i16'
+; AVX1-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I16 = fptoui float undef to i16
+; AVX1-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2I16 = fptoui <2 x float> undef to <2 x i16>
+; AVX1-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V4I16 = fptoui <4 x float> undef to <4 x i16>
+; AVX1-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V8I16 = fptoui <8 x float> undef to <8 x i16>
+; AVX1-NEXT: Cost Model: Found an estimated cost of 10 for instruction: %V16I16 = fptoui <16 x float> undef to <16 x i16>
+; AVX1-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+;
+; AVX2-LABEL: 'fptoui_float_i16'
+; AVX2-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I16 = fptoui float undef to i16
+; AVX2-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2I16 = fptoui <2 x float> undef to <2 x i16>
+; AVX2-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V4I16 = fptoui <4 x float> undef to <4 x i16>
+; AVX2-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V8I16 = fptoui <8 x float> undef to <8 x i16>
+; AVX2-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V16I16 = fptoui <16 x float> undef to <16 x i16>
+; AVX2-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; AVX512-LABEL: 'fptoui_float_i16'
; AVX512-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I16 = fptoui float undef to i16
@@ -334,25 +370,33 @@ define i32 @fptoui_float_i8(i32 %arg) {
; SSE2-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I8 = fptoui float undef to i8
; SSE2-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2I8 = fptoui <2 x float> undef to <2 x i8>
; SSE2-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V4I8 = fptoui <4 x float> undef to <4 x i8>
-; SSE2-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V8I8 = fptoui <8 x float> undef to <8 x i8>
-; SSE2-NEXT: Cost Model: Found an estimated cost of 15 for instruction: %V16I8 = fptoui <16 x float> undef to <16 x i8>
+; SSE2-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V8I8 = fptoui <8 x float> undef to <8 x i8>
+; SSE2-NEXT: Cost Model: Found an estimated cost of 11 for instruction: %V16I8 = fptoui <16 x float> undef to <16 x i8>
; SSE2-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; SSE42-LABEL: 'fptoui_float_i8'
; SSE42-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I8 = fptoui float undef to i8
; SSE42-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V2I8 = fptoui <2 x float> undef to <2 x i8>
; SSE42-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V4I8 = fptoui <4 x float> undef to <4 x i8>
-; SSE42-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V8I8 = fptoui <8 x float> undef to <8 x i8>
-; SSE42-NEXT: Cost Model: Found an estimated cost of 15 for instruction: %V16I8 = fptoui <16 x float> undef to <16 x i8>
+; SSE42-NEXT: Cost Model: Found an estimated cost of 5 for instruction: %V8I8 = fptoui <8 x float> undef to <8 x i8>
+; SSE42-NEXT: Cost Model: Found an estimated cost of 11 for instruction: %V16I8 = fptoui <16 x float> undef to <16 x i8>
; SSE42-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
-; AVX-LABEL: 'fptoui_float_i8'
-; AVX-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I8 = fptoui float undef to i8
-; AVX-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V2I8 = fptoui <2 x float> undef to <2 x i8>
-; AVX-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V4I8 = fptoui <4 x float> undef to <4 x i8>
-; AVX-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V8I8 = fptoui <8 x float> undef to <8 x i8>
-; AVX-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V16I8 = fptoui <16 x float> undef to <16 x i8>
-; AVX-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+; AVX1-LABEL: 'fptoui_float_i8'
+; AVX1-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I8 = fptoui float undef to i8
+; AVX1-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V2I8 = fptoui <2 x float> undef to <2 x i8>
+; AVX1-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V4I8 = fptoui <4 x float> undef to <4 x i8>
+; AVX1-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V8I8 = fptoui <8 x float> undef to <8 x i8>
+; AVX1-NEXT: Cost Model: Found an estimated cost of 11 for instruction: %V16I8 = fptoui <16 x float> undef to <16 x i8>
+; AVX1-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
+;
+; AVX2-LABEL: 'fptoui_float_i8'
+; AVX2-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I8 = fptoui float undef to i8
+; AVX2-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V2I8 = fptoui <2 x float> undef to <2 x i8>
+; AVX2-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V4I8 = fptoui <4 x float> undef to <4 x i8>
+; AVX2-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V8I8 = fptoui <8 x float> undef to <8 x i8>
+; AVX2-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V16I8 = fptoui <16 x float> undef to <16 x i8>
+; AVX2-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; AVX512-LABEL: 'fptoui_float_i8'
; AVX512-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I8 = fptoui float undef to i8
@@ -366,8 +410,8 @@ define i32 @fptoui_float_i8(i32 %arg) {
; SLM-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %I8 = fptoui float undef to i8
; SLM-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V2I8 = fptoui <2 x float> undef to <2 x i8>
; SLM-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V4I8 = fptoui <4 x float> undef to <4 x i8>
-; SLM-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V8I8 = fptoui <8 x float> undef to <8 x i8>
-; SLM-NEXT: Cost Model: Found an estimated cost of 15 for instruction: %V16I8 = fptoui <16 x float> undef to <16 x i8>
+; SLM-NEXT: Cost Model: Found an estimated cost of 5 for instruction: %V8I8 = fptoui <8 x float> undef to <8 x i8>
+; SLM-NEXT: Cost Model: Found an estimated cost of 11 for instruction: %V16I8 = fptoui <16 x float> undef to <16 x i8>
; SLM-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
%I8 = fptoui float undef to i8
More information about the llvm-commits
mailing list