shylie/llvm-project
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

[CostModel][X86] Fix fpext conversion cost for 16 elements (#76278)

Browse Source 
The fpext conversion cost for 16 elements should be 4 from Znver4.
This commit is contained in:
HaohaiWen 2024-01-09 09:05:11 +08:00 committed by GitHub
parent eee71ed3f7
commit 4147b72301
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
2 changed files with 2 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
llvm/lib/Target/X86/X86TargetTransformInfo.cpp
View File

@ -2232,6 +2232,7 @@ InstructionCost X86TTIImpl::getCastInstrCost(unsigned Opcode, Type *Dst,
static const TypeConversionCostTblEntry AVX512FConversionTbl[] = {
{ ISD::FP_EXTEND, MVT::v8f64, MVT::v8f32, 1 },
{ ISD::FP_EXTEND, MVT::v8f64, MVT::v16f32, 3 },
{ ISD::FP_EXTEND, MVT::v16f64, MVT::v16f32, 4 }, // 2*vcvtps2pd+vextractf64x4
{ ISD::FP_ROUND, MVT::v8f32, MVT::v8f64, 1 },
{ ISD::TRUNCATE, MVT::v2i1, MVT::v2i8, 3 }, // sext+vpslld+vptestmd

2
llvm/test/Analysis/CostModel/X86/cast.ll
View File

@ -632,7 +632,7 @@ define void @fp_conv(<8 x float> %a, <16 x float>%b, <4 x float> %c) {
; AVX512-LABEL: 'fp_conv'
; AVX512-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %A1 = fpext <4 x float> %c to <4 x double>
; AVX512-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %A2 = fpext <8 x float> %a to <8 x double>
; AVX512-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %A3 = fpext <16 x float> %b to <16 x double>
; AVX512-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %A3 = fpext <16 x float> %b to <16 x double>
; AVX512-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %A4 = fptrunc <4 x double> undef to <4 x float>
; AVX512-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %A5 = fptrunc <8 x double> undef to <8 x float>
; AVX512-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void