maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity

[X86][AVX2] Fix costs for v4i64 ashr by splat 

llvm-svn: 303022
This commit is contained in:
Simon Pilgrim 2017-05-14 20:25:42 +00:00
parent de4467b182
commit f96b4ab92d
2 changed files with 7 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
llvm/lib/Target/X86/X86TargetTransformInfo.cpp
View File

@ -471,6 +471,11 @@ int X86TTIImpl::getArithmeticInstrCost(
if (ST->hasSSE2() &&
((Op2Info == TargetTransformInfo::OK_UniformConstantValue) ||
(Op2Info == TargetTransformInfo::OK_UniformValue))) {
// Handle AVX2 uniform v4i64 ISD::SRA, it's not worth a table.
if (ISD == ISD::SRA && LT.second == MVT::v4i64 && ST->hasAVX2())
return LT.first * 4; // 2*psrad + shuffle.
if (const auto *Entry =
CostTableLookup(SSE2UniformShiftCostTable, ISD, LT.second))
return LT.first * Entry->Cost;

4
llvm/test/Analysis/CostModel/X86/vshift-ashr-cost.ll
View File

@ -192,7 +192,7 @@ define <4 x i64> @splatvar_shift_v4i64(<4 x i64> %a, i64 %b) {
; SSE2: Found an estimated cost of 8 for instruction: %shift
; SSE41: Found an estimated cost of 8 for instruction: %shift
; AVX: Found an estimated cost of 10 for instruction: %shift
; AVX2: Found an estimated cost of 10 for instruction: %shift
; AVX2: Found an estimated cost of 4 for instruction: %shift
; AVX512: Found an estimated cost of 1 for instruction: %shift
; XOP: Found an estimated cost of 6 for instruction: %shift
%insert = insertelement <4 x i64> undef, i64 %b, i32 0
@ -206,7 +206,7 @@ define <8 x i64> @splatvar_shift_v8i64(<8 x i64> %a, i64 %b) {
; SSE2: Found an estimated cost of 16 for instruction: %shift
; SSE41: Found an estimated cost of 16 for instruction: %shift
; AVX: Found an estimated cost of 20 for instruction: %shift
; AVX2: Found an estimated cost of 20 for instruction: %shift
; AVX2: Found an estimated cost of 8 for instruction: %shift
; AVX512: Found an estimated cost of 1 for instruction: %shift
; XOP: Found an estimated cost of 12 for instruction: %shift
%insert = insertelement <8 x i64> undef, i64 %b, i32 0