Skip to content

Commit

Permalink
[CostModel][X86] Treat lrint/llrint as fptosi calls (#90883)
Browse files Browse the repository at this point in the history 
X86 can use the CVTP2SI instructions to lower lrint/llrint calls, which have the same costs as the CVTTP2SI (fptosi) instructions

Followup to #90065
  • Loading branch information
@RKSimon
RKSimon committed May 3, 2024
1 parent 3563af6 commit 8a0073a
Show file tree
Hide file tree
Showing 6 changed files with 508 additions and 484 deletions.
10 changes: 10 additions & 0 deletions llvm/lib/Target/X86/X86TargetTransformInfo.cpp
View file
Original file line number Diff line number Diff line change
Expand Up @@ -4153,6 +4153,16 @@ X86TTIImpl::getIntrinsicInstrCost(const IntrinsicCostAttributes &ICA,
}
}
break;
case Intrinsic::lrint:
case Intrinsic::llrint:
// X86 can use the CVTP2SI instructions to lower lrint/llrint calls, which
// have the same costs as the CVTTP2SI (fptosi) instructions
if (!ICA.isTypeBasedOnly()) {
const SmallVectorImpl<Type *> &ArgTys = ICA.getArgTypes();
return getCastInstrCost(Instruction::FPToSI, RetTy, ArgTys[0],
TTI::CastContextHint::None, CostKind);
}
break;
case Intrinsic::maxnum:
case Intrinsic::minnum:
// FMINNUM has same costs so don't duplicate.
Expand Down
89 changes: 39 additions & 50 deletions llvm/test/Analysis/CostModel/X86/arith-fp-codesize.ll
View file
Original file line number Diff line number Diff line change
Expand Up @@ -666,35 +666,46 @@ define i32 @rint(i32 %arg) {
define i32 @lrint(i32 %arg) {
; SSE1-LABEL: 'lrint'
; SSE1-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %F32 = call i32 @llvm.lrint.i32.f32(float undef)
; SSE1-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4F32 = call <4 x i32> @llvm.lrint.v4i32.v4f32(<4 x float> undef)
; SSE1-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8F32 = call <8 x i32> @llvm.lrint.v8i32.v8f32(<8 x float> undef)
; SSE1-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V16F32 = call <16 x i32> @llvm.lrint.v16i32.v16f32(<16 x float> undef)
; SSE1-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = call <4 x i32> @llvm.lrint.v4i32.v4f32(<4 x float> undef)
; SSE1-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = call <8 x i32> @llvm.lrint.v8i32.v8f32(<8 x float> undef)
; SSE1-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = call <16 x i32> @llvm.lrint.v16i32.v16f32(<16 x float> undef)
; SSE1-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %F64 = call i32 @llvm.lrint.i32.f64(double undef)
; SSE1-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2F64 = call <2 x i32> @llvm.lrint.v2i32.v2f64(<2 x double> undef)
; SSE1-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4F64 = call <4 x i32> @llvm.lrint.v4i32.v4f64(<4 x double> undef)
; SSE1-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8F64 = call <8 x i32> @llvm.lrint.v8i32.v8f64(<8 x double> undef)
; SSE1-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = call <2 x i32> @llvm.lrint.v2i32.v2f64(<2 x double> undef)
; SSE1-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = call <4 x i32> @llvm.lrint.v4i32.v4f64(<4 x double> undef)
; SSE1-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = call <8 x i32> @llvm.lrint.v8i32.v8f64(<8 x double> undef)
; SSE1-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
;
; SSE2-LABEL: 'lrint'
; SSE2-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %F32 = call i32 @llvm.lrint.i32.f32(float undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = call <4 x i32> @llvm.lrint.v4i32.v4f32(<4 x float> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V8F32 = call <8 x i32> @llvm.lrint.v8i32.v8f32(<8 x float> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V16F32 = call <16 x i32> @llvm.lrint.v16i32.v16f32(<16 x float> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %F64 = call i32 @llvm.lrint.i32.f64(double undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = call <2 x i32> @llvm.lrint.v2i32.v2f64(<2 x double> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V4F64 = call <4 x i32> @llvm.lrint.v4i32.v4f64(<4 x double> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V8F64 = call <8 x i32> @llvm.lrint.v8i32.v8f64(<8 x double> undef)
; SSE2-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
;
; AVX-LABEL: 'lrint'
; AVX-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %F32 = call i32 @llvm.lrint.i32.f32(float undef)
; AVX-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4F32 = call <4 x i32> @llvm.lrint.v4i32.v4f32(<4 x float> undef)
; AVX-NEXT: Cost Model: Found an estimated cost of 17 for instruction: %V8F32 = call <8 x i32> @llvm.lrint.v8i32.v8f32(<8 x float> undef)
; AVX-NEXT: Cost Model: Found an estimated cost of 34 for instruction: %V16F32 = call <16 x i32> @llvm.lrint.v16i32.v16f32(<16 x float> undef)
; AVX-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = call <4 x i32> @llvm.lrint.v4i32.v4f32(<4 x float> undef)
; AVX-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = call <8 x i32> @llvm.lrint.v8i32.v8f32(<8 x float> undef)
; AVX-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V16F32 = call <16 x i32> @llvm.lrint.v16i32.v16f32(<16 x float> undef)
; AVX-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %F64 = call i32 @llvm.lrint.i32.f64(double undef)
; AVX-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2F64 = call <2 x i32> @llvm.lrint.v2i32.v2f64(<2 x double> undef)
; AVX-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4F64 = call <4 x i32> @llvm.lrint.v4i32.v4f64(<4 x double> undef)
; AVX-NEXT: Cost Model: Found an estimated cost of 17 for instruction: %V8F64 = call <8 x i32> @llvm.lrint.v8i32.v8f64(<8 x double> undef)
; AVX-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = call <2 x i32> @llvm.lrint.v2i32.v2f64(<2 x double> undef)
; AVX-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = call <4 x i32> @llvm.lrint.v4i32.v4f64(<4 x double> undef)
; AVX-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = call <8 x i32> @llvm.lrint.v8i32.v8f64(<8 x double> undef)
; AVX-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
;
; AVX512-LABEL: 'lrint'
; AVX512-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %F32 = call i32 @llvm.lrint.i32.f32(float undef)
; AVX512-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4F32 = call <4 x i32> @llvm.lrint.v4i32.v4f32(<4 x float> undef)
; AVX512-NEXT: Cost Model: Found an estimated cost of 17 for instruction: %V8F32 = call <8 x i32> @llvm.lrint.v8i32.v8f32(<8 x float> undef)
; AVX512-NEXT: Cost Model: Found an estimated cost of 35 for instruction: %V16F32 = call <16 x i32> @llvm.lrint.v16i32.v16f32(<16 x float> undef)
; AVX512-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = call <4 x i32> @llvm.lrint.v4i32.v4f32(<4 x float> undef)
; AVX512-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = call <8 x i32> @llvm.lrint.v8i32.v8f32(<8 x float> undef)
; AVX512-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = call <16 x i32> @llvm.lrint.v16i32.v16f32(<16 x float> undef)
; AVX512-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %F64 = call i32 @llvm.lrint.i32.f64(double undef)
; AVX512-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2F64 = call <2 x i32> @llvm.lrint.v2i32.v2f64(<2 x double> undef)
; AVX512-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4F64 = call <4 x i32> @llvm.lrint.v4i32.v4f64(<4 x double> undef)
; AVX512-NEXT: Cost Model: Found an estimated cost of 17 for instruction: %V8F64 = call <8 x i32> @llvm.lrint.v8i32.v8f64(<8 x double> undef)
; AVX512-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = call <2 x i32> @llvm.lrint.v2i32.v2f64(<2 x double> undef)
; AVX512-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = call <4 x i32> @llvm.lrint.v4i32.v4f64(<4 x double> undef)
; AVX512-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = call <8 x i32> @llvm.lrint.v8i32.v8f64(<8 x double> undef)
; AVX512-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
;
%F32 = call i32 @llvm.lrint.i32.f32(float undef)
Expand All @@ -711,38 +722,16 @@ define i32 @lrint(i32 %arg) {
}

define i32 @llrint(i32 %arg) {
; SSE1-LABEL: 'llrint'
; SSE1-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %F32 = call i64 @llvm.llrint.i64.f32(float undef)
; SSE1-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4F32 = call <4 x i64> @llvm.llrint.v4i64.v4f32(<4 x float> undef)
; SSE1-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8F32 = call <8 x i64> @llvm.llrint.v8i64.v8f32(<8 x float> undef)
; SSE1-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V16F32 = call <16 x i64> @llvm.llrint.v16i64.v16f32(<16 x float> undef)
; SSE1-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %F64 = call i64 @llvm.llrint.i64.f64(double undef)
; SSE1-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2F64 = call <2 x i64> @llvm.llrint.v2i64.v2f64(<2 x double> undef)
; SSE1-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4F64 = call <4 x i64> @llvm.llrint.v4i64.v4f64(<4 x double> undef)
; SSE1-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8F64 = call <8 x i64> @llvm.llrint.v8i64.v8f64(<8 x double> undef)
; SSE1-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
;
; AVX-LABEL: 'llrint'
; AVX-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %F32 = call i64 @llvm.llrint.i64.f32(float undef)
; AVX-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V4F32 = call <4 x i64> @llvm.llrint.v4i64.v4f32(<4 x float> undef)
; AVX-NEXT: Cost Model: Found an estimated cost of 18 for instruction: %V8F32 = call <8 x i64> @llvm.llrint.v8i64.v8f32(<8 x float> undef)
; AVX-NEXT: Cost Model: Found an estimated cost of 36 for instruction: %V16F32 = call <16 x i64> @llvm.llrint.v16i64.v16f32(<16 x float> undef)
; AVX-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %F64 = call i64 @llvm.llrint.i64.f64(double undef)
; AVX-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2F64 = call <2 x i64> @llvm.llrint.v2i64.v2f64(<2 x double> undef)
; AVX-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V4F64 = call <4 x i64> @llvm.llrint.v4i64.v4f64(<4 x double> undef)
; AVX-NEXT: Cost Model: Found an estimated cost of 18 for instruction: %V8F64 = call <8 x i64> @llvm.llrint.v8i64.v8f64(<8 x double> undef)
; AVX-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
;
; AVX512-LABEL: 'llrint'
; AVX512-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %F32 = call i64 @llvm.llrint.i64.f32(float undef)
; AVX512-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V4F32 = call <4 x i64> @llvm.llrint.v4i64.v4f32(<4 x float> undef)
; AVX512-NEXT: Cost Model: Found an estimated cost of 19 for instruction: %V8F32 = call <8 x i64> @llvm.llrint.v8i64.v8f32(<8 x float> undef)
; AVX512-NEXT: Cost Model: Found an estimated cost of 38 for instruction: %V16F32 = call <16 x i64> @llvm.llrint.v16i64.v16f32(<16 x float> undef)
; AVX512-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %F64 = call i64 @llvm.llrint.i64.f64(double undef)
; AVX512-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2F64 = call <2 x i64> @llvm.llrint.v2i64.v2f64(<2 x double> undef)
; AVX512-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V4F64 = call <4 x i64> @llvm.llrint.v4i64.v4f64(<4 x double> undef)
; AVX512-NEXT: Cost Model: Found an estimated cost of 19 for instruction: %V8F64 = call <8 x i64> @llvm.llrint.v8i64.v8f64(<8 x double> undef)
; AVX512-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
; CHECK-LABEL: 'llrint'
; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %F32 = call i64 @llvm.llrint.i64.f32(float undef)
; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V4F32 = call <4 x i64> @llvm.llrint.v4i64.v4f32(<4 x float> undef)
; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V8F32 = call <8 x i64> @llvm.llrint.v8i64.v8f32(<8 x float> undef)
; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V16F32 = call <16 x i64> @llvm.llrint.v16i64.v16f32(<16 x float> undef)
; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %F64 = call i64 @llvm.llrint.i64.f64(double undef)
; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V2F64 = call <2 x i64> @llvm.llrint.v2i64.v2f64(<2 x double> undef)
; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V4F64 = call <4 x i64> @llvm.llrint.v4i64.v4f64(<4 x double> undef)
; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %V8F64 = call <8 x i64> @llvm.llrint.v8i64.v8f64(<8 x double> undef)
; CHECK-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
;
%F32 = call i64 @llvm.llrint.i64.f32(float undef)
%V4F32 = call <4 x i64> @llvm.llrint.v4i64.v4f32(<4 x float> undef)
Expand Down
Loading

0 comments on commit 8a0073a

Please sign in to comment.