Skip to content

AMDGPU: Add baseline test for cost of abs intrinsics #100522

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Merged
merged 1 commit into from
Jul 25, 2024
Merged

AMDGPU: Add baseline test for cost of abs intrinsics #100522

merged 1 commit into from
Jul 25, 2024

Conversation

arsenm
Copy link
Contributor

@arsenm arsenm commented Jul 25, 2024

No description provided.

This was referenced Jul 25, 2024
Merged
Merged
Merged
Merged
Merged
Merged
@arsenm Graphite App
Copy link
Contributor Author

arsenm commented Jul 25, 2024 •
edited
Loading

This stack of pull requests is managed by Graphite. Learn more about stacking.

Join @arsenm and the rest of your teammates on Graphite Graphite

This was referenced Jul 25, 2024
Merged
Merged
@arsenm arsenm added backend:AMDGPU vectorizers llvm:analysis Includes value tracking, cost tables and constant folding llvm:transforms labels Jul 25, 2024 — with Graphite App
@arsenm arsenm marked this pull request as ready for review July 25, 2024 07:05
@llvmbot
Copy link
Member

llvmbot commented Jul 25, 2024 •
edited
Loading

@llvm/pr-subscribers-llvm-transforms
@llvm/pr-subscribers-llvm-analysis

@llvm/pr-subscribers-backend-amdgpu

Author: Matt Arsenault (arsenm)

Changes

Patch is 36.04 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/100522.diff

1 Files Affected:

  • (added) llvm/test/Analysis/CostModel/AMDGPU/abs.ll (+341)
diff --git a/llvm/test/Analysis/CostModel/AMDGPU/abs.ll b/llvm/test/Analysis/CostModel/AMDGPU/abs.ll
new file mode 100644
index 0000000000000..133b95609bc15
--- /dev/null
+++ b/llvm/test/Analysis/CostModel/AMDGPU/abs.ll
@@ -0,0 +1,341 @@
+; NOTE: Assertions have been autogenerated by utils/update_analyze_test_checks.py
+; RUN: opt -passes="print<cost-model>" 2>&1 -disable-output -mtriple=amdgcn-unknown-amdhsa -mcpu=gfx1010 < %s | FileCheck -check-prefixes=FAST %s
+; RUN: opt -passes="print<cost-model>" 2>&1 -disable-output -mtriple=amdgcn-unknown-amdhsa -mcpu=gfx90a < %s | FileCheck -check-prefixes=FAST %s
+; RUN: opt -passes="print<cost-model>" 2>&1 -disable-output -mtriple=amdgcn-unknown-amdhsa -mcpu=gfx900 < %s | FileCheck -check-prefixes=FAST %s
+; RUN: opt -passes="print<cost-model>" 2>&1 -disable-output -mtriple=amdgcn-unknown-amdhsa < %s | FileCheck -check-prefixes=SLOW %s
+; RUN: opt -passes="print<cost-model>" -cost-kind=code-size 2>&1 -disable-output -mtriple=amdgcn-unknown-amdhsa -mcpu=gfx1010 < %s | FileCheck -check-prefixes=FAST-SIZE %s
+; RUN: opt -passes="print<cost-model>" -cost-kind=code-size 2>&1 -disable-output -mtriple=amdgcn-unknown-amdhsa -mcpu=gfx90a < %s | FileCheck -check-prefixes=FAST-SIZE %s
+; RUN: opt -passes="print<cost-model>" -cost-kind=code-size 2>&1 -disable-output -mtriple=amdgcn-unknown-amdhsa -mcpu=gfx900 < %s | FileCheck -check-prefixes=FAST-SIZE %s
+; RUN: opt -passes="print<cost-model>" -cost-kind=code-size 2>&1 -disable-output -mtriple=amdgcn-unknown-amdhsa < %s | FileCheck -check-prefixes=SLOW-SIZE %s
+; END.
+
+declare i64        @llvm.abs.i64(i64, i1 immarg)
+declare <2 x i64>  @llvm.abs.v2i64(<2 x i64>, i1 immarg)
+declare <4 x i64>  @llvm.abs.v4i64(<4 x i64>, i1 immarg)
+declare <5 x i64>  @llvm.abs.v5i64(<5 x i64>, i1 immarg)
+declare <8 x i64>  @llvm.abs.v8i64(<8 x i64>, i1 immarg)
+
+declare i32        @llvm.abs.i32(i32, i1 immarg)
+declare <2 x i32>  @llvm.abs.v2i32(<2 x i32>, i1 immarg)
+declare <4 x i32>  @llvm.abs.v4i32(<4 x i32>, i1 immarg)
+declare <8 x i32>  @llvm.abs.v8i32(<8 x i32>, i1 immarg)
+declare <9 x i32>  @llvm.abs.v9i32(<9 x i32>, i1 immarg)
+declare <16 x i32> @llvm.abs.v16i32(<16 x i32>, i1 immarg)
+
+declare i16        @llvm.abs.i16(i16, i1 immarg)
+declare <2 x i16>  @llvm.abs.v2i16(<2 x i16>, i1 immarg)
+declare <4 x i16>  @llvm.abs.v4i16(<4 x i16>, i1 immarg)
+declare <8 x i16>  @llvm.abs.v8i16(<8 x i16>, i1 immarg)
+declare <16 x i16> @llvm.abs.v16i16(<16 x i16>, i1 immarg)
+declare <17 x i16> @llvm.abs.v17i16(<17 x i16>, i1 immarg)
+declare <32 x i16> @llvm.abs.v32i16(<32 x i16>, i1 immarg)
+
+declare i8         @llvm.abs.i8(i8, i1 immarg)
+declare <2 x i8>   @llvm.abs.v2i8(<2 x i8>, i1 immarg)
+declare <4 x i8>   @llvm.abs.v4i8(<4 x i8>, i1 immarg)
+declare <8 x i8>   @llvm.abs.v8i8(<8 x i8>, i1 immarg)
+declare <16 x i8>  @llvm.abs.v16i8(<16 x i8>, i1 immarg)
+declare <32 x i8>  @llvm.abs.v32i8(<32 x i8>, i1 immarg)
+declare <33 x i8>  @llvm.abs.v33i8(<33 x i8>, i1 immarg)
+declare <64 x i8>  @llvm.abs.v64i8(<64 x i8>, i1 immarg)
+
+define i32 @abs_nonpoison(i32 %arg) {
+; FAST-LABEL: 'abs_nonpoison'
+; FAST-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %I64 = call i64 @llvm.abs.i64(i64 undef, i1 false)
+; FAST-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V2I64 = call <2 x i64> @llvm.abs.v2i64(<2 x i64> undef, i1 false)
+; FAST-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V4I64 = call <4 x i64> @llvm.abs.v4i64(<4 x i64> undef, i1 false)
+; FAST-NEXT:  Cost Model: Found an estimated cost of 58 for instruction: %V5I64 = call <5 x i64> @llvm.abs.v5i64(<5 x i64> undef, i1 false)
+; FAST-NEXT:  Cost Model: Found an estimated cost of 64 for instruction: %V8I64 = call <8 x i64> @llvm.abs.v8i64(<8 x i64> undef, i1 false)
+; FAST-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %I32 = call i32 @llvm.abs.i32(i32 undef, i1 false)
+; FAST-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V2I32 = call <2 x i32> @llvm.abs.v2i32(<2 x i32> undef, i1 false)
+; FAST-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V4I32 = call <4 x i32> @llvm.abs.v4i32(<4 x i32> undef, i1 false)
+; FAST-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V8I32 = call <8 x i32> @llvm.abs.v8i32(<8 x i32> undef, i1 false)
+; FAST-NEXT:  Cost Model: Found an estimated cost of 45 for instruction: %V9I32 = call <9 x i32> @llvm.abs.v9i32(<9 x i32> undef, i1 false)
+; FAST-NEXT:  Cost Model: Found an estimated cost of 80 for instruction: %V16I32 = call <16 x i32> @llvm.abs.v16i32(<16 x i32> undef, i1 false)
+; FAST-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %I16 = call i16 @llvm.abs.i16(i16 undef, i1 false)
+; FAST-NEXT:  Cost Model: Found an estimated cost of 7 for instruction: %V2I16 = call <2 x i16> @llvm.abs.v2i16(<2 x i16> undef, i1 false)
+; FAST-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V4I16 = call <4 x i16> @llvm.abs.v4i16(<4 x i16> undef, i1 false)
+; FAST-NEXT:  Cost Model: Found an estimated cost of 34 for instruction: %V8I16 = call <8 x i16> @llvm.abs.v8i16(<8 x i16> undef, i1 false)
+; FAST-NEXT:  Cost Model: Found an estimated cost of 70 for instruction: %V16I16 = call <16 x i16> @llvm.abs.v16i16(<16 x i16> undef, i1 false)
+; FAST-NEXT:  Cost Model: Found an estimated cost of 114 for instruction: %V17I16 = call <17 x i16> @llvm.abs.v17i16(<17 x i16> undef, i1 false)
+; FAST-NEXT:  Cost Model: Found an estimated cost of 174 for instruction: %V32I16 = call <32 x i16> @llvm.abs.v32i16(<32 x i16> undef, i1 false)
+; FAST-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %I8 = call i8 @llvm.abs.i8(i8 undef, i1 false)
+; FAST-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V2I8 = call <2 x i8> @llvm.abs.v2i8(<2 x i8> undef, i1 false)
+; FAST-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V4I8 = call <4 x i8> @llvm.abs.v4i8(<4 x i8> undef, i1 false)
+; FAST-NEXT:  Cost Model: Found an estimated cost of 40 for instruction: %V8I8 = call <8 x i8> @llvm.abs.v8i8(<8 x i8> undef, i1 false)
+; FAST-NEXT:  Cost Model: Found an estimated cost of 80 for instruction: %V16I8 = call <16 x i8> @llvm.abs.v16i8(<16 x i8> undef, i1 false)
+; FAST-NEXT:  Cost Model: Found an estimated cost of 160 for instruction: %V32I8 = call <32 x i8> @llvm.abs.v32i8(<32 x i8> undef, i1 false)
+; FAST-NEXT:  Cost Model: Found an estimated cost of 198 for instruction: %V33I8 = call <33 x i8> @llvm.abs.v33i8(<33 x i8> undef, i1 false)
+; FAST-NEXT:  Cost Model: Found an estimated cost of 322 for instruction: %V64I8 = call <64 x i8> @llvm.abs.v64i8(<64 x i8> undef, i1 false)
+; FAST-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: ret i32 undef
+;
+; SLOW-LABEL: 'abs_nonpoison'
+; SLOW-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %I64 = call i64 @llvm.abs.i64(i64 undef, i1 false)
+; SLOW-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V2I64 = call <2 x i64> @llvm.abs.v2i64(<2 x i64> undef, i1 false)
+; SLOW-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V4I64 = call <4 x i64> @llvm.abs.v4i64(<4 x i64> undef, i1 false)
+; SLOW-NEXT:  Cost Model: Found an estimated cost of 58 for instruction: %V5I64 = call <5 x i64> @llvm.abs.v5i64(<5 x i64> undef, i1 false)
+; SLOW-NEXT:  Cost Model: Found an estimated cost of 64 for instruction: %V8I64 = call <8 x i64> @llvm.abs.v8i64(<8 x i64> undef, i1 false)
+; SLOW-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %I32 = call i32 @llvm.abs.i32(i32 undef, i1 false)
+; SLOW-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V2I32 = call <2 x i32> @llvm.abs.v2i32(<2 x i32> undef, i1 false)
+; SLOW-NEXT:  Cost Model: Found an estimated cost of 12 for instruction: %V4I32 = call <4 x i32> @llvm.abs.v4i32(<4 x i32> undef, i1 false)
+; SLOW-NEXT:  Cost Model: Found an estimated cost of 24 for instruction: %V8I32 = call <8 x i32> @llvm.abs.v8i32(<8 x i32> undef, i1 false)
+; SLOW-NEXT:  Cost Model: Found an estimated cost of 45 for instruction: %V9I32 = call <9 x i32> @llvm.abs.v9i32(<9 x i32> undef, i1 false)
+; SLOW-NEXT:  Cost Model: Found an estimated cost of 80 for instruction: %V16I32 = call <16 x i32> @llvm.abs.v16i32(<16 x i32> undef, i1 false)
+; SLOW-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %I16 = call i16 @llvm.abs.i16(i16 undef, i1 false)
+; SLOW-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V2I16 = call <2 x i16> @llvm.abs.v2i16(<2 x i16> undef, i1 false)
+; SLOW-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V4I16 = call <4 x i16> @llvm.abs.v4i16(<4 x i16> undef, i1 false)
+; SLOW-NEXT:  Cost Model: Found an estimated cost of 40 for instruction: %V8I16 = call <8 x i16> @llvm.abs.v8i16(<8 x i16> undef, i1 false)
+; SLOW-NEXT:  Cost Model: Found an estimated cost of 80 for instruction: %V16I16 = call <16 x i16> @llvm.abs.v16i16(<16 x i16> undef, i1 false)
+; SLOW-NEXT:  Cost Model: Found an estimated cost of 102 for instruction: %V17I16 = call <17 x i16> @llvm.abs.v17i16(<17 x i16> undef, i1 false)
+; SLOW-NEXT:  Cost Model: Found an estimated cost of 162 for instruction: %V32I16 = call <32 x i16> @llvm.abs.v32i16(<32 x i16> undef, i1 false)
+; SLOW-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %I8 = call i8 @llvm.abs.i8(i8 undef, i1 false)
+; SLOW-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V2I8 = call <2 x i8> @llvm.abs.v2i8(<2 x i8> undef, i1 false)
+; SLOW-NEXT:  Cost Model: Found an estimated cost of 20 for instruction: %V4I8 = call <4 x i8> @llvm.abs.v4i8(<4 x i8> undef, i1 false)
+; SLOW-NEXT:  Cost Model: Found an estimated cost of 40 for instruction: %V8I8 = call <8 x i8> @llvm.abs.v8i8(<8 x i8> undef, i1 false)
+; SLOW-NEXT:  Cost Model: Found an estimated cost of 80 for instruction: %V16I8 = call <16 x i8> @llvm.abs.v16i8(<16 x i8> undef, i1 false)
+; SLOW-NEXT:  Cost Model: Found an estimated cost of 160 for instruction: %V32I8 = call <32 x i8> @llvm.abs.v32i8(<32 x i8> undef, i1 false)
+; SLOW-NEXT:  Cost Model: Found an estimated cost of 198 for instruction: %V33I8 = call <33 x i8> @llvm.abs.v33i8(<33 x i8> undef, i1 false)
+; SLOW-NEXT:  Cost Model: Found an estimated cost of 322 for instruction: %V64I8 = call <64 x i8> @llvm.abs.v64i8(<64 x i8> undef, i1 false)
+; SLOW-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: ret i32 undef
+;
+; FAST-SIZE-LABEL: 'abs_nonpoison'
+; FAST-SIZE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %I64 = call i64 @llvm.abs.i64(i64 undef, i1 false)
+; FAST-SIZE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V2I64 = call <2 x i64> @llvm.abs.v2i64(<2 x i64> undef, i1 false)
+; FAST-SIZE-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V4I64 = call <4 x i64> @llvm.abs.v4i64(<4 x i64> undef, i1 false)
+; FAST-SIZE-NEXT:  Cost Model: Found an estimated cost of 50 for instruction: %V5I64 = call <5 x i64> @llvm.abs.v5i64(<5 x i64> undef, i1 false)
+; FAST-SIZE-NEXT:  Cost Model: Found an estimated cost of 50 for instruction: %V8I64 = call <8 x i64> @llvm.abs.v8i64(<8 x i64> undef, i1 false)
+; FAST-SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %I32 = call i32 @llvm.abs.i32(i32 undef, i1 false)
+; FAST-SIZE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2I32 = call <2 x i32> @llvm.abs.v2i32(<2 x i32> undef, i1 false)
+; FAST-SIZE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4I32 = call <4 x i32> @llvm.abs.v4i32(<4 x i32> undef, i1 false)
+; FAST-SIZE-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V8I32 = call <8 x i32> @llvm.abs.v8i32(<8 x i32> undef, i1 false)
+; FAST-SIZE-NEXT:  Cost Model: Found an estimated cost of 29 for instruction: %V9I32 = call <9 x i32> @llvm.abs.v9i32(<9 x i32> undef, i1 false)
+; FAST-SIZE-NEXT:  Cost Model: Found an estimated cost of 50 for instruction: %V16I32 = call <16 x i32> @llvm.abs.v16i32(<16 x i32> undef, i1 false)
+; FAST-SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %I16 = call i16 @llvm.abs.i16(i16 undef, i1 false)
+; FAST-SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %V2I16 = call <2 x i16> @llvm.abs.v2i16(<2 x i16> undef, i1 false)
+; FAST-SIZE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V4I16 = call <4 x i16> @llvm.abs.v4i16(<4 x i16> undef, i1 false)
+; FAST-SIZE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V8I16 = call <8 x i16> @llvm.abs.v8i16(<8 x i16> undef, i1 false)
+; FAST-SIZE-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V16I16 = call <16 x i16> @llvm.abs.v16i16(<16 x i16> undef, i1 false)
+; FAST-SIZE-NEXT:  Cost Model: Found an estimated cost of 50 for instruction: %V17I16 = call <17 x i16> @llvm.abs.v17i16(<17 x i16> undef, i1 false)
+; FAST-SIZE-NEXT:  Cost Model: Found an estimated cost of 50 for instruction: %V32I16 = call <32 x i16> @llvm.abs.v32i16(<32 x i16> undef, i1 false)
+; FAST-SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %I8 = call i8 @llvm.abs.i8(i8 undef, i1 false)
+; FAST-SIZE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2I8 = call <2 x i8> @llvm.abs.v2i8(<2 x i8> undef, i1 false)
+; FAST-SIZE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4I8 = call <4 x i8> @llvm.abs.v4i8(<4 x i8> undef, i1 false)
+; FAST-SIZE-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V8I8 = call <8 x i8> @llvm.abs.v8i8(<8 x i8> undef, i1 false)
+; FAST-SIZE-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V16I8 = call <16 x i8> @llvm.abs.v16i8(<16 x i8> undef, i1 false)
+; FAST-SIZE-NEXT:  Cost Model: Found an estimated cost of 34 for instruction: %V32I8 = call <32 x i8> @llvm.abs.v32i8(<32 x i8> undef, i1 false)
+; FAST-SIZE-NEXT:  Cost Model: Found an estimated cost of 68 for instruction: %V33I8 = call <33 x i8> @llvm.abs.v33i8(<33 x i8> undef, i1 false)
+; FAST-SIZE-NEXT:  Cost Model: Found an estimated cost of 68 for instruction: %V64I8 = call <64 x i8> @llvm.abs.v64i8(<64 x i8> undef, i1 false)
+; FAST-SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+; SLOW-SIZE-LABEL: 'abs_nonpoison'
+; SLOW-SIZE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %I64 = call i64 @llvm.abs.i64(i64 undef, i1 false)
+; SLOW-SIZE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V2I64 = call <2 x i64> @llvm.abs.v2i64(<2 x i64> undef, i1 false)
+; SLOW-SIZE-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V4I64 = call <4 x i64> @llvm.abs.v4i64(<4 x i64> undef, i1 false)
+; SLOW-SIZE-NEXT:  Cost Model: Found an estimated cost of 50 for instruction: %V5I64 = call <5 x i64> @llvm.abs.v5i64(<5 x i64> undef, i1 false)
+; SLOW-SIZE-NEXT:  Cost Model: Found an estimated cost of 50 for instruction: %V8I64 = call <8 x i64> @llvm.abs.v8i64(<8 x i64> undef, i1 false)
+; SLOW-SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %I32 = call i32 @llvm.abs.i32(i32 undef, i1 false)
+; SLOW-SIZE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2I32 = call <2 x i32> @llvm.abs.v2i32(<2 x i32> undef, i1 false)
+; SLOW-SIZE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4I32 = call <4 x i32> @llvm.abs.v4i32(<4 x i32> undef, i1 false)
+; SLOW-SIZE-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V8I32 = call <8 x i32> @llvm.abs.v8i32(<8 x i32> undef, i1 false)
+; SLOW-SIZE-NEXT:  Cost Model: Found an estimated cost of 29 for instruction: %V9I32 = call <9 x i32> @llvm.abs.v9i32(<9 x i32> undef, i1 false)
+; SLOW-SIZE-NEXT:  Cost Model: Found an estimated cost of 50 for instruction: %V16I32 = call <16 x i32> @llvm.abs.v16i32(<16 x i32> undef, i1 false)
+; SLOW-SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %I16 = call i16 @llvm.abs.i16(i16 undef, i1 false)
+; SLOW-SIZE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2I16 = call <2 x i16> @llvm.abs.v2i16(<2 x i16> undef, i1 false)
+; SLOW-SIZE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4I16 = call <4 x i16> @llvm.abs.v4i16(<4 x i16> undef, i1 false)
+; SLOW-SIZE-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V8I16 = call <8 x i16> @llvm.abs.v8i16(<8 x i16> undef, i1 false)
+; SLOW-SIZE-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V16I16 = call <16 x i16> @llvm.abs.v16i16(<16 x i16> undef, i1 false)
+; SLOW-SIZE-NEXT:  Cost Model: Found an estimated cost of 36 for instruction: %V17I16 = call <17 x i16> @llvm.abs.v17i16(<17 x i16> undef, i1 false)
+; SLOW-SIZE-NEXT:  Cost Model: Found an estimated cost of 36 for instruction: %V32I16 = call <32 x i16> @llvm.abs.v32i16(<32 x i16> undef, i1 false)
+; SLOW-SIZE-NEXT:  Cost Model: Found an estimated cost of 3 for instruction: %I8 = call i8 @llvm.abs.i8(i8 undef, i1 false)
+; SLOW-SIZE-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %V2I8 = call <2 x i8> @llvm.abs.v2i8(<2 x i8> undef, i1 false)
+; SLOW-SIZE-NEXT:  Cost Model: Found an estimated cost of 6 for instruction: %V4I8 = call <4 x i8> @llvm.abs.v4i8(<4 x i8> undef, i1 false)
+; SLOW-SIZE-NEXT:  Cost Model: Found an estimated cost of 10 for instruction: %V8I8 = call <8 x i8> @llvm.abs.v8i8(<8 x i8> undef, i1 false)
+; SLOW-SIZE-NEXT:  Cost Model: Found an estimated cost of 18 for instruction: %V16I8 = call <16 x i8> @llvm.abs.v16i8(<16 x i8> undef, i1 false)
+; SLOW-SIZE-NEXT:  Cost Model: Found an estimated cost of 34 for instruction: %V32I8 = call <32 x i8> @llvm.abs.v32i8(<32 x i8> undef, i1 false)
+; SLOW-SIZE-NEXT:  Cost Model: Found an estimated cost of 68 for instruction: %V33I8 = call <33 x i8> @llvm.abs.v33i8(<33 x i8> undef, i1 false)
+; SLOW-SIZE-NEXT:  Cost Model: Found an estimated cost of 68 for instruction: %V64I8 = call <64 x i8> @llvm.abs.v64i8(<64 x i8> undef, i1 false)
+; SLOW-SIZE-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
+;
+  %I64 = call i64 @llvm.abs.i64(i64 undef, i1 false)
+  %V2I64 = call <2 x i64> @llvm.abs.v2i64(<2 x i64> undef, i1 false)
+  %V4I64 = call <4 x i64> @llvm.abs.v4i64(<4 x i64> undef, i1 false)
+  %V5I64 = call <5 x i64> @llvm.abs.v5i64(<5 x i64> undef, i1 false)
+  %V8I64 = call <8 x i64> @llvm.abs.v8i64(<8 x i64> undef, i1 false)
+
+  %I32 = call i32 @llvm.abs.i32(i32 undef, i1 false)
+  %V2I32  = call <2 x i32>  @llvm.abs.v2i32(<2 x i32> undef, i1 false)
+  %V4I32  = call <4 x i32>  @llvm.abs.v4i32(<4 x i32> undef, i1 false)
+  %V8I32  = call <8 x i32>  @llvm.abs.v8i32(<8 x i32> undef, i1 false)
+  %V9I32  = call <9 x i32>  @llvm.abs.v9i32(<9 x i32> undef, i1 false)
+  %V16I32 = call <16 x i32> @llvm.abs.v16i32(<16 x i32> undef, i1 false)
+
+  %I16 = call i16 @llvm.abs.i16(i16 undef, i1 false)
+  %V2I16  = call <2 x i16>  @llvm.abs.v2i16(<2 x i16> undef, i1 false)
+  %V4I16  = call <4 x i16>  @llvm.abs.v4i16(<4 x i16> undef, i1 false)
+  %V8I16  = call <8 x i16>  @llvm.abs.v8i16(<8 x i16> undef, i1 false)
+  %V16I16 = call <16 x i16> @llvm.abs.v16i16(<16 x i16> undef, i1 false)
+  %V17I16 = call <17 x i16> @llvm.abs.v17i16(<17 x i16> undef, i1 false)
+  %V32I16 = call <32 x i16> @llvm.abs.v32i16(<32 x i16> undef, i1 false)
+
+  %I8 = call i8 @llvm.abs.i8(i8 undef, i1 false)
+  %V2I8  = call <2 x i8>  @llvm.abs.v2i8(<2 x i8> undef, i1 false)
+  %V4I8  = call <4 x i8>  @llvm.abs.v4i8(<4 x i8> undef, i1 false)
+  %V8I8  = call <8 x i8>  @llvm.abs.v8i8(<8 x i8> undef, i1 false)
+  %V16I8 = call <16 x i8> @llvm.abs.v16i8(<16 x i8> undef, i1 false)
+  %V32I8 = call <32 x i8> @llvm.abs.v32i8(<32 x i8> undef, i1 false)
+  %V33I8 = call <33 x i8> @llvm.abs.v33i8(<33 x i8> undef, i1 false)
+  %V64I8 = call <64 x i8> @llvm.abs.v64i8(<64 x i8> undef, i1 false)
+
+  ret i32 undef
+}
+
+define i32 @abs_poison(i32 %arg) {
+; FAST-LABEL: 'abs_poison'
+; FAST-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %I64 = call i64 @llvm.abs.i64(i64 undef, i1 true)
+; FAST-NEXT:  Cost Model: Found an estimated cost of 8 for instruction: %V2I64 = call <2 x i64> @llvm.abs.v2i64(<2 x i64> undef, i1 true)
+; FAST-NEXT:  Cost Model: Found an estimated cost of 16 for instruction: %V4I64 = call <4 x i64> ...
[truncated]

@arsenm arsenm force-pushed the users/arsenm/amdgpu-add-baseline-tti-cost-abs branch from 330c0e2 to df2b6b7 Compare July 25, 2024 17:27
@arsenm arsenm changed the base branch from users/arsenm/tti-check-fptoi-sat-legalize-costs to users/arsenm/amdgpu-add-slp-vectorize-test-integer-min-max July 25, 2024 17:27
@arsenm Graphite App
Copy link
Contributor Author

arsenm commented Jul 25, 2024 •
edited
Loading

Merge activity

  • Jul 25, 4:25 PM EDT: @arsenm started a stack merge that includes this pull request via Graphite.
  • Jul 25, 4:46 PM EDT: Graphite rebased this pull request as part of a merge.
  • Jul 25, 4:49 PM EDT: Graphite rebased this pull request as part of a merge.
  • Jul 25, 4:51 PM EDT: @arsenm merged this pull request with Graphite.

@arsenm arsenm force-pushed the users/arsenm/amdgpu-add-slp-vectorize-test-integer-min-max branch 2 times, most recently from 9876e76 to e7d5b89 Compare July 25, 2024 20:31
@ldionne ldionne removed request for a team, ayermolo and tbaederr July 25, 2024 20:31
@arsenm arsenm force-pushed the users/arsenm/amdgpu-add-slp-vectorize-test-integer-min-max branch 3 times, most recently from 6561bfd to cc1245c Compare July 25, 2024 20:41
Base automatically changed from users/arsenm/amdgpu-add-slp-vectorize-test-integer-min-max to main July 25, 2024 20:45
@arsenm arsenm force-pushed the users/arsenm/amdgpu-add-baseline-tti-cost-abs branch from df2b6b7 to eb121f6 Compare July 25, 2024 20:45
@arsenm arsenm force-pushed the users/arsenm/amdgpu-add-baseline-tti-cost-abs branch from eb121f6 to c9b0504 Compare July 25, 2024 20:49
@arsenm arsenm merged commit 9720690 into main Jul 25, 2024
4 of 6 checks passed
@arsenm arsenm deleted the users/arsenm/amdgpu-add-baseline-tti-cost-abs branch July 25, 2024 20:51
@arsenm arsenm mentioned this pull request Jul 26, 2024
Merged
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@alexey-bataev alexey-bataev Awaiting requested review from alexey-bataev

@nikic nikic Awaiting requested review from nikic

@RKSimon RKSimon Awaiting requested review from RKSimon

@rotateright rotateright Awaiting requested review from rotateright

@sdesmalen-arm sdesmalen-arm Awaiting requested review from sdesmalen-arm

@aaupov aaupov Awaiting requested review from aaupov

@maksfb maksfb Awaiting requested review from maksfb

@rafaelauler rafaelauler Awaiting requested review from rafaelauler

@dcci dcci Awaiting requested review from dcci

@banach-space banach-space Awaiting requested review from banach-space

@dcaballe dcaballe Awaiting requested review from dcaballe

@nicolasvasilache nicolasvasilache Awaiting requested review from nicolasvasilache

@hanhanW hanhanW Awaiting requested review from hanhanW

@matthias-springer matthias-springer Awaiting requested review from matthias-springer

@JDevlieghere JDevlieghere Awaiting requested review from JDevlieghere

Assignees
No one assigned
Labels
backend:AMDGPU llvm:analysis Includes value tracking, cost tables and constant folding llvm:transforms vectorizers
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
2 participants
@arsenm @llvmbot