mirror of
https://github.com/RPCS3/llvm-mirror.git
synced 2024-11-22 18:54:02 +01:00
6e7877566a
As noticed on D105432 we didn't have any coverage to distinguish between fast/exact float reductions
166 lines
14 KiB
LLVM
166 lines
14 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_analyze_test_checks.py
|
|
; RUN: opt < %s -cost-model -mtriple=x86_64-apple-darwin -analyze -mattr=+sse2 | FileCheck %s --check-prefixes=SSE
|
|
; RUN: opt < %s -cost-model -mtriple=x86_64-apple-darwin -analyze -mattr=+ssse3 | FileCheck %s --check-prefixes=SSE
|
|
; RUN: opt < %s -cost-model -mtriple=x86_64-apple-darwin -analyze -mattr=+sse4.1 | FileCheck %s --check-prefixes=SSE
|
|
; RUN: opt < %s -cost-model -mtriple=x86_64-apple-darwin -analyze -mattr=+sse4.2 | FileCheck %s --check-prefixes=SSE
|
|
; RUN: opt < %s -cost-model -mtriple=x86_64-apple-darwin -analyze -mattr=+avx | FileCheck %s --check-prefixes=AVX
|
|
; RUN: opt < %s -cost-model -mtriple=x86_64-apple-darwin -analyze -mattr=+avx2 | FileCheck %s --check-prefixes=AVX
|
|
; RUN: opt < %s -cost-model -mtriple=x86_64-apple-darwin -analyze -mattr=+avx512f | FileCheck %s --check-prefixes=AVX512
|
|
; RUN: opt < %s -cost-model -mtriple=x86_64-apple-darwin -analyze -mattr=+avx512f,+avx512bw | FileCheck %s --check-prefixes=AVX512
|
|
; RUN: opt < %s -cost-model -mtriple=x86_64-apple-darwin -analyze -mattr=+avx512f,+avx512dq | FileCheck %s --check-prefixes=AVX512
|
|
|
|
define i32 @reduce_f64(i32 %arg) {
|
|
; SSE-LABEL: 'reduce_f64'
|
|
; SSE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %V1 = call double @llvm.vector.reduce.fmin.v1f64(<1 x double> undef)
|
|
; SSE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2 = call double @llvm.vector.reduce.fmin.v2f64(<2 x double> undef)
|
|
; SSE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V4 = call double @llvm.vector.reduce.fmin.v4f64(<4 x double> undef)
|
|
; SSE-NEXT: Cost Model: Found an estimated cost of 5 for instruction: %V8 = call double @llvm.vector.reduce.fmin.v8f64(<8 x double> undef)
|
|
; SSE-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V16 = call double @llvm.vector.reduce.fmin.v16f64(<16 x double> undef)
|
|
; SSE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
|
|
;
|
|
; AVX-LABEL: 'reduce_f64'
|
|
; AVX-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %V1 = call double @llvm.vector.reduce.fmin.v1f64(<1 x double> undef)
|
|
; AVX-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2 = call double @llvm.vector.reduce.fmin.v2f64(<2 x double> undef)
|
|
; AVX-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V4 = call double @llvm.vector.reduce.fmin.v4f64(<4 x double> undef)
|
|
; AVX-NEXT: Cost Model: Found an estimated cost of 5 for instruction: %V8 = call double @llvm.vector.reduce.fmin.v8f64(<8 x double> undef)
|
|
; AVX-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V16 = call double @llvm.vector.reduce.fmin.v16f64(<16 x double> undef)
|
|
; AVX-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
|
|
;
|
|
; AVX512-LABEL: 'reduce_f64'
|
|
; AVX512-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %V1 = call double @llvm.vector.reduce.fmin.v1f64(<1 x double> undef)
|
|
; AVX512-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2 = call double @llvm.vector.reduce.fmin.v2f64(<2 x double> undef)
|
|
; AVX512-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V4 = call double @llvm.vector.reduce.fmin.v4f64(<4 x double> undef)
|
|
; AVX512-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V8 = call double @llvm.vector.reduce.fmin.v8f64(<8 x double> undef)
|
|
; AVX512-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V16 = call double @llvm.vector.reduce.fmin.v16f64(<16 x double> undef)
|
|
; AVX512-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
|
|
;
|
|
%V1 = call double @llvm.vector.reduce.fmin.v1f64(<1 x double> undef)
|
|
%V2 = call double @llvm.vector.reduce.fmin.v2f64(<2 x double> undef)
|
|
%V4 = call double @llvm.vector.reduce.fmin.v4f64(<4 x double> undef)
|
|
%V8 = call double @llvm.vector.reduce.fmin.v8f64(<8 x double> undef)
|
|
%V16 = call double @llvm.vector.reduce.fmin.v16f64(<16 x double> undef)
|
|
ret i32 undef
|
|
}
|
|
|
|
define i32 @reduce_f32(i32 %arg) {
|
|
; SSE-LABEL: 'reduce_f32'
|
|
; SSE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %V1 = call float @llvm.vector.reduce.fmin.v1f32(<1 x float> undef)
|
|
; SSE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2 = call float @llvm.vector.reduce.fmin.v2f32(<2 x float> undef)
|
|
; SSE-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V4 = call float @llvm.vector.reduce.fmin.v4f32(<4 x float> undef)
|
|
; SSE-NEXT: Cost Model: Found an estimated cost of 5 for instruction: %V8 = call float @llvm.vector.reduce.fmin.v8f32(<8 x float> undef)
|
|
; SSE-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V16 = call float @llvm.vector.reduce.fmin.v16f32(<16 x float> undef)
|
|
; SSE-NEXT: Cost Model: Found an estimated cost of 11 for instruction: %V32 = call float @llvm.vector.reduce.fmin.v32f32(<32 x float> undef)
|
|
; SSE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
|
|
;
|
|
; AVX-LABEL: 'reduce_f32'
|
|
; AVX-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %V1 = call float @llvm.vector.reduce.fmin.v1f32(<1 x float> undef)
|
|
; AVX-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2 = call float @llvm.vector.reduce.fmin.v2f32(<2 x float> undef)
|
|
; AVX-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V4 = call float @llvm.vector.reduce.fmin.v4f32(<4 x float> undef)
|
|
; AVX-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V8 = call float @llvm.vector.reduce.fmin.v8f32(<8 x float> undef)
|
|
; AVX-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V16 = call float @llvm.vector.reduce.fmin.v16f32(<16 x float> undef)
|
|
; AVX-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V32 = call float @llvm.vector.reduce.fmin.v32f32(<32 x float> undef)
|
|
; AVX-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
|
|
;
|
|
; AVX512-LABEL: 'reduce_f32'
|
|
; AVX512-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %V1 = call float @llvm.vector.reduce.fmin.v1f32(<1 x float> undef)
|
|
; AVX512-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2 = call float @llvm.vector.reduce.fmin.v2f32(<2 x float> undef)
|
|
; AVX512-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V4 = call float @llvm.vector.reduce.fmin.v4f32(<4 x float> undef)
|
|
; AVX512-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V8 = call float @llvm.vector.reduce.fmin.v8f32(<8 x float> undef)
|
|
; AVX512-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V16 = call float @llvm.vector.reduce.fmin.v16f32(<16 x float> undef)
|
|
; AVX512-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V32 = call float @llvm.vector.reduce.fmin.v32f32(<32 x float> undef)
|
|
; AVX512-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
|
|
;
|
|
%V1 = call float @llvm.vector.reduce.fmin.v1f32(<1 x float> undef)
|
|
%V2 = call float @llvm.vector.reduce.fmin.v2f32(<2 x float> undef)
|
|
%V4 = call float @llvm.vector.reduce.fmin.v4f32(<4 x float> undef)
|
|
%V8 = call float @llvm.vector.reduce.fmin.v8f32(<8 x float> undef)
|
|
%V16 = call float @llvm.vector.reduce.fmin.v16f32(<16 x float> undef)
|
|
%V32 = call float @llvm.vector.reduce.fmin.v32f32(<32 x float> undef)
|
|
ret i32 undef
|
|
}
|
|
|
|
; Fast Reductions
|
|
|
|
define i32 @reduce_f64_fast(i32 %arg) {
|
|
; SSE-LABEL: 'reduce_f64_fast'
|
|
; SSE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %V1 = call fast double @llvm.vector.reduce.fmin.v1f64(<1 x double> undef)
|
|
; SSE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2 = call fast double @llvm.vector.reduce.fmin.v2f64(<2 x double> undef)
|
|
; SSE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V4 = call fast double @llvm.vector.reduce.fmin.v4f64(<4 x double> undef)
|
|
; SSE-NEXT: Cost Model: Found an estimated cost of 5 for instruction: %V8 = call fast double @llvm.vector.reduce.fmin.v8f64(<8 x double> undef)
|
|
; SSE-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V16 = call fast double @llvm.vector.reduce.fmin.v16f64(<16 x double> undef)
|
|
; SSE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
|
|
;
|
|
; AVX-LABEL: 'reduce_f64_fast'
|
|
; AVX-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %V1 = call fast double @llvm.vector.reduce.fmin.v1f64(<1 x double> undef)
|
|
; AVX-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2 = call fast double @llvm.vector.reduce.fmin.v2f64(<2 x double> undef)
|
|
; AVX-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V4 = call fast double @llvm.vector.reduce.fmin.v4f64(<4 x double> undef)
|
|
; AVX-NEXT: Cost Model: Found an estimated cost of 5 for instruction: %V8 = call fast double @llvm.vector.reduce.fmin.v8f64(<8 x double> undef)
|
|
; AVX-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V16 = call fast double @llvm.vector.reduce.fmin.v16f64(<16 x double> undef)
|
|
; AVX-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
|
|
;
|
|
; AVX512-LABEL: 'reduce_f64_fast'
|
|
; AVX512-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %V1 = call fast double @llvm.vector.reduce.fmin.v1f64(<1 x double> undef)
|
|
; AVX512-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2 = call fast double @llvm.vector.reduce.fmin.v2f64(<2 x double> undef)
|
|
; AVX512-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V4 = call fast double @llvm.vector.reduce.fmin.v4f64(<4 x double> undef)
|
|
; AVX512-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V8 = call fast double @llvm.vector.reduce.fmin.v8f64(<8 x double> undef)
|
|
; AVX512-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V16 = call fast double @llvm.vector.reduce.fmin.v16f64(<16 x double> undef)
|
|
; AVX512-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
|
|
;
|
|
%V1 = call fast double @llvm.vector.reduce.fmin.v1f64(<1 x double> undef)
|
|
%V2 = call fast double @llvm.vector.reduce.fmin.v2f64(<2 x double> undef)
|
|
%V4 = call fast double @llvm.vector.reduce.fmin.v4f64(<4 x double> undef)
|
|
%V8 = call fast double @llvm.vector.reduce.fmin.v8f64(<8 x double> undef)
|
|
%V16 = call fast double @llvm.vector.reduce.fmin.v16f64(<16 x double> undef)
|
|
ret i32 undef
|
|
}
|
|
|
|
define i32 @reduce_f32_fast(i32 %arg) {
|
|
; SSE-LABEL: 'reduce_f32_fast'
|
|
; SSE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %V1 = call fast float @llvm.vector.reduce.fmin.v1f32(<1 x float> undef)
|
|
; SSE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2 = call fast float @llvm.vector.reduce.fmin.v2f32(<2 x float> undef)
|
|
; SSE-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V4 = call fast float @llvm.vector.reduce.fmin.v4f32(<4 x float> undef)
|
|
; SSE-NEXT: Cost Model: Found an estimated cost of 5 for instruction: %V8 = call fast float @llvm.vector.reduce.fmin.v8f32(<8 x float> undef)
|
|
; SSE-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V16 = call fast float @llvm.vector.reduce.fmin.v16f32(<16 x float> undef)
|
|
; SSE-NEXT: Cost Model: Found an estimated cost of 11 for instruction: %V32 = call fast float @llvm.vector.reduce.fmin.v32f32(<32 x float> undef)
|
|
; SSE-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
|
|
;
|
|
; AVX-LABEL: 'reduce_f32_fast'
|
|
; AVX-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %V1 = call fast float @llvm.vector.reduce.fmin.v1f32(<1 x float> undef)
|
|
; AVX-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2 = call fast float @llvm.vector.reduce.fmin.v2f32(<2 x float> undef)
|
|
; AVX-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V4 = call fast float @llvm.vector.reduce.fmin.v4f32(<4 x float> undef)
|
|
; AVX-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V8 = call fast float @llvm.vector.reduce.fmin.v8f32(<8 x float> undef)
|
|
; AVX-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V16 = call fast float @llvm.vector.reduce.fmin.v16f32(<16 x float> undef)
|
|
; AVX-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V32 = call fast float @llvm.vector.reduce.fmin.v32f32(<32 x float> undef)
|
|
; AVX-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
|
|
;
|
|
; AVX512-LABEL: 'reduce_f32_fast'
|
|
; AVX512-NEXT: Cost Model: Found an estimated cost of 0 for instruction: %V1 = call fast float @llvm.vector.reduce.fmin.v1f32(<1 x float> undef)
|
|
; AVX512-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2 = call fast float @llvm.vector.reduce.fmin.v2f32(<2 x float> undef)
|
|
; AVX512-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V4 = call fast float @llvm.vector.reduce.fmin.v4f32(<4 x float> undef)
|
|
; AVX512-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V8 = call fast float @llvm.vector.reduce.fmin.v8f32(<8 x float> undef)
|
|
; AVX512-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V16 = call fast float @llvm.vector.reduce.fmin.v16f32(<16 x float> undef)
|
|
; AVX512-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V32 = call fast float @llvm.vector.reduce.fmin.v32f32(<32 x float> undef)
|
|
; AVX512-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
|
|
;
|
|
%V1 = call fast float @llvm.vector.reduce.fmin.v1f32(<1 x float> undef)
|
|
%V2 = call fast float @llvm.vector.reduce.fmin.v2f32(<2 x float> undef)
|
|
%V4 = call fast float @llvm.vector.reduce.fmin.v4f32(<4 x float> undef)
|
|
%V8 = call fast float @llvm.vector.reduce.fmin.v8f32(<8 x float> undef)
|
|
%V16 = call fast float @llvm.vector.reduce.fmin.v16f32(<16 x float> undef)
|
|
%V32 = call fast float @llvm.vector.reduce.fmin.v32f32(<32 x float> undef)
|
|
ret i32 undef
|
|
}
|
|
|
|
declare double @llvm.vector.reduce.fmin.v1f64(<1 x double>)
|
|
declare double @llvm.vector.reduce.fmin.v2f64(<2 x double>)
|
|
declare double @llvm.vector.reduce.fmin.v4f64(<4 x double>)
|
|
declare double @llvm.vector.reduce.fmin.v8f64(<8 x double>)
|
|
declare double @llvm.vector.reduce.fmin.v16f64(<16 x double>)
|
|
|
|
declare float @llvm.vector.reduce.fmin.v1f32(<1 x float>)
|
|
declare float @llvm.vector.reduce.fmin.v2f32(<2 x float>)
|
|
declare float @llvm.vector.reduce.fmin.v4f32(<4 x float>)
|
|
declare float @llvm.vector.reduce.fmin.v8f32(<8 x float>)
|
|
declare float @llvm.vector.reduce.fmin.v16f32(<16 x float>)
|
|
declare float @llvm.vector.reduce.fmin.v32f32(<32 x float>)
|