derek.holloway
7502bf64a4
CodeQL / Analyze (actions) (push) Canceled after 0s
CodeQL / Analyze (python) (push) Canceled after 0s
pre-commit / pre-commit (push) Canceled after 0s
Multi-Arch CI / setup (push) Canceled after 0s
Publish build_manylinux_x86_64 images / publish_build_manylinux_x86_64 (push) Canceled after 0s
Setup (Multi-Arch) / setup (push) Canceled after 0s
Multi-Arch CI / Linux::${{ fromJSON(needs.setup.outputs.linux_build_config || '{}').build_variant_label || 'skip' }} (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish no_rocm_image_ubi10 image / publish_no_rocm_image_ubi10 (push) Canceled after 0s
Multi-Arch CI / Windows::${{ fromJSON(needs.setup.outputs.windows_build_config || '{}').build_variant_label || 'skip' }} (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Multi-Arch CI / Manifest Diff (push) Canceled after 0s
TheRock Manifest Diff Report / Generate Manifest Diff Report (push) Canceled after 0s
Multi-Arch CI / CI Summary (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04 (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04_ocl_rt images / publish_no_rocm_image_ubuntu24_04_ocl_rt (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_rocgdb (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04_rocgdb images / publish_no_rocm_image_ubuntu24_04_rocgdb (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_ocl_rt (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Unit Tests / Unit Tests :: ubuntu-24.04 (push) Canceled after 0s
Unit Tests / Unit Tests :: windows-2022 (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_media (push) Canceled after 0s
Unit Tests / Unit Tests Summary (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_openmpi (push) Canceled after 0s
190 lines
7.5 KiB
LLVM
190 lines
7.5 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --filter "(callq|ptrue)" --version 5
|
|
; RUN: llc -mtriple=x86_64-unknown-linux-gnu -mcpu=x86-64 -vector-library=AMDLIBM < %s | FileCheck %s --check-prefixes=CHECK,AMD,AMD-SSE
|
|
; RUN: llc -mtriple=x86_64-unknown-linux-gnu -mcpu=x86-64-v2 -vector-library=AMDLIBM < %s | FileCheck %s --check-prefixes=CHECK,AMD,AMD-AVX
|
|
; RUN: llc -mtriple=x86_64-unknown-linux-gnu -mcpu=x86-64-v3 -vector-library=AMDLIBM < %s | FileCheck %s --check-prefixes=CHECK,AMD,AMD-AVX2
|
|
; RUN: llc -mtriple=x86_64-unknown-linux-gnu -mcpu=x86-64-v4 -vector-library=AMDLIBM < %s | FileCheck %s --check-prefixes=CHECK,AMD,AMD-AVX512
|
|
; RUN: llc -mtriple=x86_64-unknown-linux-gnu -mcpu=x86-64 -vector-library=LIBMVEC < %s | FileCheck %s --check-prefixes=CHECK,GLIBC,GLIBC-SSE
|
|
; RUN: llc -mtriple=x86_64-unknown-linux-gnu -mcpu=x86-64-v2 -vector-library=LIBMVEC < %s | FileCheck %s --check-prefixes=CHECK,GLIBC,GLIBC-AVX
|
|
; RUN: llc -mtriple=x86_64-unknown-linux-gnu -mcpu=x86-64-v3 -vector-library=LIBMVEC < %s | FileCheck %s --check-prefixes=CHECK,GLIBC,GLIBC-AVX2
|
|
; RUN: llc -mtriple=x86_64-unknown-linux-gnu -mcpu=x86-64-v4 -vector-library=LIBMVEC < %s | FileCheck %s --check-prefixes=CHECK,GLIBC,GLIBC-AVX512
|
|
|
|
define void @test_sincos_v4f32(<4 x float> %x, ptr noalias %out_sin, ptr noalias %out_cos) {
|
|
; AMD-LABEL: test_sincos_v4f32:
|
|
; AMD: callq amd_vrs4_sincosf@PLT
|
|
;
|
|
; GLIBC-LABEL: test_sincos_v4f32:
|
|
; GLIBC: callq sincosf@PLT
|
|
; GLIBC: callq sincosf@PLT
|
|
; GLIBC: callq sincosf@PLT
|
|
; GLIBC: callq sincosf@PLT
|
|
%result = call { <4 x float>, <4 x float> } @llvm.sincos.v4f32(<4 x float> %x)
|
|
%result.0 = extractvalue { <4 x float>, <4 x float> } %result, 0
|
|
%result.1 = extractvalue { <4 x float>, <4 x float> } %result, 1
|
|
store <4 x float> %result.0, ptr %out_sin, align 16
|
|
store <4 x float> %result.1, ptr %out_cos, align 16
|
|
ret void
|
|
}
|
|
|
|
define void @test_sincos_v8f32(<8 x float> %x, ptr noalias %out_sin, ptr noalias %out_cos) {
|
|
; AMD-SSE-LABEL: test_sincos_v8f32:
|
|
; AMD-SSE: callq amd_vrs4_sincosf@PLT
|
|
; AMD-SSE: callq amd_vrs4_sincosf@PLT
|
|
;
|
|
; AMD-AVX-LABEL: test_sincos_v8f32:
|
|
; AMD-AVX: callq amd_vrs4_sincosf@PLT
|
|
; AMD-AVX: callq amd_vrs4_sincosf@PLT
|
|
;
|
|
; AMD-AVX2-LABEL: test_sincos_v8f32:
|
|
; AMD-AVX2: callq amd_vrs8_sincosf@PLT
|
|
;
|
|
; AMD-AVX512-LABEL: test_sincos_v8f32:
|
|
; AMD-AVX512: callq amd_vrs8_sincosf@PLT
|
|
;
|
|
; GLIBC-LABEL: test_sincos_v8f32:
|
|
; GLIBC: callq sincosf@PLT
|
|
; GLIBC: callq sincosf@PLT
|
|
; GLIBC: callq sincosf@PLT
|
|
; GLIBC: callq sincosf@PLT
|
|
; GLIBC: callq sincosf@PLT
|
|
; GLIBC: callq sincosf@PLT
|
|
; GLIBC: callq sincosf@PLT
|
|
; GLIBC: callq sincosf@PLT
|
|
%result = call { <8 x float>, <8 x float> } @llvm.sincos.v8f32(<8 x float> %x)
|
|
%result.0 = extractvalue { <8 x float>, <8 x float> } %result, 0
|
|
%result.1 = extractvalue { <8 x float>, <8 x float> } %result, 1
|
|
store <8 x float> %result.0, ptr %out_sin, align 32
|
|
store <8 x float> %result.1, ptr %out_cos, align 32
|
|
ret void
|
|
}
|
|
|
|
define void @test_sincos_v16f32(<16 x float> %x, ptr noalias %out_sin, ptr noalias %out_cos) {
|
|
; AMD-SSE-LABEL: test_sincos_v16f32:
|
|
; AMD-SSE: callq amd_vrs4_sincosf@PLT
|
|
; AMD-SSE: callq amd_vrs4_sincosf@PLT
|
|
; AMD-SSE: callq amd_vrs4_sincosf@PLT
|
|
; AMD-SSE: callq amd_vrs4_sincosf@PLT
|
|
;
|
|
; AMD-AVX-LABEL: test_sincos_v16f32:
|
|
; AMD-AVX: callq amd_vrs4_sincosf@PLT
|
|
; AMD-AVX: callq amd_vrs4_sincosf@PLT
|
|
; AMD-AVX: callq amd_vrs4_sincosf@PLT
|
|
; AMD-AVX: callq amd_vrs4_sincosf@PLT
|
|
;
|
|
; AMD-AVX2-LABEL: test_sincos_v16f32:
|
|
; AMD-AVX2: callq amd_vrs8_sincosf@PLT
|
|
; AMD-AVX2: callq amd_vrs8_sincosf@PLT
|
|
;
|
|
; AMD-AVX512-LABEL: test_sincos_v16f32:
|
|
; AMD-AVX512: callq amd_vrs16_sincosf@PLT
|
|
;
|
|
; GLIBC-LABEL: test_sincos_v16f32:
|
|
; GLIBC: callq sincosf@PLT
|
|
; GLIBC: callq sincosf@PLT
|
|
; GLIBC: callq sincosf@PLT
|
|
; GLIBC: callq sincosf@PLT
|
|
; GLIBC: callq sincosf@PLT
|
|
; GLIBC: callq sincosf@PLT
|
|
; GLIBC: callq sincosf@PLT
|
|
; GLIBC: callq sincosf@PLT
|
|
; GLIBC: callq sincosf@PLT
|
|
; GLIBC: callq sincosf@PLT
|
|
; GLIBC: callq sincosf@PLT
|
|
; GLIBC: callq sincosf@PLT
|
|
; GLIBC: callq sincosf@PLT
|
|
; GLIBC: callq sincosf@PLT
|
|
; GLIBC: callq sincosf@PLT
|
|
; GLIBC: callq sincosf@PLT
|
|
%result = call { <16 x float>, <16 x float> } @llvm.sincos.v16f32(<16 x float> %x)
|
|
%result.0 = extractvalue { <16 x float>, <16 x float> } %result, 0
|
|
%result.1 = extractvalue { <16 x float>, <16 x float> } %result, 1
|
|
store <16 x float> %result.0, ptr %out_sin, align 64
|
|
store <16 x float> %result.1, ptr %out_cos, align 64
|
|
ret void
|
|
}
|
|
|
|
define void @test_sincos_v2f64(<2 x double> %x, ptr noalias %out_sin, ptr noalias %out_cos) {
|
|
; AMD-LABEL: test_sincos_v2f64:
|
|
; AMD: callq amd_vrd2_sincos@PLT
|
|
;
|
|
; GLIBC-LABEL: test_sincos_v2f64:
|
|
; GLIBC: callq sincos@PLT
|
|
; GLIBC: callq sincos@PLT
|
|
%result = call { <2 x double>, <2 x double> } @llvm.sincos.v2f64(<2 x double> %x)
|
|
%result.0 = extractvalue { <2 x double>, <2 x double> } %result, 0
|
|
%result.1 = extractvalue { <2 x double>, <2 x double> } %result, 1
|
|
store <2 x double> %result.0, ptr %out_sin, align 16
|
|
store <2 x double> %result.1, ptr %out_cos, align 16
|
|
ret void
|
|
}
|
|
|
|
define void @test_sincos_v4f64(<4 x double> %x, ptr noalias %out_sin, ptr noalias %out_cos) {
|
|
; AMD-SSE-LABEL: test_sincos_v4f64:
|
|
; AMD-SSE: callq amd_vrd2_sincos@PLT
|
|
; AMD-SSE: callq amd_vrd2_sincos@PLT
|
|
;
|
|
; AMD-AVX-LABEL: test_sincos_v4f64:
|
|
; AMD-AVX: callq amd_vrd2_sincos@PLT
|
|
; AMD-AVX: callq amd_vrd2_sincos@PLT
|
|
;
|
|
; AMD-AVX2-LABEL: test_sincos_v4f64:
|
|
; AMD-AVX2: callq amd_vrd4_sincos@PLT
|
|
;
|
|
; AMD-AVX512-LABEL: test_sincos_v4f64:
|
|
; AMD-AVX512: callq amd_vrd4_sincos@PLT
|
|
;
|
|
; GLIBC-LABEL: test_sincos_v4f64:
|
|
; GLIBC: callq sincos@PLT
|
|
; GLIBC: callq sincos@PLT
|
|
; GLIBC: callq sincos@PLT
|
|
; GLIBC: callq sincos@PLT
|
|
%result = call { <4 x double>, <4 x double> } @llvm.sincos.v4f64(<4 x double> %x)
|
|
%result.0 = extractvalue { <4 x double>, <4 x double> } %result, 0
|
|
%result.1 = extractvalue { <4 x double>, <4 x double> } %result, 1
|
|
store <4 x double> %result.0, ptr %out_sin, align 32
|
|
store <4 x double> %result.1, ptr %out_cos, align 32
|
|
ret void
|
|
}
|
|
|
|
define void @test_sincos_v8f64(<8 x double> %x, ptr noalias %out_sin, ptr noalias %out_cos) {
|
|
; AMD-SSE-LABEL: test_sincos_v8f64:
|
|
; AMD-SSE: callq amd_vrd2_sincos@PLT
|
|
; AMD-SSE: callq amd_vrd2_sincos@PLT
|
|
; AMD-SSE: callq amd_vrd2_sincos@PLT
|
|
; AMD-SSE: callq amd_vrd2_sincos@PLT
|
|
;
|
|
; AMD-AVX-LABEL: test_sincos_v8f64:
|
|
; AMD-AVX: callq amd_vrd2_sincos@PLT
|
|
; AMD-AVX: callq amd_vrd2_sincos@PLT
|
|
; AMD-AVX: callq amd_vrd2_sincos@PLT
|
|
; AMD-AVX: callq amd_vrd2_sincos@PLT
|
|
;
|
|
; AMD-AVX2-LABEL: test_sincos_v8f64:
|
|
; AMD-AVX2: callq amd_vrd4_sincos@PLT
|
|
; AMD-AVX2: callq amd_vrd4_sincos@PLT
|
|
;
|
|
; AMD-AVX512-LABEL: test_sincos_v8f64:
|
|
; AMD-AVX512: callq amd_vrd8_sincos@PLT
|
|
;
|
|
; GLIBC-LABEL: test_sincos_v8f64:
|
|
; GLIBC: callq sincos@PLT
|
|
; GLIBC: callq sincos@PLT
|
|
; GLIBC: callq sincos@PLT
|
|
; GLIBC: callq sincos@PLT
|
|
; GLIBC: callq sincos@PLT
|
|
; GLIBC: callq sincos@PLT
|
|
; GLIBC: callq sincos@PLT
|
|
; GLIBC: callq sincos@PLT
|
|
%result = call { <8 x double>, <8 x double> } @llvm.sincos.v8f64(<8 x double> %x)
|
|
%result.0 = extractvalue { <8 x double>, <8 x double> } %result, 0
|
|
%result.1 = extractvalue { <8 x double>, <8 x double> } %result, 1
|
|
store <8 x double> %result.0, ptr %out_sin, align 64
|
|
store <8 x double> %result.1, ptr %out_cos, align 64
|
|
ret void
|
|
}
|
|
;; NOTE: These prefixes are unused and the list is autogenerated. Do not add tests below this line:
|
|
; CHECK: {{.*}}
|
|
; GLIBC-AVX: {{.*}}
|
|
; GLIBC-AVX2: {{.*}}
|
|
; GLIBC-AVX512: {{.*}}
|
|
; GLIBC-SSE: {{.*}}
|