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
177 lines
9.2 KiB
LLVM
177 lines
9.2 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 4
|
|
; RUN: opt -passes=pre-isel-intrinsic-lowering -S < %s | FileCheck %s
|
|
|
|
define void @vp_fadd_v4f32(<4 x float> %a0, <4 x float> %a1, ptr %out, i32 %vp) nounwind {
|
|
; CHECK-LABEL: define void @vp_fadd_v4f32(
|
|
; CHECK-SAME: <4 x float> [[A0:%.*]], <4 x float> [[A1:%.*]], ptr [[OUT:%.*]], i32 [[VP:%.*]]) #[[ATTR0:[0-9]+]] {
|
|
; CHECK-NEXT: [[RES1:%.*]] = fadd <4 x float> [[A0]], [[A1]]
|
|
; CHECK-NEXT: store <4 x float> [[RES1]], ptr [[OUT]], align 16
|
|
; CHECK-NEXT: ret void
|
|
;
|
|
%res = call <4 x float> @llvm.vp.fadd.v4f32(<4 x float> %a0, <4 x float> %a1, <4 x i1> <i1 -1, i1 -1, i1 -1, i1 -1>, i32 %vp)
|
|
store <4 x float> %res, ptr %out
|
|
ret void
|
|
}
|
|
declare <4 x float> @llvm.vp.fadd.v4f32(<4 x float>, <4 x float>, <4 x i1>, i32)
|
|
|
|
define void @vp_fsub_v4f32(<4 x float> %a0, <4 x float> %a1, ptr %out, i32 %vp) nounwind {
|
|
; CHECK-LABEL: define void @vp_fsub_v4f32(
|
|
; CHECK-SAME: <4 x float> [[A0:%.*]], <4 x float> [[A1:%.*]], ptr [[OUT:%.*]], i32 [[VP:%.*]]) #[[ATTR0]] {
|
|
; CHECK-NEXT: [[RES1:%.*]] = fsub <4 x float> [[A0]], [[A1]]
|
|
; CHECK-NEXT: store <4 x float> [[RES1]], ptr [[OUT]], align 16
|
|
; CHECK-NEXT: ret void
|
|
;
|
|
%res = call <4 x float> @llvm.vp.fsub.v4f32(<4 x float> %a0, <4 x float> %a1, <4 x i1> <i1 -1, i1 -1, i1 -1, i1 -1>, i32 %vp)
|
|
store <4 x float> %res, ptr %out
|
|
ret void
|
|
}
|
|
declare <4 x float> @llvm.vp.fsub.v4f32(<4 x float>, <4 x float>, <4 x i1>, i32)
|
|
|
|
define void @vp_fmul_v4f32(<4 x float> %a0, <4 x float> %a1, ptr %out, i32 %vp) nounwind {
|
|
; CHECK-LABEL: define void @vp_fmul_v4f32(
|
|
; CHECK-SAME: <4 x float> [[A0:%.*]], <4 x float> [[A1:%.*]], ptr [[OUT:%.*]], i32 [[VP:%.*]]) #[[ATTR0]] {
|
|
; CHECK-NEXT: [[RES1:%.*]] = fmul <4 x float> [[A0]], [[A1]]
|
|
; CHECK-NEXT: store <4 x float> [[RES1]], ptr [[OUT]], align 16
|
|
; CHECK-NEXT: ret void
|
|
;
|
|
%res = call <4 x float> @llvm.vp.fmul.v4f32(<4 x float> %a0, <4 x float> %a1, <4 x i1> <i1 -1, i1 -1, i1 -1, i1 -1>, i32 %vp)
|
|
store <4 x float> %res, ptr %out
|
|
ret void
|
|
}
|
|
declare <4 x float> @llvm.vp.fmul.v4f32(<4 x float>, <4 x float>, <4 x i1>, i32)
|
|
|
|
define void @vp_fdiv_v4f32(<4 x float> %a0, <4 x float> %a1, ptr %out, i32 %vp) nounwind {
|
|
; CHECK-LABEL: define void @vp_fdiv_v4f32(
|
|
; CHECK-SAME: <4 x float> [[A0:%.*]], <4 x float> [[A1:%.*]], ptr [[OUT:%.*]], i32 [[VP:%.*]]) #[[ATTR0]] {
|
|
; CHECK-NEXT: [[RES1:%.*]] = fdiv <4 x float> [[A0]], [[A1]]
|
|
; CHECK-NEXT: store <4 x float> [[RES1]], ptr [[OUT]], align 16
|
|
; CHECK-NEXT: ret void
|
|
;
|
|
%res = call <4 x float> @llvm.vp.fdiv.v4f32(<4 x float> %a0, <4 x float> %a1, <4 x i1> <i1 -1, i1 -1, i1 -1, i1 -1>, i32 %vp)
|
|
store <4 x float> %res, ptr %out
|
|
ret void
|
|
}
|
|
declare <4 x float> @llvm.vp.fdiv.v4f32(<4 x float>, <4 x float>, <4 x i1>, i32)
|
|
|
|
define void @vp_frem_v4f32(<4 x float> %a0, <4 x float> %a1, ptr %out, i32 %vp) nounwind {
|
|
; CHECK-LABEL: define void @vp_frem_v4f32(
|
|
; CHECK-SAME: <4 x float> [[A0:%.*]], <4 x float> [[A1:%.*]], ptr [[OUT:%.*]], i32 [[VP:%.*]]) #[[ATTR0]] {
|
|
; CHECK-NEXT: [[RES1:%.*]] = frem <4 x float> [[A0]], [[A1]]
|
|
; CHECK-NEXT: store <4 x float> [[RES1]], ptr [[OUT]], align 16
|
|
; CHECK-NEXT: ret void
|
|
;
|
|
%res = call <4 x float> @llvm.vp.frem.v4f32(<4 x float> %a0, <4 x float> %a1, <4 x i1> <i1 -1, i1 -1, i1 -1, i1 -1>, i32 %vp)
|
|
store <4 x float> %res, ptr %out
|
|
ret void
|
|
}
|
|
declare <4 x float> @llvm.vp.frem.v4f32(<4 x float>, <4 x float>, <4 x i1>, i32)
|
|
|
|
define void @vp_fabs_v4f32(<4 x float> %a0, <4 x float> %a1, ptr %out, i32 %vp) nounwind {
|
|
; CHECK-LABEL: define void @vp_fabs_v4f32(
|
|
; CHECK-SAME: <4 x float> [[A0:%.*]], <4 x float> [[A1:%.*]], ptr [[OUT:%.*]], i32 [[VP:%.*]]) #[[ATTR0]] {
|
|
; CHECK-NEXT: [[RES1:%.*]] = call <4 x float> @llvm.fabs.v4f32(<4 x float> [[A0]])
|
|
; CHECK-NEXT: store <4 x float> [[RES1]], ptr [[OUT]], align 16
|
|
; CHECK-NEXT: ret void
|
|
;
|
|
%res = call <4 x float> @llvm.vp.fabs.v4f32(<4 x float> %a0, <4 x i1> <i1 -1, i1 -1, i1 -1, i1 -1>, i32 %vp)
|
|
store <4 x float> %res, ptr %out
|
|
ret void
|
|
}
|
|
declare <4 x float> @llvm.vp.fabs.v4f32(<4 x float>, <4 x i1>, i32)
|
|
|
|
define void @vp_sqrt_v4f32(<4 x float> %a0, <4 x float> %a1, ptr %out, i32 %vp) nounwind {
|
|
; CHECK-LABEL: define void @vp_sqrt_v4f32(
|
|
; CHECK-SAME: <4 x float> [[A0:%.*]], <4 x float> [[A1:%.*]], ptr [[OUT:%.*]], i32 [[VP:%.*]]) #[[ATTR0]] {
|
|
; CHECK-NEXT: [[RES1:%.*]] = call <4 x float> @llvm.sqrt.v4f32(<4 x float> [[A0]])
|
|
; CHECK-NEXT: store <4 x float> [[RES1]], ptr [[OUT]], align 16
|
|
; CHECK-NEXT: ret void
|
|
;
|
|
%res = call <4 x float> @llvm.vp.sqrt.v4f32(<4 x float> %a0, <4 x i1> <i1 -1, i1 -1, i1 -1, i1 -1>, i32 %vp)
|
|
store <4 x float> %res, ptr %out
|
|
ret void
|
|
}
|
|
declare <4 x float> @llvm.vp.sqrt.v4f32(<4 x float>, <4 x i1>, i32)
|
|
|
|
define void @vp_fneg_v4f32(<4 x float> %a0, <4 x float> %a1, ptr %out, i32 %vp) nounwind {
|
|
; CHECK-LABEL: define void @vp_fneg_v4f32(
|
|
; CHECK-SAME: <4 x float> [[A0:%.*]], <4 x float> [[A1:%.*]], ptr [[OUT:%.*]], i32 [[VP:%.*]]) #[[ATTR0]] {
|
|
; CHECK-NEXT: [[RES1:%.*]] = fneg <4 x float> [[A0]]
|
|
; CHECK-NEXT: store <4 x float> [[RES1]], ptr [[OUT]], align 16
|
|
; CHECK-NEXT: ret void
|
|
;
|
|
%res = call <4 x float> @llvm.vp.fneg.v4f32(<4 x float> %a0, <4 x i1> <i1 -1, i1 -1, i1 -1, i1 -1>, i32 %vp)
|
|
store <4 x float> %res, ptr %out
|
|
ret void
|
|
}
|
|
declare <4 x float> @llvm.vp.fneg.v4f32(<4 x float>, <4 x i1>, i32)
|
|
|
|
define void @vp_fma_v4f32(<4 x float> %a0, <4 x float> %a1, ptr %out, i4 %a5) nounwind {
|
|
; CHECK-LABEL: define void @vp_fma_v4f32(
|
|
; CHECK-SAME: <4 x float> [[A0:%.*]], <4 x float> [[A1:%.*]], ptr [[OUT:%.*]], i4 [[A5:%.*]]) #[[ATTR0]] {
|
|
; CHECK-NEXT: [[RES1:%.*]] = call <4 x float> @llvm.fma.v4f32(<4 x float> [[A0]], <4 x float> [[A1]], <4 x float> [[A1]])
|
|
; CHECK-NEXT: store <4 x float> [[RES1]], ptr [[OUT]], align 16
|
|
; CHECK-NEXT: ret void
|
|
;
|
|
%res = call <4 x float> @llvm.vp.fma.v4f32(<4 x float> %a0, <4 x float> %a1, <4 x float> %a1, <4 x i1> <i1 -1, i1 -1, i1 -1, i1 -1>, i32 4)
|
|
store <4 x float> %res, ptr %out
|
|
ret void
|
|
}
|
|
declare <4 x float> @llvm.vp.fma.v4f32(<4 x float>, <4 x float>, <4 x float>, <4 x i1>, i32)
|
|
|
|
define void @vp_fmuladd_v4f32(<4 x float> %a0, <4 x float> %a1, ptr %out, i4 %a5) nounwind {
|
|
; CHECK-LABEL: define void @vp_fmuladd_v4f32(
|
|
; CHECK-SAME: <4 x float> [[A0:%.*]], <4 x float> [[A1:%.*]], ptr [[OUT:%.*]], i4 [[A5:%.*]]) #[[ATTR0]] {
|
|
; CHECK-NEXT: [[RES1:%.*]] = call <4 x float> @llvm.fmuladd.v4f32(<4 x float> [[A0]], <4 x float> [[A1]], <4 x float> [[A1]])
|
|
; CHECK-NEXT: store <4 x float> [[RES1]], ptr [[OUT]], align 16
|
|
; CHECK-NEXT: ret void
|
|
;
|
|
%res = call <4 x float> @llvm.vp.fmuladd.v4f32(<4 x float> %a0, <4 x float> %a1, <4 x float> %a1, <4 x i1> <i1 -1, i1 -1, i1 -1, i1 -1>, i32 4)
|
|
store <4 x float> %res, ptr %out
|
|
ret void
|
|
}
|
|
declare <4 x float> @llvm.vp.fmuladd.v4f32(<4 x float>, <4 x float>, <4 x float>, <4 x i1>, i32)
|
|
|
|
declare <4 x float> @llvm.vp.maxnum.v4f32(<4 x float>, <4 x float>, <4 x i1>, i32)
|
|
define <4 x float> @vfmax_vv_v4f32(<4 x float> %va, <4 x float> %vb, <4 x i1> %m, i32 zeroext %evl) {
|
|
; CHECK-LABEL: define <4 x float> @vfmax_vv_v4f32(
|
|
; CHECK-SAME: <4 x float> [[VA:%.*]], <4 x float> [[VB:%.*]], <4 x i1> [[M:%.*]], i32 zeroext [[EVL:%.*]]) {
|
|
; CHECK-NEXT: [[V1:%.*]] = call <4 x float> @llvm.maxnum.v4f32(<4 x float> [[VA]], <4 x float> [[VB]])
|
|
; CHECK-NEXT: ret <4 x float> [[V1]]
|
|
;
|
|
%v = call <4 x float> @llvm.vp.maxnum.v4f32(<4 x float> %va, <4 x float> %vb, <4 x i1> %m, i32 %evl)
|
|
ret <4 x float> %v
|
|
}
|
|
|
|
declare <8 x float> @llvm.vp.maxnum.v8f32(<8 x float>, <8 x float>, <8 x i1>, i32)
|
|
define <8 x float> @vfmax_vv_v8f32(<8 x float> %va, <8 x float> %vb, <8 x i1> %m, i32 zeroext %evl) {
|
|
; CHECK-LABEL: define <8 x float> @vfmax_vv_v8f32(
|
|
; CHECK-SAME: <8 x float> [[VA:%.*]], <8 x float> [[VB:%.*]], <8 x i1> [[M:%.*]], i32 zeroext [[EVL:%.*]]) {
|
|
; CHECK-NEXT: [[V1:%.*]] = call <8 x float> @llvm.maxnum.v8f32(<8 x float> [[VA]], <8 x float> [[VB]])
|
|
; CHECK-NEXT: ret <8 x float> [[V1]]
|
|
;
|
|
%v = call <8 x float> @llvm.vp.maxnum.v8f32(<8 x float> %va, <8 x float> %vb, <8 x i1> %m, i32 %evl)
|
|
ret <8 x float> %v
|
|
}
|
|
|
|
declare <4 x float> @llvm.vp.minnum.v4f32(<4 x float>, <4 x float>, <4 x i1>, i32)
|
|
define <4 x float> @vfmin_vv_v4f32(<4 x float> %va, <4 x float> %vb, <4 x i1> %m, i32 zeroext %evl) {
|
|
; CHECK-LABEL: define <4 x float> @vfmin_vv_v4f32(
|
|
; CHECK-SAME: <4 x float> [[VA:%.*]], <4 x float> [[VB:%.*]], <4 x i1> [[M:%.*]], i32 zeroext [[EVL:%.*]]) {
|
|
; CHECK-NEXT: [[V1:%.*]] = call <4 x float> @llvm.minnum.v4f32(<4 x float> [[VA]], <4 x float> [[VB]])
|
|
; CHECK-NEXT: ret <4 x float> [[V1]]
|
|
;
|
|
%v = call <4 x float> @llvm.vp.minnum.v4f32(<4 x float> %va, <4 x float> %vb, <4 x i1> %m, i32 %evl)
|
|
ret <4 x float> %v
|
|
}
|
|
|
|
declare <8 x float> @llvm.vp.minnum.v8f32(<8 x float>, <8 x float>, <8 x i1>, i32)
|
|
define <8 x float> @vfmin_vv_v8f32(<8 x float> %va, <8 x float> %vb, <8 x i1> %m, i32 zeroext %evl) {
|
|
; CHECK-LABEL: define <8 x float> @vfmin_vv_v8f32(
|
|
; CHECK-SAME: <8 x float> [[VA:%.*]], <8 x float> [[VB:%.*]], <8 x i1> [[M:%.*]], i32 zeroext [[EVL:%.*]]) {
|
|
; CHECK-NEXT: [[V1:%.*]] = call <8 x float> @llvm.minnum.v8f32(<8 x float> [[VA]], <8 x float> [[VB]])
|
|
; CHECK-NEXT: ret <8 x float> [[V1]]
|
|
;
|
|
%v = call <8 x float> @llvm.vp.minnum.v8f32(<8 x float> %va, <8 x float> %vb, <8 x i1> %m, i32 %evl)
|
|
ret <8 x float> %v
|
|
}
|