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
148 lines
4.9 KiB
LLVM
148 lines
4.9 KiB
LLVM
; RUN: llc -mtriple=arm-eabi -mattr=+neon %s -o - | FileCheck %s
|
|
|
|
define <8 x i8> @vpmins8(ptr %A, ptr %B) nounwind {
|
|
;CHECK-LABEL: vpmins8:
|
|
;CHECK: vpmin.s8
|
|
%tmp1 = load <8 x i8>, ptr %A
|
|
%tmp2 = load <8 x i8>, ptr %B
|
|
%tmp3 = call <8 x i8> @llvm.arm.neon.vpmins.v8i8(<8 x i8> %tmp1, <8 x i8> %tmp2)
|
|
ret <8 x i8> %tmp3
|
|
}
|
|
|
|
define <4 x i16> @vpmins16(ptr %A, ptr %B) nounwind {
|
|
;CHECK-LABEL: vpmins16:
|
|
;CHECK: vpmin.s16
|
|
%tmp1 = load <4 x i16>, ptr %A
|
|
%tmp2 = load <4 x i16>, ptr %B
|
|
%tmp3 = call <4 x i16> @llvm.arm.neon.vpmins.v4i16(<4 x i16> %tmp1, <4 x i16> %tmp2)
|
|
ret <4 x i16> %tmp3
|
|
}
|
|
|
|
define <2 x i32> @vpmins32(ptr %A, ptr %B) nounwind {
|
|
;CHECK-LABEL: vpmins32:
|
|
;CHECK: vpmin.s32
|
|
%tmp1 = load <2 x i32>, ptr %A
|
|
%tmp2 = load <2 x i32>, ptr %B
|
|
%tmp3 = call <2 x i32> @llvm.arm.neon.vpmins.v2i32(<2 x i32> %tmp1, <2 x i32> %tmp2)
|
|
ret <2 x i32> %tmp3
|
|
}
|
|
|
|
define <8 x i8> @vpminu8(ptr %A, ptr %B) nounwind {
|
|
;CHECK-LABEL: vpminu8:
|
|
;CHECK: vpmin.u8
|
|
%tmp1 = load <8 x i8>, ptr %A
|
|
%tmp2 = load <8 x i8>, ptr %B
|
|
%tmp3 = call <8 x i8> @llvm.arm.neon.vpminu.v8i8(<8 x i8> %tmp1, <8 x i8> %tmp2)
|
|
ret <8 x i8> %tmp3
|
|
}
|
|
|
|
define <4 x i16> @vpminu16(ptr %A, ptr %B) nounwind {
|
|
;CHECK-LABEL: vpminu16:
|
|
;CHECK: vpmin.u16
|
|
%tmp1 = load <4 x i16>, ptr %A
|
|
%tmp2 = load <4 x i16>, ptr %B
|
|
%tmp3 = call <4 x i16> @llvm.arm.neon.vpminu.v4i16(<4 x i16> %tmp1, <4 x i16> %tmp2)
|
|
ret <4 x i16> %tmp3
|
|
}
|
|
|
|
define <2 x i32> @vpminu32(ptr %A, ptr %B) nounwind {
|
|
;CHECK-LABEL: vpminu32:
|
|
;CHECK: vpmin.u32
|
|
%tmp1 = load <2 x i32>, ptr %A
|
|
%tmp2 = load <2 x i32>, ptr %B
|
|
%tmp3 = call <2 x i32> @llvm.arm.neon.vpminu.v2i32(<2 x i32> %tmp1, <2 x i32> %tmp2)
|
|
ret <2 x i32> %tmp3
|
|
}
|
|
|
|
define <2 x float> @vpminf32(ptr %A, ptr %B) nounwind {
|
|
;CHECK-LABEL: vpminf32:
|
|
;CHECK: vpmin.f32
|
|
%tmp1 = load <2 x float>, ptr %A
|
|
%tmp2 = load <2 x float>, ptr %B
|
|
%tmp3 = call <2 x float> @llvm.arm.neon.vpmins.v2f32(<2 x float> %tmp1, <2 x float> %tmp2)
|
|
ret <2 x float> %tmp3
|
|
}
|
|
|
|
declare <8 x i8> @llvm.arm.neon.vpmins.v8i8(<8 x i8>, <8 x i8>) nounwind readnone
|
|
declare <4 x i16> @llvm.arm.neon.vpmins.v4i16(<4 x i16>, <4 x i16>) nounwind readnone
|
|
declare <2 x i32> @llvm.arm.neon.vpmins.v2i32(<2 x i32>, <2 x i32>) nounwind readnone
|
|
|
|
declare <8 x i8> @llvm.arm.neon.vpminu.v8i8(<8 x i8>, <8 x i8>) nounwind readnone
|
|
declare <4 x i16> @llvm.arm.neon.vpminu.v4i16(<4 x i16>, <4 x i16>) nounwind readnone
|
|
declare <2 x i32> @llvm.arm.neon.vpminu.v2i32(<2 x i32>, <2 x i32>) nounwind readnone
|
|
|
|
declare <2 x float> @llvm.arm.neon.vpmins.v2f32(<2 x float>, <2 x float>) nounwind readnone
|
|
|
|
define <8 x i8> @vpmaxs8(ptr %A, ptr %B) nounwind {
|
|
;CHECK-LABEL: vpmaxs8:
|
|
;CHECK: vpmax.s8
|
|
%tmp1 = load <8 x i8>, ptr %A
|
|
%tmp2 = load <8 x i8>, ptr %B
|
|
%tmp3 = call <8 x i8> @llvm.arm.neon.vpmaxs.v8i8(<8 x i8> %tmp1, <8 x i8> %tmp2)
|
|
ret <8 x i8> %tmp3
|
|
}
|
|
|
|
define <4 x i16> @vpmaxs16(ptr %A, ptr %B) nounwind {
|
|
;CHECK-LABEL: vpmaxs16:
|
|
;CHECK: vpmax.s16
|
|
%tmp1 = load <4 x i16>, ptr %A
|
|
%tmp2 = load <4 x i16>, ptr %B
|
|
%tmp3 = call <4 x i16> @llvm.arm.neon.vpmaxs.v4i16(<4 x i16> %tmp1, <4 x i16> %tmp2)
|
|
ret <4 x i16> %tmp3
|
|
}
|
|
|
|
define <2 x i32> @vpmaxs32(ptr %A, ptr %B) nounwind {
|
|
;CHECK-LABEL: vpmaxs32:
|
|
;CHECK: vpmax.s32
|
|
%tmp1 = load <2 x i32>, ptr %A
|
|
%tmp2 = load <2 x i32>, ptr %B
|
|
%tmp3 = call <2 x i32> @llvm.arm.neon.vpmaxs.v2i32(<2 x i32> %tmp1, <2 x i32> %tmp2)
|
|
ret <2 x i32> %tmp3
|
|
}
|
|
|
|
define <8 x i8> @vpmaxu8(ptr %A, ptr %B) nounwind {
|
|
;CHECK-LABEL: vpmaxu8:
|
|
;CHECK: vpmax.u8
|
|
%tmp1 = load <8 x i8>, ptr %A
|
|
%tmp2 = load <8 x i8>, ptr %B
|
|
%tmp3 = call <8 x i8> @llvm.arm.neon.vpmaxu.v8i8(<8 x i8> %tmp1, <8 x i8> %tmp2)
|
|
ret <8 x i8> %tmp3
|
|
}
|
|
|
|
define <4 x i16> @vpmaxu16(ptr %A, ptr %B) nounwind {
|
|
;CHECK-LABEL: vpmaxu16:
|
|
;CHECK: vpmax.u16
|
|
%tmp1 = load <4 x i16>, ptr %A
|
|
%tmp2 = load <4 x i16>, ptr %B
|
|
%tmp3 = call <4 x i16> @llvm.arm.neon.vpmaxu.v4i16(<4 x i16> %tmp1, <4 x i16> %tmp2)
|
|
ret <4 x i16> %tmp3
|
|
}
|
|
|
|
define <2 x i32> @vpmaxu32(ptr %A, ptr %B) nounwind {
|
|
;CHECK-LABEL: vpmaxu32:
|
|
;CHECK: vpmax.u32
|
|
%tmp1 = load <2 x i32>, ptr %A
|
|
%tmp2 = load <2 x i32>, ptr %B
|
|
%tmp3 = call <2 x i32> @llvm.arm.neon.vpmaxu.v2i32(<2 x i32> %tmp1, <2 x i32> %tmp2)
|
|
ret <2 x i32> %tmp3
|
|
}
|
|
|
|
define <2 x float> @vpmaxf32(ptr %A, ptr %B) nounwind {
|
|
;CHECK-LABEL: vpmaxf32:
|
|
;CHECK: vpmax.f32
|
|
%tmp1 = load <2 x float>, ptr %A
|
|
%tmp2 = load <2 x float>, ptr %B
|
|
%tmp3 = call <2 x float> @llvm.arm.neon.vpmaxs.v2f32(<2 x float> %tmp1, <2 x float> %tmp2)
|
|
ret <2 x float> %tmp3
|
|
}
|
|
|
|
declare <8 x i8> @llvm.arm.neon.vpmaxs.v8i8(<8 x i8>, <8 x i8>) nounwind readnone
|
|
declare <4 x i16> @llvm.arm.neon.vpmaxs.v4i16(<4 x i16>, <4 x i16>) nounwind readnone
|
|
declare <2 x i32> @llvm.arm.neon.vpmaxs.v2i32(<2 x i32>, <2 x i32>) nounwind readnone
|
|
|
|
declare <8 x i8> @llvm.arm.neon.vpmaxu.v8i8(<8 x i8>, <8 x i8>) nounwind readnone
|
|
declare <4 x i16> @llvm.arm.neon.vpmaxu.v4i16(<4 x i16>, <4 x i16>) nounwind readnone
|
|
declare <2 x i32> @llvm.arm.neon.vpmaxu.v2i32(<2 x i32>, <2 x i32>) nounwind readnone
|
|
|
|
declare <2 x float> @llvm.arm.neon.vpmaxs.v2f32(<2 x float>, <2 x float>) nounwind readnone
|