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
126 lines
4.2 KiB
LLVM
126 lines
4.2 KiB
LLVM
; RUN: llc -mtriple=arm-eabi -mattr=+neon %s -o - | FileCheck %s
|
|
|
|
define <8 x i8> @vhsubs8(ptr %A, ptr %B) nounwind {
|
|
;CHECK-LABEL: vhsubs8:
|
|
;CHECK: vhsub.s8
|
|
%tmp1 = load <8 x i8>, ptr %A
|
|
%tmp2 = load <8 x i8>, ptr %B
|
|
%tmp3 = call <8 x i8> @llvm.arm.neon.vhsubs.v8i8(<8 x i8> %tmp1, <8 x i8> %tmp2)
|
|
ret <8 x i8> %tmp3
|
|
}
|
|
|
|
define <4 x i16> @vhsubs16(ptr %A, ptr %B) nounwind {
|
|
;CHECK-LABEL: vhsubs16:
|
|
;CHECK: vhsub.s16
|
|
%tmp1 = load <4 x i16>, ptr %A
|
|
%tmp2 = load <4 x i16>, ptr %B
|
|
%tmp3 = call <4 x i16> @llvm.arm.neon.vhsubs.v4i16(<4 x i16> %tmp1, <4 x i16> %tmp2)
|
|
ret <4 x i16> %tmp3
|
|
}
|
|
|
|
define <2 x i32> @vhsubs32(ptr %A, ptr %B) nounwind {
|
|
;CHECK-LABEL: vhsubs32:
|
|
;CHECK: vhsub.s32
|
|
%tmp1 = load <2 x i32>, ptr %A
|
|
%tmp2 = load <2 x i32>, ptr %B
|
|
%tmp3 = call <2 x i32> @llvm.arm.neon.vhsubs.v2i32(<2 x i32> %tmp1, <2 x i32> %tmp2)
|
|
ret <2 x i32> %tmp3
|
|
}
|
|
|
|
define <8 x i8> @vhsubu8(ptr %A, ptr %B) nounwind {
|
|
;CHECK-LABEL: vhsubu8:
|
|
;CHECK: vhsub.u8
|
|
%tmp1 = load <8 x i8>, ptr %A
|
|
%tmp2 = load <8 x i8>, ptr %B
|
|
%tmp3 = call <8 x i8> @llvm.arm.neon.vhsubu.v8i8(<8 x i8> %tmp1, <8 x i8> %tmp2)
|
|
ret <8 x i8> %tmp3
|
|
}
|
|
|
|
define <4 x i16> @vhsubu16(ptr %A, ptr %B) nounwind {
|
|
;CHECK-LABEL: vhsubu16:
|
|
;CHECK: vhsub.u16
|
|
%tmp1 = load <4 x i16>, ptr %A
|
|
%tmp2 = load <4 x i16>, ptr %B
|
|
%tmp3 = call <4 x i16> @llvm.arm.neon.vhsubu.v4i16(<4 x i16> %tmp1, <4 x i16> %tmp2)
|
|
ret <4 x i16> %tmp3
|
|
}
|
|
|
|
define <2 x i32> @vhsubu32(ptr %A, ptr %B) nounwind {
|
|
;CHECK-LABEL: vhsubu32:
|
|
;CHECK: vhsub.u32
|
|
%tmp1 = load <2 x i32>, ptr %A
|
|
%tmp2 = load <2 x i32>, ptr %B
|
|
%tmp3 = call <2 x i32> @llvm.arm.neon.vhsubu.v2i32(<2 x i32> %tmp1, <2 x i32> %tmp2)
|
|
ret <2 x i32> %tmp3
|
|
}
|
|
|
|
define <16 x i8> @vhsubQs8(ptr %A, ptr %B) nounwind {
|
|
;CHECK-LABEL: vhsubQs8:
|
|
;CHECK: vhsub.s8
|
|
%tmp1 = load <16 x i8>, ptr %A
|
|
%tmp2 = load <16 x i8>, ptr %B
|
|
%tmp3 = call <16 x i8> @llvm.arm.neon.vhsubs.v16i8(<16 x i8> %tmp1, <16 x i8> %tmp2)
|
|
ret <16 x i8> %tmp3
|
|
}
|
|
|
|
define <8 x i16> @vhsubQs16(ptr %A, ptr %B) nounwind {
|
|
;CHECK-LABEL: vhsubQs16:
|
|
;CHECK: vhsub.s16
|
|
%tmp1 = load <8 x i16>, ptr %A
|
|
%tmp2 = load <8 x i16>, ptr %B
|
|
%tmp3 = call <8 x i16> @llvm.arm.neon.vhsubs.v8i16(<8 x i16> %tmp1, <8 x i16> %tmp2)
|
|
ret <8 x i16> %tmp3
|
|
}
|
|
|
|
define <4 x i32> @vhsubQs32(ptr %A, ptr %B) nounwind {
|
|
;CHECK-LABEL: vhsubQs32:
|
|
;CHECK: vhsub.s32
|
|
%tmp1 = load <4 x i32>, ptr %A
|
|
%tmp2 = load <4 x i32>, ptr %B
|
|
%tmp3 = call <4 x i32> @llvm.arm.neon.vhsubs.v4i32(<4 x i32> %tmp1, <4 x i32> %tmp2)
|
|
ret <4 x i32> %tmp3
|
|
}
|
|
|
|
define <16 x i8> @vhsubQu8(ptr %A, ptr %B) nounwind {
|
|
;CHECK-LABEL: vhsubQu8:
|
|
;CHECK: vhsub.u8
|
|
%tmp1 = load <16 x i8>, ptr %A
|
|
%tmp2 = load <16 x i8>, ptr %B
|
|
%tmp3 = call <16 x i8> @llvm.arm.neon.vhsubu.v16i8(<16 x i8> %tmp1, <16 x i8> %tmp2)
|
|
ret <16 x i8> %tmp3
|
|
}
|
|
|
|
define <8 x i16> @vhsubQu16(ptr %A, ptr %B) nounwind {
|
|
;CHECK-LABEL: vhsubQu16:
|
|
;CHECK: vhsub.u16
|
|
%tmp1 = load <8 x i16>, ptr %A
|
|
%tmp2 = load <8 x i16>, ptr %B
|
|
%tmp3 = call <8 x i16> @llvm.arm.neon.vhsubu.v8i16(<8 x i16> %tmp1, <8 x i16> %tmp2)
|
|
ret <8 x i16> %tmp3
|
|
}
|
|
|
|
define <4 x i32> @vhsubQu32(ptr %A, ptr %B) nounwind {
|
|
;CHECK-LABEL: vhsubQu32:
|
|
;CHECK: vhsub.u32
|
|
%tmp1 = load <4 x i32>, ptr %A
|
|
%tmp2 = load <4 x i32>, ptr %B
|
|
%tmp3 = call <4 x i32> @llvm.arm.neon.vhsubu.v4i32(<4 x i32> %tmp1, <4 x i32> %tmp2)
|
|
ret <4 x i32> %tmp3
|
|
}
|
|
|
|
declare <8 x i8> @llvm.arm.neon.vhsubs.v8i8(<8 x i8>, <8 x i8>) nounwind readnone
|
|
declare <4 x i16> @llvm.arm.neon.vhsubs.v4i16(<4 x i16>, <4 x i16>) nounwind readnone
|
|
declare <2 x i32> @llvm.arm.neon.vhsubs.v2i32(<2 x i32>, <2 x i32>) nounwind readnone
|
|
|
|
declare <8 x i8> @llvm.arm.neon.vhsubu.v8i8(<8 x i8>, <8 x i8>) nounwind readnone
|
|
declare <4 x i16> @llvm.arm.neon.vhsubu.v4i16(<4 x i16>, <4 x i16>) nounwind readnone
|
|
declare <2 x i32> @llvm.arm.neon.vhsubu.v2i32(<2 x i32>, <2 x i32>) nounwind readnone
|
|
|
|
declare <16 x i8> @llvm.arm.neon.vhsubs.v16i8(<16 x i8>, <16 x i8>) nounwind readnone
|
|
declare <8 x i16> @llvm.arm.neon.vhsubs.v8i16(<8 x i16>, <8 x i16>) nounwind readnone
|
|
declare <4 x i32> @llvm.arm.neon.vhsubs.v4i32(<4 x i32>, <4 x i32>) nounwind readnone
|
|
|
|
declare <16 x i8> @llvm.arm.neon.vhsubu.v16i8(<16 x i8>, <16 x i8>) nounwind readnone
|
|
declare <8 x i16> @llvm.arm.neon.vhsubu.v8i16(<8 x i16>, <8 x i16>) nounwind readnone
|
|
declare <4 x i32> @llvm.arm.neon.vhsubu.v4i32(<4 x i32>, <4 x i32>) nounwind readnone
|