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
117 lines
3.2 KiB
LLVM
117 lines
3.2 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
|
|
; RUN: llc -mtriple=x86_64-linux-gnu -mcpu=skx -global-isel -verify-machineinstrs < %s -o - | FileCheck %s --check-prefix=SKX
|
|
; RUN: llc -mtriple=x86_64-linux-gnu -mcpu=skx -regbankselect-greedy -global-isel -verify-machineinstrs < %s -o - | FileCheck %s --check-prefix=SKX
|
|
|
|
define <4 x i32> @test_load_v4i32_noalign(ptr %p1) {
|
|
; SKX-LABEL: test_load_v4i32_noalign:
|
|
; SKX: # %bb.0:
|
|
; SKX-NEXT: vmovups (%rdi), %xmm0
|
|
; SKX-NEXT: retq
|
|
%r = load <4 x i32>, ptr %p1, align 1
|
|
ret <4 x i32> %r
|
|
}
|
|
|
|
define <4 x i32> @test_load_v4i32_align(ptr %p1) {
|
|
; SKX-LABEL: test_load_v4i32_align:
|
|
; SKX: # %bb.0:
|
|
; SKX-NEXT: vmovaps (%rdi), %xmm0
|
|
; SKX-NEXT: retq
|
|
%r = load <4 x i32>, ptr %p1, align 16
|
|
ret <4 x i32> %r
|
|
}
|
|
|
|
define <8 x i32> @test_load_v8i32_noalign(ptr %p1) {
|
|
; SKX-LABEL: test_load_v8i32_noalign:
|
|
; SKX: # %bb.0:
|
|
; SKX-NEXT: vmovups (%rdi), %ymm0
|
|
; SKX-NEXT: retq
|
|
%r = load <8 x i32>, ptr %p1, align 1
|
|
ret <8 x i32> %r
|
|
}
|
|
|
|
define <8 x i32> @test_load_v8i32_align(ptr %p1) {
|
|
; SKX-LABEL: test_load_v8i32_align:
|
|
; SKX: # %bb.0:
|
|
; SKX-NEXT: vmovaps (%rdi), %ymm0
|
|
; SKX-NEXT: retq
|
|
%r = load <8 x i32>, ptr %p1, align 32
|
|
ret <8 x i32> %r
|
|
}
|
|
|
|
define <16 x i32> @test_load_v16i32_noalign(ptr %p1) {
|
|
; SKX-LABEL: test_load_v16i32_noalign:
|
|
; SKX: # %bb.0:
|
|
; SKX-NEXT: vmovups (%rdi), %zmm0
|
|
; SKX-NEXT: retq
|
|
%r = load <16 x i32>, ptr %p1, align 1
|
|
ret <16 x i32> %r
|
|
}
|
|
|
|
define <16 x i32> @test_load_v16i32_align(ptr %p1) {
|
|
; SKX-LABEL: test_load_v16i32_align:
|
|
; SKX: # %bb.0:
|
|
; SKX-NEXT: vmovups (%rdi), %zmm0
|
|
; SKX-NEXT: retq
|
|
%r = load <16 x i32>, ptr %p1, align 32
|
|
ret <16 x i32> %r
|
|
}
|
|
|
|
define void @test_store_v4i32_noalign(<4 x i32> %val, ptr %p1) {
|
|
; SKX-LABEL: test_store_v4i32_noalign:
|
|
; SKX: # %bb.0:
|
|
; SKX-NEXT: vmovups %xmm0, (%rdi)
|
|
; SKX-NEXT: retq
|
|
store <4 x i32> %val, ptr %p1, align 1
|
|
ret void
|
|
}
|
|
|
|
define void @test_store_v4i32_align(<4 x i32> %val, ptr %p1) {
|
|
; SKX-LABEL: test_store_v4i32_align:
|
|
; SKX: # %bb.0:
|
|
; SKX-NEXT: vmovaps %xmm0, (%rdi)
|
|
; SKX-NEXT: retq
|
|
store <4 x i32> %val, ptr %p1, align 16
|
|
ret void
|
|
}
|
|
|
|
define void @test_store_v8i32_noalign(<8 x i32> %val, ptr %p1) {
|
|
; SKX-LABEL: test_store_v8i32_noalign:
|
|
; SKX: # %bb.0:
|
|
; SKX-NEXT: vmovups %ymm0, (%rdi)
|
|
; SKX-NEXT: vzeroupper
|
|
; SKX-NEXT: retq
|
|
store <8 x i32> %val, ptr %p1, align 1
|
|
ret void
|
|
}
|
|
|
|
define void @test_store_v8i32_align(<8 x i32> %val, ptr %p1) {
|
|
; SKX-LABEL: test_store_v8i32_align:
|
|
; SKX: # %bb.0:
|
|
; SKX-NEXT: vmovaps %ymm0, (%rdi)
|
|
; SKX-NEXT: vzeroupper
|
|
; SKX-NEXT: retq
|
|
store <8 x i32> %val, ptr %p1, align 32
|
|
ret void
|
|
}
|
|
|
|
define void @test_store_v16i32_noalign(<16 x i32> %val, ptr %p1) {
|
|
; SKX-LABEL: test_store_v16i32_noalign:
|
|
; SKX: # %bb.0:
|
|
; SKX-NEXT: vmovups %zmm0, (%rdi)
|
|
; SKX-NEXT: vzeroupper
|
|
; SKX-NEXT: retq
|
|
store <16 x i32> %val, ptr %p1, align 1
|
|
ret void
|
|
}
|
|
|
|
define void @test_store_v16i32_align(<16 x i32> %val, ptr %p1) {
|
|
; SKX-LABEL: test_store_v16i32_align:
|
|
; SKX: # %bb.0:
|
|
; SKX-NEXT: vmovaps %zmm0, (%rdi)
|
|
; SKX-NEXT: vzeroupper
|
|
; SKX-NEXT: retq
|
|
store <16 x i32> %val, ptr %p1, align 64
|
|
ret void
|
|
}
|
|
|