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
80 lines
2.2 KiB
LLVM
80 lines
2.2 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
|
|
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=-slow-unaligned-mem-16 | FileCheck %s --check-prefix=FAST
|
|
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+slow-unaligned-mem-16 | FileCheck %s --check-prefix=SLOW
|
|
|
|
; Verify that the DAGCombiner is creating unaligned 16-byte loads and stores
|
|
; if and only if those are fast.
|
|
|
|
define void @merge_const_vec_store(ptr %ptr) {
|
|
; FAST-LABEL: merge_const_vec_store:
|
|
; FAST: # %bb.0:
|
|
; FAST-NEXT: xorps %xmm0, %xmm0
|
|
; FAST-NEXT: movups %xmm0, (%rdi)
|
|
; FAST-NEXT: retq
|
|
;
|
|
; SLOW-LABEL: merge_const_vec_store:
|
|
; SLOW: # %bb.0:
|
|
; SLOW-NEXT: movq $0, (%rdi)
|
|
; SLOW-NEXT: movq $0, 8(%rdi)
|
|
; SLOW-NEXT: retq
|
|
|
|
%idx1 = getelementptr i64, ptr %ptr, i64 1
|
|
|
|
store i64 0, ptr %ptr, align 8
|
|
store i64 0, ptr %idx1, align 8
|
|
ret void
|
|
}
|
|
|
|
|
|
define void @merge_vec_element_store(<4 x double> %v, ptr %ptr) {
|
|
; FAST-LABEL: merge_vec_element_store:
|
|
; FAST: # %bb.0:
|
|
; FAST-NEXT: movups %xmm0, (%rdi)
|
|
; FAST-NEXT: retq
|
|
;
|
|
; SLOW-LABEL: merge_vec_element_store:
|
|
; SLOW: # %bb.0:
|
|
; SLOW-NEXT: movlps %xmm0, (%rdi)
|
|
; SLOW-NEXT: movhps %xmm0, 8(%rdi)
|
|
; SLOW-NEXT: retq
|
|
|
|
%vecext0 = extractelement <4 x double> %v, i32 0
|
|
%vecext1 = extractelement <4 x double> %v, i32 1
|
|
|
|
%idx1 = getelementptr double, ptr %ptr, i64 1
|
|
|
|
store double %vecext0, ptr %ptr, align 8
|
|
store double %vecext1, ptr %idx1, align 8
|
|
ret void
|
|
}
|
|
|
|
|
|
define void @merge_vec_load_and_stores(ptr%ptr) {
|
|
; FAST-LABEL: merge_vec_load_and_stores:
|
|
; FAST: # %bb.0:
|
|
; FAST-NEXT: movups (%rdi), %xmm0
|
|
; FAST-NEXT: movups %xmm0, 40(%rdi)
|
|
; FAST-NEXT: retq
|
|
;
|
|
; SLOW-LABEL: merge_vec_load_and_stores:
|
|
; SLOW: # %bb.0:
|
|
; SLOW-NEXT: movq (%rdi), %rax
|
|
; SLOW-NEXT: movq 8(%rdi), %rcx
|
|
; SLOW-NEXT: movq %rax, 40(%rdi)
|
|
; SLOW-NEXT: movq %rcx, 48(%rdi)
|
|
; SLOW-NEXT: retq
|
|
|
|
%idx1 = getelementptr i64, ptr %ptr, i64 1
|
|
|
|
%ld0 = load i64, ptr %ptr, align 4
|
|
%ld1 = load i64, ptr %idx1, align 4
|
|
|
|
%idx4 = getelementptr i64, ptr %ptr, i64 5
|
|
%idx5 = getelementptr i64, ptr %ptr, i64 6
|
|
|
|
store i64 %ld0, ptr %idx4, align 4
|
|
store i64 %ld1, ptr %idx5, align 4
|
|
ret void
|
|
}
|
|
|