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
66 lines
3.4 KiB
LLVM
66 lines
3.4 KiB
LLVM
; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
|
|
; RUN: opt -S -mtriple=amdgcn-amd-amdhsa -passes=amdgpu-promote-alloca -disable-promote-alloca-to-lds=1 < %s | FileCheck %s
|
|
|
|
; Check that the extracted index is correctly sign-extended when 32-bit scratch
|
|
; address arithmetic is promoted to 64-bit vector index arithmetic.
|
|
|
|
define amdgpu_kernel void @negative_index_byte(ptr %out, i64 %offset) {
|
|
; CHECK-LABEL: @negative_index_byte(
|
|
; CHECK-NEXT: [[STACK:%.*]] = freeze <4 x i8> poison
|
|
; CHECK-NEXT: [[TMP1:%.*]] = insertelement <4 x i8> [[STACK]], i8 0, i32 0
|
|
; CHECK-NEXT: [[TMP2:%.*]] = insertelement <4 x i8> [[TMP1]], i8 1, i32 1
|
|
; CHECK-NEXT: [[TMP3:%.*]] = insertelement <4 x i8> [[TMP2]], i8 2, i32 2
|
|
; CHECK-NEXT: [[TMP4:%.*]] = insertelement <4 x i8> [[TMP3]], i8 3, i32 3
|
|
; CHECK-NEXT: [[TMP5:%.*]] = trunc i64 [[OFFSET:%.*]] to i32
|
|
; CHECK-NEXT: [[TMP7:%.*]] = add i32 [[TMP5]], -1
|
|
; CHECK-NEXT: [[TMP6:%.*]] = extractelement <4 x i8> [[TMP4]], i32 [[TMP7]]
|
|
; CHECK-NEXT: store i8 [[TMP6]], ptr [[OUT:%.*]], align 1
|
|
; CHECK-NEXT: ret void
|
|
;
|
|
%stack = alloca [4 x i8], align 4, addrspace(5)
|
|
%gep.0 = getelementptr inbounds [4 x i8], ptr addrspace(5) %stack, i64 0, i64 0
|
|
%gep.1 = getelementptr inbounds [4 x i8], ptr addrspace(5) %stack, i64 0, i64 1
|
|
%gep.2 = getelementptr inbounds [4 x i8], ptr addrspace(5) %stack, i64 0, i64 2
|
|
%gep.3 = getelementptr inbounds [4 x i8], ptr addrspace(5) %stack, i64 0, i64 3
|
|
store i8 0, ptr addrspace(5) %gep.0
|
|
store i8 1, ptr addrspace(5) %gep.1
|
|
store i8 2, ptr addrspace(5) %gep.2
|
|
store i8 3, ptr addrspace(5) %gep.3
|
|
%vgep = getelementptr inbounds [4 x i8], ptr addrspace(5) %stack, i64 0, i64 %offset
|
|
%cgep = getelementptr inbounds [4 x i8], ptr addrspace(5) %vgep, i64 0, i64 -1
|
|
%load = load i8, ptr addrspace(5) %cgep
|
|
store i8 %load, ptr %out
|
|
ret void
|
|
}
|
|
|
|
define amdgpu_kernel void @negative_index_word(ptr %out, i64 %offset) {
|
|
; CHECK-LABEL: @negative_index_word(
|
|
; CHECK-NEXT: [[STACK:%.*]] = freeze <4 x i32> poison
|
|
; CHECK-NEXT: [[TMP1:%.*]] = insertelement <4 x i32> [[STACK]], i32 0, i32 0
|
|
; CHECK-NEXT: [[TMP2:%.*]] = insertelement <4 x i32> [[TMP1]], i32 1, i32 1
|
|
; CHECK-NEXT: [[TMP3:%.*]] = insertelement <4 x i32> [[TMP2]], i32 2, i32 2
|
|
; CHECK-NEXT: [[TMP4:%.*]] = insertelement <4 x i32> [[TMP3]], i32 3, i32 3
|
|
; CHECK-NEXT: [[TMP5:%.*]] = trunc i64 [[OFFSET:%.*]] to i32
|
|
; CHECK-NEXT: [[TMP7:%.*]] = add i32 [[TMP5]], -1
|
|
; CHECK-NEXT: [[TMP6:%.*]] = extractelement <4 x i32> [[TMP4]], i32 [[TMP7]]
|
|
; CHECK-NEXT: store i32 [[TMP6]], ptr [[OUT:%.*]], align 4
|
|
; CHECK-NEXT: ret void
|
|
;
|
|
%stack = alloca [4 x i32], align 4, addrspace(5)
|
|
%gep.0 = getelementptr inbounds [4 x i32], ptr addrspace(5) %stack, i64 0, i64 0
|
|
%gep.1 = getelementptr inbounds [4 x i32], ptr addrspace(5) %stack, i64 0, i64 1
|
|
%gep.2 = getelementptr inbounds [4 x i32], ptr addrspace(5) %stack, i64 0, i64 2
|
|
%gep.3 = getelementptr inbounds [4 x i32], ptr addrspace(5) %stack, i64 0, i64 3
|
|
store i32 0, ptr addrspace(5) %gep.0
|
|
store i32 1, ptr addrspace(5) %gep.1
|
|
store i32 2, ptr addrspace(5) %gep.2
|
|
store i32 3, ptr addrspace(5) %gep.3
|
|
%vgep = getelementptr inbounds [4 x i32], ptr addrspace(5) %stack, i64 0, i64 %offset
|
|
%cgep = getelementptr inbounds [4 x i32], ptr addrspace(5) %vgep, i64 0, i64 -1
|
|
%load = load i32, ptr addrspace(5) %cgep
|
|
store i32 %load, ptr %out
|
|
ret void
|
|
}
|
|
|
|
|