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
57 lines
2.1 KiB
LLVM
57 lines
2.1 KiB
LLVM
; RUN: opt -S -mtriple=amdgcn-unknown-amdhsa -passes=loop-unroll < %s | FileCheck %s
|
|
|
|
; Test that max iterations count to analyze (specific for the target)
|
|
; is enough to make the inner loop completely unrolled
|
|
; CHECK-LABEL: foo
|
|
define void @foo(ptr addrspace(5) %ptrB, ptr addrspace(5) %ptrC, i32 %A, i32 %A2, float %M) {
|
|
bb:
|
|
br label %bb2
|
|
|
|
bb2: ; preds = %bb7, %bb
|
|
%i = phi i32 [ 0, %bb ], [ %i8, %bb7 ]
|
|
br label %bb4
|
|
|
|
bb3: ; preds = %bb7
|
|
ret void
|
|
|
|
bb4: ; preds = %bb10, %bb2
|
|
%i5 = phi i32 [ 0, %bb2 ], [ %i11, %bb10 ]
|
|
%i6 = add nuw nsw i32 %i5, %i
|
|
br label %for.body
|
|
|
|
bb7: ; preds = %bb10
|
|
%i8 = add nuw nsw i32 %i, 1
|
|
%i9 = icmp eq i32 %i8, 8
|
|
br i1 %i9, label %bb3, label %bb2
|
|
|
|
bb10: ; preds = %for.body
|
|
%i11 = add nuw nsw i32 %i5, 1
|
|
%cmpj = icmp ult i32 %i11, 8
|
|
br i1 %cmpj, label %bb7, label %bb4
|
|
|
|
; CHECK-LABEL: for.body
|
|
; CHECK-NOT: %phi = phi {{.*}}
|
|
for.body: ; preds = %bb4, %for.body
|
|
%phi = phi i32 [ 0, %bb4 ], [ %inc, %for.body ]
|
|
%mul = shl nuw nsw i32 %phi, 6
|
|
%add = add i32 %A, %mul
|
|
%arrayidx = getelementptr inbounds float, ptr addrspace(5) %ptrC, i32 %add
|
|
%ld1 = load float, ptr addrspace(5) %arrayidx, align 4
|
|
%mul2 = shl nuw nsw i32 %phi, 3
|
|
%add2 = add i32 %A2, %mul2
|
|
%arrayidx2 = getelementptr inbounds float, ptr addrspace(5) %ptrB, i32 %add2
|
|
%ld2 = load float, ptr addrspace(5) %arrayidx2, align 4
|
|
%mul3 = fmul contract float %M, %ld2
|
|
%add3 = fadd contract float %ld1, %mul3
|
|
store float %add3, ptr addrspace(5) %arrayidx, align 4
|
|
%add1 = add nuw nsw i32 %add, 2048
|
|
%arrayidx3 = getelementptr inbounds float, ptr addrspace(5) %ptrC, i32 %add1
|
|
%ld3 = load float, ptr addrspace(5) %arrayidx3, align 4
|
|
%mul4 = fmul contract float %ld2, %M
|
|
%add4 = fadd contract float %ld3, %mul4
|
|
store float %add4, ptr addrspace(5) %arrayidx3, align 4
|
|
%inc = add nuw nsw i32 %phi, 1
|
|
%cmpi = icmp ult i32 %phi, 31
|
|
br i1 %cmpi, label %for.body, label %bb10
|
|
}
|