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
3.9 KiB
LLVM
57 lines
3.9 KiB
LLVM
; RUN: llc < %s -mtriple=r600 -mcpu=r600 | FileCheck %s
|
|
; RUN: llc < %s -mtriple=r600 -mcpu=rs880 | FileCheck %s
|
|
; RUN: llc < %s -mtriple=r600 -mcpu=rv670 | FileCheck %s
|
|
|
|
; R600 supports 8 fetches in a clause
|
|
; CHECK: {{^}}fetch_limits_r600:
|
|
; CHECK: Fetch clause
|
|
; CHECK: Fetch clause
|
|
|
|
define amdgpu_ps void @fetch_limits_r600() {
|
|
entry:
|
|
%tmp = load <4 x float>, ptr addrspace(8) null
|
|
%tmp1 = load <4 x float>, ptr addrspace(8) getelementptr ([1024 x <4 x float>], ptr addrspace(8) null, i64 0, i32 1)
|
|
%tmp2 = load <4 x float>, ptr addrspace(8) getelementptr ([1024 x <4 x float>], ptr addrspace(8) null, i64 0, i32 2)
|
|
%tmp3 = load <4 x float>, ptr addrspace(8) getelementptr ([1024 x <4 x float>], ptr addrspace(8) null, i64 0, i32 3)
|
|
%tmp4 = load <4 x float>, ptr addrspace(8) getelementptr ([1024 x <4 x float>], ptr addrspace(8) null, i64 0, i32 4)
|
|
%tmp5 = load <4 x float>, ptr addrspace(8) getelementptr ([1024 x <4 x float>], ptr addrspace(8) null, i64 0, i32 5)
|
|
%tmp6 = load <4 x float>, ptr addrspace(8) getelementptr ([1024 x <4 x float>], ptr addrspace(8) null, i64 0, i32 6)
|
|
%tmp7 = load <4 x float>, ptr addrspace(8) getelementptr ([1024 x <4 x float>], ptr addrspace(8) null, i64 0, i32 7)
|
|
%tmp8 = load <4 x float>, ptr addrspace(8) getelementptr ([1024 x <4 x float>], ptr addrspace(8) null, i64 0, i32 8)
|
|
%tmp9 = shufflevector <4 x float> %tmp, <4 x float> %tmp, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
|
|
%tmp10 = call <4 x float> @llvm.r600.tex(<4 x float> %tmp9, i32 0, i32 0, i32 0, i32 0, i32 0, i32 1, i32 1, i32 1, i32 1)
|
|
%tmp11 = shufflevector <4 x float> %tmp1, <4 x float> %tmp1, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
|
|
%tmp12 = call <4 x float> @llvm.r600.tex(<4 x float> %tmp11, i32 0, i32 0, i32 0, i32 0, i32 0, i32 1, i32 1, i32 1, i32 1)
|
|
%tmp13 = shufflevector <4 x float> %tmp2, <4 x float> %tmp2, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
|
|
%tmp14 = call <4 x float> @llvm.r600.tex(<4 x float> %tmp13, i32 0, i32 0, i32 0, i32 0, i32 0, i32 1, i32 1, i32 1, i32 1)
|
|
%tmp15 = shufflevector <4 x float> %tmp3, <4 x float> %tmp3, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
|
|
%tmp16 = call <4 x float> @llvm.r600.tex(<4 x float> %tmp15, i32 0, i32 0, i32 0, i32 0, i32 0, i32 1, i32 1, i32 1, i32 1)
|
|
%tmp17 = shufflevector <4 x float> %tmp4, <4 x float> %tmp4, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
|
|
%tmp18 = call <4 x float> @llvm.r600.tex(<4 x float> %tmp17, i32 0, i32 0, i32 0, i32 0, i32 0, i32 1, i32 1, i32 1, i32 1)
|
|
%tmp19 = shufflevector <4 x float> %tmp5, <4 x float> %tmp5, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
|
|
%tmp20 = call <4 x float> @llvm.r600.tex(<4 x float> %tmp19, i32 0, i32 0, i32 0, i32 0, i32 0, i32 1, i32 1, i32 1, i32 1)
|
|
%tmp21 = shufflevector <4 x float> %tmp6, <4 x float> %tmp6, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
|
|
%tmp22 = call <4 x float> @llvm.r600.tex(<4 x float> %tmp21, i32 0, i32 0, i32 0, i32 0, i32 0, i32 1, i32 1, i32 1, i32 1)
|
|
%tmp23 = shufflevector <4 x float> %tmp7, <4 x float> %tmp7, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
|
|
%tmp24 = call <4 x float> @llvm.r600.tex(<4 x float> %tmp23, i32 0, i32 0, i32 0, i32 0, i32 0, i32 1, i32 1, i32 1, i32 1)
|
|
%tmp25 = shufflevector <4 x float> %tmp8, <4 x float> %tmp8, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
|
|
%tmp26 = call <4 x float> @llvm.r600.tex(<4 x float> %tmp25, i32 0, i32 0, i32 0, i32 0, i32 0, i32 1, i32 1, i32 1, i32 1)
|
|
%a = fadd <4 x float> %tmp10, %tmp12
|
|
%b = fadd <4 x float> %tmp14, %tmp16
|
|
%c = fadd <4 x float> %tmp18, %tmp20
|
|
%d = fadd <4 x float> %tmp22, %tmp24
|
|
%e = fadd <4 x float> %tmp26, %a
|
|
%bc = fadd <4 x float> %b, %c
|
|
%de = fadd <4 x float> %d, %e
|
|
%bcde = fadd <4 x float> %bc, %de
|
|
call void @llvm.r600.store.swizzle(<4 x float> %bcde, i32 0, i32 1)
|
|
ret void
|
|
}
|
|
|
|
declare void @llvm.r600.store.swizzle(<4 x float>, i32, i32)
|
|
|
|
; Function Attrs: readnone
|
|
declare <4 x float> @llvm.r600.tex(<4 x float>, i32, i32, i32, i32, i32, i32, i32, i32, i32) #0
|
|
|
|
attributes #0 = { nounwind readnone }
|