Files
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
Flatten subprojects into a single base monorepo
2026-09-02 16:34:47 -07:00

148 lines
6.4 KiB
LLVM

; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -global-isel -new-reg-bank-select -new-reg-bank-select -mtriple=amdgcn -o - %s | FileCheck %s
define amdgpu_cs void @test1(i32 %arg1, <4 x i32> inreg %arg2, i32, ptr addrspace(6) inreg %arg3) {
; CHECK-LABEL: test1:
; CHECK: ; %bb.0: ; %.entry
; CHECK-NEXT: v_and_b32_e32 v3, 0x3ffffffc, v0
; CHECK-NEXT: v_mov_b32_e32 v0, 11
; CHECK-NEXT: v_mov_b32_e32 v1, 22
; CHECK-NEXT: v_mov_b32_e32 v2, 33
; CHECK-NEXT: v_lshlrev_b32_e32 v3, 2, v3
; CHECK-NEXT: v_add_i32_e32 v4, vcc, s4, v3
; CHECK-NEXT: v_mov_b32_e32 v3, 44
; CHECK-NEXT: buffer_store_dwordx4 v[0:3], v4, s[0:3], 0 offen
; CHECK-NEXT: s_endpgm
.entry:
%bs1 = and i32 %arg1, 1073741820
%ep1 = getelementptr i32, ptr addrspace(6) %arg3, i32 %bs1
%ad1 = ptrtoint ptr addrspace(6) %ep1 to i32
call void @llvm.amdgcn.raw.buffer.store.i32(i32 11, <4 x i32> %arg2, i32 %ad1, i32 0, i32 0)
%bs2 = or disjoint i32 %bs1, 1
%ep2 = getelementptr i32, ptr addrspace(6) %arg3, i32 %bs2
%ad2 = ptrtoint ptr addrspace(6) %ep2 to i32
call void @llvm.amdgcn.raw.buffer.store.i32(i32 22, <4 x i32> %arg2, i32 %ad2, i32 0, i32 0)
%bs3 = or disjoint i32 %bs1, 2
%ep3 = getelementptr i32, ptr addrspace(6) %arg3, i32 %bs3
%ad3 = ptrtoint ptr addrspace(6) %ep3 to i32
call void @llvm.amdgcn.raw.buffer.store.i32(i32 33, <4 x i32> %arg2, i32 %ad3, i32 0, i32 0)
%bs4 = or disjoint i32 %bs1, 3
%ep4 = getelementptr i32, ptr addrspace(6) %arg3, i32 %bs4
%ad4 = ptrtoint ptr addrspace(6) %ep4 to i32
call void @llvm.amdgcn.raw.buffer.store.i32(i32 44, <4 x i32> %arg2, i32 %ad4, i32 0, i32 0)
ret void
}
define amdgpu_cs void @test1_ptr(i32 %arg1, ptr addrspace(8) inreg %arg2, i32, ptr addrspace(6) inreg %arg3) {
; CHECK-LABEL: test1_ptr:
; CHECK: ; %bb.0: ; %.entry
; CHECK-NEXT: v_and_b32_e32 v3, 0x3ffffffc, v0
; CHECK-NEXT: v_mov_b32_e32 v0, 11
; CHECK-NEXT: v_mov_b32_e32 v1, 22
; CHECK-NEXT: v_mov_b32_e32 v2, 33
; CHECK-NEXT: v_lshlrev_b32_e32 v3, 2, v3
; CHECK-NEXT: v_add_i32_e32 v4, vcc, s4, v3
; CHECK-NEXT: v_mov_b32_e32 v3, 44
; CHECK-NEXT: buffer_store_dwordx4 v[0:3], v4, s[0:3], 0 offen
; CHECK-NEXT: s_endpgm
.entry:
%bs1 = and i32 %arg1, 1073741820
%ep1 = getelementptr i32, ptr addrspace(6) %arg3, i32 %bs1
%ad1 = ptrtoint ptr addrspace(6) %ep1 to i32
call void @llvm.amdgcn.raw.ptr.buffer.store.i32(i32 11, ptr addrspace(8) %arg2, i32 %ad1, i32 0, i32 0)
%bs2 = or disjoint i32 %bs1, 1
%ep2 = getelementptr i32, ptr addrspace(6) %arg3, i32 %bs2
%ad2 = ptrtoint ptr addrspace(6) %ep2 to i32
call void @llvm.amdgcn.raw.ptr.buffer.store.i32(i32 22, ptr addrspace(8) %arg2, i32 %ad2, i32 0, i32 0)
%bs3 = or disjoint i32 %bs1, 2
%ep3 = getelementptr i32, ptr addrspace(6) %arg3, i32 %bs3
%ad3 = ptrtoint ptr addrspace(6) %ep3 to i32
call void @llvm.amdgcn.raw.ptr.buffer.store.i32(i32 33, ptr addrspace(8) %arg2, i32 %ad3, i32 0, i32 0)
%bs4 = or disjoint i32 %bs1, 3
%ep4 = getelementptr i32, ptr addrspace(6) %arg3, i32 %bs4
%ad4 = ptrtoint ptr addrspace(6) %ep4 to i32
call void @llvm.amdgcn.raw.ptr.buffer.store.i32(i32 44, ptr addrspace(8) %arg2, i32 %ad4, i32 0, i32 0)
ret void
}
define amdgpu_cs void @test2(i32 %arg1, <4 x i32> inreg %arg2) {
; CHECK-LABEL: test2:
; CHECK: ; %bb.0: ; %.entry
; CHECK-NEXT: v_and_b32_e32 v3, 0x3ffffffc, v0
; CHECK-NEXT: v_mov_b32_e32 v0, 11
; CHECK-NEXT: v_mov_b32_e32 v1, 22
; CHECK-NEXT: v_mov_b32_e32 v2, 33
; CHECK-NEXT: v_lshlrev_b32_e32 v4, 2, v3
; CHECK-NEXT: v_mov_b32_e32 v3, 44
; CHECK-NEXT: buffer_store_dwordx4 v[0:3], v4, s[0:3], 0 offen
; CHECK-NEXT: s_endpgm
.entry:
%bs1 = and i32 %arg1, 1073741820
%ep1 = getelementptr <{ [64 x i32] }>, ptr addrspace(6) null, i32 0, i32 0, i32 %bs1
%ad1 = ptrtoint ptr addrspace(6) %ep1 to i32
call void @llvm.amdgcn.raw.buffer.store.i32(i32 11, <4 x i32> %arg2, i32 %ad1, i32 0, i32 0)
%bs2 = or disjoint i32 %bs1, 1
%ep2 = getelementptr <{ [64 x i32] }>, ptr addrspace(6) null, i32 0, i32 0, i32 %bs2
%ad2 = ptrtoint ptr addrspace(6) %ep2 to i32
call void @llvm.amdgcn.raw.buffer.store.i32(i32 22, <4 x i32> %arg2, i32 %ad2, i32 0, i32 0)
%bs3 = or disjoint i32 %bs1, 2
%ep3 = getelementptr <{ [64 x i32] }>, ptr addrspace(6) null, i32 0, i32 0, i32 %bs3
%ad3 = ptrtoint ptr addrspace(6) %ep3 to i32
call void @llvm.amdgcn.raw.buffer.store.i32(i32 33, <4 x i32> %arg2, i32 %ad3, i32 0, i32 0)
%bs4 = or disjoint i32 %bs1, 3
%ep4 = getelementptr <{ [64 x i32] }>, ptr addrspace(6) null, i32 0, i32 0, i32 %bs4
%ad4 = ptrtoint ptr addrspace(6) %ep4 to i32
call void @llvm.amdgcn.raw.buffer.store.i32(i32 44, <4 x i32> %arg2, i32 %ad4, i32 0, i32 0)
ret void
}
define amdgpu_cs void @test2_ptr(i32 %arg1, ptr addrspace(8) inreg %arg2) {
; CHECK-LABEL: test2_ptr:
; CHECK: ; %bb.0: ; %.entry
; CHECK-NEXT: v_and_b32_e32 v3, 0x3ffffffc, v0
; CHECK-NEXT: v_mov_b32_e32 v0, 11
; CHECK-NEXT: v_mov_b32_e32 v1, 22
; CHECK-NEXT: v_mov_b32_e32 v2, 33
; CHECK-NEXT: v_lshlrev_b32_e32 v4, 2, v3
; CHECK-NEXT: v_mov_b32_e32 v3, 44
; CHECK-NEXT: buffer_store_dwordx4 v[0:3], v4, s[0:3], 0 offen
; CHECK-NEXT: s_endpgm
.entry:
%bs1 = and i32 %arg1, 1073741820
%ep1 = getelementptr <{ [64 x i32] }>, ptr addrspace(6) null, i32 0, i32 0, i32 %bs1
%ad1 = ptrtoint ptr addrspace(6) %ep1 to i32
call void @llvm.amdgcn.raw.ptr.buffer.store.i32(i32 11, ptr addrspace(8) %arg2, i32 %ad1, i32 0, i32 0)
%bs2 = or disjoint i32 %bs1, 1
%ep2 = getelementptr <{ [64 x i32] }>, ptr addrspace(6) null, i32 0, i32 0, i32 %bs2
%ad2 = ptrtoint ptr addrspace(6) %ep2 to i32
call void @llvm.amdgcn.raw.ptr.buffer.store.i32(i32 22, ptr addrspace(8) %arg2, i32 %ad2, i32 0, i32 0)
%bs3 = or disjoint i32 %bs1, 2
%ep3 = getelementptr <{ [64 x i32] }>, ptr addrspace(6) null, i32 0, i32 0, i32 %bs3
%ad3 = ptrtoint ptr addrspace(6) %ep3 to i32
call void @llvm.amdgcn.raw.ptr.buffer.store.i32(i32 33, ptr addrspace(8) %arg2, i32 %ad3, i32 0, i32 0)
%bs4 = or disjoint i32 %bs1, 3
%ep4 = getelementptr <{ [64 x i32] }>, ptr addrspace(6) null, i32 0, i32 0, i32 %bs4
%ad4 = ptrtoint ptr addrspace(6) %ep4 to i32
call void @llvm.amdgcn.raw.ptr.buffer.store.i32(i32 44, ptr addrspace(8) %arg2, i32 %ad4, i32 0, i32 0)
ret void
}
declare void @llvm.amdgcn.raw.buffer.store.i32(i32, <4 x i32>, i32, i32, i32 immarg)
declare void @llvm.amdgcn.raw.ptr.buffer.store.i32(i32, ptr addrspace(8), i32, i32, i32 immarg)