66 lines · plain
1; RUN: opt -S -mtriple=amdgcn-- -amdgpu-lower-module-lds --amdgpu-lower-module-lds-strategy=module < %s | FileCheck %s2; RUN: opt -S -mtriple=amdgcn-- -passes=amdgpu-lower-module-lds --amdgpu-lower-module-lds-strategy=module < %s | FileCheck %s3 4@lds.size.1.align.1 = internal unnamed_addr addrspace(3) global [1 x i8] poison, align 15@lds.size.2.align.2 = internal unnamed_addr addrspace(3) global [2 x i8] poison, align 26@lds.size.4.align.4 = internal unnamed_addr addrspace(3) global [4 x i8] poison, align 47@lds.size.16.align.16 = internal unnamed_addr addrspace(3) global [16 x i8] poison, align 168 9; CHECK: %llvm.amdgcn.kernel.k0.lds.t = type { [16 x i8], [4 x i8], [2 x i8], [1 x i8] }10; CHECK: %llvm.amdgcn.kernel.k1.lds.t = type { [16 x i8], [4 x i8], [2 x i8] }11 12;.13; CHECK: @lds.k2 = addrspace(3) global [1 x i8] poison, align 114; CHECK: @llvm.amdgcn.kernel.k0.lds = internal addrspace(3) global %llvm.amdgcn.kernel.k0.lds.t poison, align 16, !absolute_symbol !015; CHECK: @llvm.amdgcn.kernel.k1.lds = internal addrspace(3) global %llvm.amdgcn.kernel.k1.lds.t poison, align 16, !absolute_symbol !016;.17define amdgpu_kernel void @k0() {18; CHECK-LABEL: @k0(19; CHECK-NEXT: store i8 1, ptr addrspace(3) getelementptr inbounds (%llvm.amdgcn.kernel.k0.lds.t, ptr addrspace(3) @llvm.amdgcn.kernel.k0.lds, i32 0, i32 3), align 2, !alias.scope !1, !noalias !420; CHECK-NEXT: store i8 2, ptr addrspace(3) getelementptr inbounds (%llvm.amdgcn.kernel.k0.lds.t, ptr addrspace(3) @llvm.amdgcn.kernel.k0.lds, i32 0, i32 2), align 4, !alias.scope !8, !noalias !921; CHECK-NEXT: store i8 4, ptr addrspace(3) getelementptr inbounds (%llvm.amdgcn.kernel.k0.lds.t, ptr addrspace(3) @llvm.amdgcn.kernel.k0.lds, i32 0, i32 1), align 16, !alias.scope !10, !noalias !1122; CHECK-NEXT: store i8 16, ptr addrspace(3) @llvm.amdgcn.kernel.k0.lds, align 16, !alias.scope !12, !noalias !1323; CHECK-NEXT: ret void24 store i8 1, ptr addrspace(3) @lds.size.1.align.1, align 125 26 store i8 2, ptr addrspace(3) @lds.size.2.align.2, align 227 28 store i8 4, ptr addrspace(3) @lds.size.4.align.4, align 429 30 store i8 16, ptr addrspace(3) @lds.size.16.align.16, align 1631 32 ret void33}34 35define amdgpu_kernel void @k1() {36; CHECK-LABEL: @k1(37; CHECK-NEXT: store i8 2, ptr addrspace(3) getelementptr inbounds (%llvm.amdgcn.kernel.k1.lds.t, ptr addrspace(3) @llvm.amdgcn.kernel.k1.lds, i32 0, i32 2), align 4, !alias.scope !14, !noalias !1738; CHECK-NEXT: store i8 4, ptr addrspace(3) getelementptr inbounds (%llvm.amdgcn.kernel.k1.lds.t, ptr addrspace(3) @llvm.amdgcn.kernel.k1.lds, i32 0, i32 1), align 16, !alias.scope !20, !noalias !2139; CHECK-NEXT: store i8 16, ptr addrspace(3) @llvm.amdgcn.kernel.k1.lds, align 16, !alias.scope !22, !noalias !2340; CHECK-NEXT: ret void41;42 store i8 2, ptr addrspace(3) @lds.size.2.align.2, align 243 44 store i8 4, ptr addrspace(3) @lds.size.4.align.4, align 445 46 store i8 16, ptr addrspace(3) @lds.size.16.align.16, align 1647 48 ret void49}50 51; Do not lower LDS for graphics shaders.52 53@lds.k2 = addrspace(3) global [1 x i8] poison, align 154 55define amdgpu_ps void @k2() {56; CHECK-LABEL: @k2(57; CHECK-NEXT: store i8 1, ptr addrspace(3) @lds.k2, align 158; CHECK-NEXT: ret void59;60 store i8 1, ptr addrspace(3) @lds.k2, align 161 62 ret void63}64 65; CHECK: !0 = !{i32 0, i32 1}66