1 ; RUN: llc -march=amdgcn -mcpu=bonaire -verify-machineinstrs -mattr=+load-store-opt < %s | FileCheck -strict-whitespace -check-prefix=SI %s
2 ; RUN: llc -march=amdgcn -mcpu=tonga -verify-machineinstrs -mattr=+load-store-opt < %s | FileCheck -strict-whitespace -check-prefix=SI %s
5 @lds = addrspace(3) global [512 x float] undef, align 4
7 ; offset0 is larger than offset1
9 ; SI-LABEL: {{^}}offset_order:
11 ; SI: ds_read2st64_b32 v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}} offset1:4{{$}}
12 ; SI: ds_read2_b32 v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}} offset0:3 offset1:2
13 ; SI: ds_read2_b32 v[{{[0-9]+}}:{{[0-9]+}}], v{{[0-9]+}} offset0:12 offset1:14
14 ; SI: ds_read_b32 v{{[0-9]+}}, v{{[0-9]+}} offset:44
16 define void @offset_order(float addrspace(1)* %out) {
18 %ptr0 = getelementptr inbounds [512 x float], [512 x float] addrspace(3)* @lds, i32 0, i32 0
19 %val0 = load float, float addrspace(3)* %ptr0
21 %ptr1 = getelementptr inbounds [512 x float], [512 x float] addrspace(3)* @lds, i32 0, i32 256
22 %val1 = load float, float addrspace(3)* %ptr1
23 %add1 = fadd float %val0, %val1
25 %ptr2 = getelementptr inbounds [512 x float], [512 x float] addrspace(3)* @lds, i32 0, i32 3
26 %val2 = load float, float addrspace(3)* %ptr2
27 %add2 = fadd float %add1, %val2
29 %ptr3 = getelementptr inbounds [512 x float], [512 x float] addrspace(3)* @lds, i32 0, i32 2
30 %val3 = load float, float addrspace(3)* %ptr3
31 %add3 = fadd float %add2, %val3
33 %ptr4 = getelementptr inbounds [512 x float], [512 x float] addrspace(3)* @lds, i32 0, i32 12
34 %val4 = load float, float addrspace(3)* %ptr4
35 %add4 = fadd float %add3, %val4
37 %ptr5 = getelementptr inbounds [512 x float], [512 x float] addrspace(3)* @lds, i32 0, i32 14
38 %val5 = load float, float addrspace(3)* %ptr5
39 %add5 = fadd float %add4, %val5
41 %ptr6 = getelementptr inbounds [512 x float], [512 x float] addrspace(3)* @lds, i32 0, i32 11
42 %val6 = load float, float addrspace(3)* %ptr6
43 %add6 = fadd float %add5, %val6
44 store float %add6, float addrspace(1)* %out