1 ;RUN: llc < %s -march=amdgcn -mcpu=verde -verify-machineinstrs | FileCheck -check-prefix=VERDE %s
2 ;RUN: llc < %s -march=amdgcn -mcpu=tonga -verify-machineinstrs | FileCheck %s
4 ;CHECK-LABEL: {{^}}buffer_store:
6 ;CHECK: buffer_store_format_xyzw v[0:3], off, s[0:3], 0
7 ;CHECK: buffer_store_format_xyzw v[4:7], off, s[0:3], 0 glc
8 ;CHECK: buffer_store_format_xyzw v[8:11], off, s[0:3], 0 slc
9 define amdgpu_ps void @buffer_store(<4 x i32> inreg, <4 x float>, <4 x float>, <4 x float>) {
11 call void @llvm.amdgcn.buffer.store.format.v4f32(<4 x float> %1, <4 x i32> %0, i32 0, i32 0, i1 0, i1 0)
12 call void @llvm.amdgcn.buffer.store.format.v4f32(<4 x float> %2, <4 x i32> %0, i32 0, i32 0, i1 1, i1 0)
13 call void @llvm.amdgcn.buffer.store.format.v4f32(<4 x float> %3, <4 x i32> %0, i32 0, i32 0, i1 0, i1 1)
17 ;CHECK-LABEL: {{^}}buffer_store_immoffs:
19 ;CHECK: buffer_store_format_xyzw v[0:3], off, s[0:3], 0 offset:42
20 define amdgpu_ps void @buffer_store_immoffs(<4 x i32> inreg, <4 x float>) {
22 call void @llvm.amdgcn.buffer.store.format.v4f32(<4 x float> %1, <4 x i32> %0, i32 0, i32 42, i1 0, i1 0)
26 ;CHECK-LABEL: {{^}}buffer_store_idx:
28 ;CHECK: buffer_store_format_xyzw v[0:3], v4, s[0:3], 0 idxen
29 define amdgpu_ps void @buffer_store_idx(<4 x i32> inreg, <4 x float>, i32) {
31 call void @llvm.amdgcn.buffer.store.format.v4f32(<4 x float> %1, <4 x i32> %0, i32 %2, i32 0, i1 0, i1 0)
35 ;CHECK-LABEL: {{^}}buffer_store_ofs:
37 ;CHECK: buffer_store_format_xyzw v[0:3], v4, s[0:3], 0 offen
38 define amdgpu_ps void @buffer_store_ofs(<4 x i32> inreg, <4 x float>, i32) {
40 call void @llvm.amdgcn.buffer.store.format.v4f32(<4 x float> %1, <4 x i32> %0, i32 0, i32 %2, i1 0, i1 0)
44 ;CHECK-LABEL: {{^}}buffer_store_both:
46 ;CHECK: buffer_store_format_xyzw v[0:3], v[4:5], s[0:3], 0 idxen offen
47 define amdgpu_ps void @buffer_store_both(<4 x i32> inreg, <4 x float>, i32, i32) {
49 call void @llvm.amdgcn.buffer.store.format.v4f32(<4 x float> %1, <4 x i32> %0, i32 %2, i32 %3, i1 0, i1 0)
53 ;CHECK-LABEL: {{^}}buffer_store_both_reversed:
54 ;CHECK: v_mov_b32_e32 v6, v4
56 ;CHECK: buffer_store_format_xyzw v[0:3], v[5:6], s[0:3], 0 idxen offen
57 define amdgpu_ps void @buffer_store_both_reversed(<4 x i32> inreg, <4 x float>, i32, i32) {
59 call void @llvm.amdgcn.buffer.store.format.v4f32(<4 x float> %1, <4 x i32> %0, i32 %3, i32 %2, i1 0, i1 0)
63 ; Ideally, the register allocator would avoid the wait here
65 ;CHECK-LABEL: {{^}}buffer_store_wait:
67 ;CHECK: buffer_store_format_xyzw v[0:3], v4, s[0:3], 0 idxen
68 ;VERDE: s_waitcnt expcnt(0)
69 ;CHECK: buffer_load_format_xyzw v[0:3], v5, s[0:3], 0 idxen
70 ;CHECK: s_waitcnt vmcnt(0)
71 ;CHECK: buffer_store_format_xyzw v[0:3], v6, s[0:3], 0 idxen
72 define amdgpu_ps void @buffer_store_wait(<4 x i32> inreg, <4 x float>, i32, i32, i32) {
74 call void @llvm.amdgcn.buffer.store.format.v4f32(<4 x float> %1, <4 x i32> %0, i32 %2, i32 0, i1 0, i1 0)
75 %data = call <4 x float> @llvm.amdgcn.buffer.load.format.v4f32(<4 x i32> %0, i32 %3, i32 0, i1 0, i1 0)
76 call void @llvm.amdgcn.buffer.store.format.v4f32(<4 x float> %data, <4 x i32> %0, i32 %4, i32 0, i1 0, i1 0)
80 ;CHECK-LABEL: {{^}}buffer_store_x1:
82 ;CHECK: buffer_store_format_x v0, v1, s[0:3], 0 idxen
83 define amdgpu_ps void @buffer_store_x1(<4 x i32> inreg %rsrc, float %data, i32 %index) {
85 call void @llvm.amdgcn.buffer.store.format.f32(float %data, <4 x i32> %rsrc, i32 %index, i32 0, i1 0, i1 0)
89 ;CHECK-LABEL: {{^}}buffer_store_x2:
91 ;CHECK: buffer_store_format_xy v[0:1], v2, s[0:3], 0 idxen
92 define amdgpu_ps void @buffer_store_x2(<4 x i32> inreg %rsrc, <2 x float> %data, i32 %index) {
94 call void @llvm.amdgcn.buffer.store.format.v2f32(<2 x float> %data, <4 x i32> %rsrc, i32 %index, i32 0, i1 0, i1 0)
98 declare void @llvm.amdgcn.buffer.store.format.f32(float, <4 x i32>, i32, i32, i1, i1) #0
99 declare void @llvm.amdgcn.buffer.store.format.v2f32(<2 x float>, <4 x i32>, i32, i32, i1, i1) #0
100 declare void @llvm.amdgcn.buffer.store.format.v4f32(<4 x float>, <4 x i32>, i32, i32, i1, i1) #0
101 declare <4 x float> @llvm.amdgcn.buffer.load.format.v4f32(<4 x i32>, i32, i32, i1, i1) #1
103 attributes #0 = { nounwind }
104 attributes #1 = { nounwind readonly }