• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1; RUN: llc -march=amdgcn -verify-machineinstrs < %s | FileCheck -check-prefix=SI -check-prefix=FUNC %s
2; RUN: llc -march=amdgcn -mcpu=tonga -verify-machineinstrs < %s | FileCheck -check-prefix=SI -check-prefix=FUNC %s
3; RUN: llc -march=r600 -mcpu=redwood < %s | FileCheck -check-prefix=R600 -check-prefix=FUNC %s
4
5; FUNC-LABEL: {{^}}v_fsub_f32:
6; SI: v_subrev_f32_e32 {{v[0-9]+}}, {{v[0-9]+}}, {{v[0-9]+}}
7define void @v_fsub_f32(float addrspace(1)* %out, float addrspace(1)* %in) {
8  %b_ptr = getelementptr float, float addrspace(1)* %in, i32 1
9  %a = load float, float addrspace(1)* %in, align 4
10  %b = load float, float addrspace(1)* %b_ptr, align 4
11  %result = fsub float %a, %b
12  store float %result, float addrspace(1)* %out, align 4
13  ret void
14}
15
16; FUNC-LABEL: {{^}}s_fsub_f32:
17; R600: ADD {{\** *}}T{{[0-9]+\.[XYZW]}}, KC0[2].Z, -KC0[2].W
18
19; SI: v_sub_f32_e32 {{v[0-9]+}}, {{s[0-9]+}}, {{v[0-9]+}}
20define void @s_fsub_f32(float addrspace(1)* %out, float %a, float %b) {
21  %sub = fsub float %a, %b
22  store float %sub, float addrspace(1)* %out, align 4
23  ret void
24}
25
26declare float @llvm.R600.load.input(i32) readnone
27
28declare void @llvm.AMDGPU.store.output(float, i32)
29
30; FUNC-LABEL: {{^}}fsub_v2f32:
31; R600-DAG: ADD {{\** *}}T{{[0-9]+\.[XYZW]}}, KC0[3].X, -KC0[3].Z
32; R600-DAG: ADD {{\** *}}T{{[0-9]+\.[XYZW]}}, KC0[2].W, -KC0[3].Y
33
34; SI: v_subrev_f32_e32 {{v[0-9]+}}, {{s[0-9]+}}, {{v[0-9]+}}
35; SI: v_subrev_f32_e32 {{v[0-9]+}}, {{s[0-9]+}}, {{v[0-9]+}}
36define void @fsub_v2f32(<2 x float> addrspace(1)* %out, <2 x float> %a, <2 x float> %b) {
37  %sub = fsub <2 x float> %a, %b
38  store <2 x float> %sub, <2 x float> addrspace(1)* %out, align 8
39  ret void
40}
41
42; FUNC-LABEL: {{^}}v_fsub_v4f32:
43; R600: ADD {{\** *}}T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW], -T[0-9]+\.[XYZW]}}
44; R600: ADD {{\** *}}T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW], -T[0-9]+\.[XYZW]}}
45; R600: ADD {{\** *}}T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW], -T[0-9]+\.[XYZW]}}
46; R600: ADD {{\** *}}T{{[0-9]+\.[XYZW], T[0-9]+\.[XYZW], -T[0-9]+\.[XYZW]}}
47
48; SI: v_subrev_f32_e32 {{v[0-9]+}}, {{v[0-9]+}}, {{v[0-9]+}}
49; SI: v_subrev_f32_e32 {{v[0-9]+}}, {{v[0-9]+}}, {{v[0-9]+}}
50; SI: v_subrev_f32_e32 {{v[0-9]+}}, {{v[0-9]+}}, {{v[0-9]+}}
51; SI: v_subrev_f32_e32 {{v[0-9]+}}, {{v[0-9]+}}, {{v[0-9]+}}
52define void @v_fsub_v4f32(<4 x float> addrspace(1)* %out, <4 x float> addrspace(1)* %in) {
53  %b_ptr = getelementptr <4 x float>, <4 x float> addrspace(1)* %in, i32 1
54  %a = load <4 x float>, <4 x float> addrspace(1)* %in, align 16
55  %b = load <4 x float>, <4 x float> addrspace(1)* %b_ptr, align 16
56  %result = fsub <4 x float> %a, %b
57  store <4 x float> %result, <4 x float> addrspace(1)* %out, align 16
58  ret void
59}
60
61; FUNC-LABEL: {{^}}s_fsub_v4f32:
62; SI: v_subrev_f32_e32 {{v[0-9]+}}, {{s[0-9]+}}, {{v[0-9]+}}
63; SI: v_subrev_f32_e32 {{v[0-9]+}}, {{s[0-9]+}}, {{v[0-9]+}}
64; SI: v_subrev_f32_e32 {{v[0-9]+}}, {{s[0-9]+}}, {{v[0-9]+}}
65; SI: v_subrev_f32_e32 {{v[0-9]+}}, {{s[0-9]+}}, {{v[0-9]+}}
66; SI: s_endpgm
67define void @s_fsub_v4f32(<4 x float> addrspace(1)* %out, <4 x float> %a, <4 x float> %b) {
68  %result = fsub <4 x float> %a, %b
69  store <4 x float> %result, <4 x float> addrspace(1)* %out, align 16
70  ret void
71}
72