1 ; RUN: llc -march=amdgcn -mcpu=verde -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=VCCZ-BUG %s
2 ; RUN: llc -march=amdgcn -mcpu=bonaire -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=VCCZ-BUG %s
3 ; RUN: llc -march=amdgcn -mcpu=tonga -mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=NOVCCZ-BUG %s
5 ; GCN-FUNC: {{^}}vccz_workaround:
6 ; GCN: s_load_dword s{{[0-9]+}}, s[{{[0-9]+:[0-9]+}}], 0x0
7 ; GCN: v_cmp_neq_f32_e64 {{[^,]*}}, s{{[0-9]+}}, 0{{$}}
8 ; VCCZ-BUG: s_waitcnt lgkmcnt(0)
9 ; VCCZ-BUG: s_mov_b64 vcc, vcc
10 ; NOVCCZ-BUG-NOT: s_mov_b64 vcc, vcc
11 ; GCN: s_cbranch_vccnz [[EXIT:[0-9A-Za-z_]+]]
12 ; GCN: buffer_store_dword
15 define amdgpu_kernel void @vccz_workaround(i32 addrspace(4)* %in, i32 addrspace(1)* %out, float %cond) {
17 %cnd = fcmp oeq float 0.0, %cond
18 %sgpr = load volatile i32, i32 addrspace(4)* %in
19 br i1 %cnd, label %if, label %endif
22 store i32 %sgpr, i32 addrspace(1)* %out
29 ; GCN-FUNC: {{^}}vccz_noworkaround:
30 ; GCN: v_cmp_neq_f32_e32 vcc, 0, v{{[0-9]+}}
31 ; GCN: s_cbranch_vccnz [[EXIT:[0-9A-Za-z_]+]]
32 ; GCN: buffer_store_dword
35 define amdgpu_kernel void @vccz_noworkaround(float addrspace(1)* %in, float addrspace(1)* %out) {
37 %vgpr = load volatile float, float addrspace(1)* %in
38 %cnd = fcmp oeq float 0.0, %vgpr
39 br i1 %cnd, label %if, label %endif
42 store float %vgpr, float addrspace(1)* %out