1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
|
; RUN: llc -mtriple=amdgcn < %s | FileCheck --check-prefixes=GCN,GCN-FMF,GCN-SAFE %s
declare i32 @llvm.amdgcn.workitem.id.x() nounwind readnone
; Test that the -enable-no-signed-zeros-fp-math flag works
; GCN-LABEL: {{^}}fneg_fsub_f32_fmf:
; GCN: v_sub_f32_e32 [[SUB:v[0-9]+]], {{v[0-9]+}}, {{v[0-9]+}}
; GCN-FMF-NOT: xor
define amdgpu_kernel void @fneg_fsub_f32_fmf(ptr addrspace(1) %out, ptr addrspace(1) %in) #0 {
%tid = call i32 @llvm.amdgcn.workitem.id.x()
%add = add i32 %tid, 1
%gep = getelementptr float, ptr addrspace(1) %in, i32 %tid
%b_ptr = getelementptr float, ptr addrspace(1) %in, i32 %add
%a = load float, ptr addrspace(1) %gep, align 4
%b = load float, ptr addrspace(1) %b_ptr, align 4
%result = fsub fast float %a, %b
%neg.result = fsub fast float -0.0, %result
store float %neg.result, ptr addrspace(1) %out, align 4
ret void
}
; GCN-LABEL: {{^}}fneg_fsub_f32_safe:
; GCN: v_sub_f32_e32 [[SUB:v[0-9]+]], {{v[0-9]+}}, {{v[0-9]+}}
; GCN-SAFE: v_xor_b32_e32 v{{[0-9]+}}, 0x80000000, [[SUB]]
define amdgpu_kernel void @fneg_fsub_f32_safe(ptr addrspace(1) %out, ptr addrspace(1) %in) #0 {
%tid = call i32 @llvm.amdgcn.workitem.id.x()
%add = add i32 %tid, 1
%gep = getelementptr float, ptr addrspace(1) %in, i32 %tid
%b_ptr = getelementptr float, ptr addrspace(1) %in, i32 %add
%a = load float, ptr addrspace(1) %gep, align 4
%b = load float, ptr addrspace(1) %b_ptr, align 4
%result = fsub float %a, %b
%neg.result = fsub float -0.0, %result
store float %neg.result, ptr addrspace(1) %out, align 4
ret void
}
attributes #0 = { nounwind }
|
