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
|
target datalayout = "e-p:64:64-p1:64:64-p2:32:32-p3:32:32-p4:64:64-p5:32:32-p6:32:32-i64:64-v16:16-v24:32-v32:32-v48:64-v96:128-v192:256-v256:256-v512:512-v1024:1024-v2048:2048-n32:64-S32-A5-G1-ni:7"
target triple = "amdgcn-amd-amdhsa"
define protected amdgpu_kernel void @div_approx(ptr addrspace(4) byref(i64) %"23", ptr addrspace(4) byref(i64) %"24") #0 {
"28":
%"8" = alloca i1, align 1, addrspace(5)
store i1 false, ptr addrspace(5) %"8", align 1
%"9" = alloca i1, align 1, addrspace(5)
store i1 false, ptr addrspace(5) %"9", align 1
%"4" = alloca i64, align 8, addrspace(5)
%"5" = alloca i64, align 8, addrspace(5)
%"6" = alloca float, align 4, addrspace(5)
%"7" = alloca float, align 4, addrspace(5)
%"10" = load i64, ptr addrspace(4) %"23", align 8
store i64 %"10", ptr addrspace(5) %"4", align 8
%"11" = load i64, ptr addrspace(4) %"24", align 8
store i64 %"11", ptr addrspace(5) %"5", align 8
%"13" = load i64, ptr addrspace(5) %"4", align 8
%"25" = inttoptr i64 %"13" to ptr
%"12" = load float, ptr %"25", align 4
store float %"12", ptr addrspace(5) %"6", align 4
%"15" = load i64, ptr addrspace(5) %"4", align 8
%"26" = inttoptr i64 %"15" to ptr
%"30" = getelementptr inbounds i8, ptr %"26", i64 4
%"14" = load float, ptr %"30", align 4
store float %"14", ptr addrspace(5) %"7", align 4
%"17" = load float, ptr addrspace(5) %"6", align 4
%"18" = load float, ptr addrspace(5) %"7", align 4
%"16" = fdiv arcp afn float %"17", %"18"
store float %"16", ptr addrspace(5) %"6", align 4
%"19" = load i64, ptr addrspace(5) %"5", align 8
%"20" = load float, ptr addrspace(5) %"6", align 4
%"27" = inttoptr i64 %"19" to ptr
store float %"20", ptr %"27", align 4
ret void
}
attributes #0 = { "amdgpu-unsafe-fp-atomics"="true" "denormal-fp-math"="ieee,ieee" "denormal-fp-math-f32"="ieee,ieee" "no-trapping-math"="true" "uniform-work-group-size"="true" }
|
