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
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
|
; RUN: llc -mtriple=amdgcn--amdpal -mcpu=tahiti -verify-machineinstrs < %s | FileCheck -check-prefix=SI %s
; RUN: llc -mtriple=amdgcn--amdpal -mcpu=tonga -mattr=-flat-for-global,-xnack -verify-machineinstrs < %s | FileCheck -check-prefix=SI %s
; SI-LABEL: {{^}}s_mulk_i32_k0:
; SI: s_load_dword [[VAL:s[0-9]+]]
; SI: s_mulk_i32 [[VAL]], 0x41
; SI: v_mov_b32_e32 [[VRESULT:v[0-9]+]], [[VAL]]
; SI: buffer_store_dword [[VRESULT]]
; SI: s_endpgm
define amdgpu_kernel void @s_mulk_i32_k0(ptr addrspace(1) %out, i32 %b) {
%mul = mul i32 %b, 65
store i32 %mul, ptr addrspace(1) %out
ret void
}
; SI-LABEL: {{^}}s_mulk_i32_k1:
; SI: s_mulk_i32 {{s[0-9]+}}, 0x7fff{{$}}
; SI: s_endpgm
define amdgpu_kernel void @s_mulk_i32_k1(ptr addrspace(1) %out, i32 %b) {
%mul = mul i32 %b, 32767 ; (1 << 15) - 1
store i32 %mul, ptr addrspace(1) %out
ret void
}
; SI-LABEL: {{^}}s_mulk_i32_k2:
; SI: s_mulk_i32 {{s[0-9]+}}, 0xffef{{$}}
; SI: s_endpgm
define amdgpu_kernel void @s_mulk_i32_k2(ptr addrspace(1) %out, i32 %b) {
%mul = mul i32 %b, -17
store i32 %mul, ptr addrspace(1) %out
ret void
}
; SI-LABEL: {{^}}no_s_mulk_i32_k0:
; SI: s_mul_i32 {{s[0-9]+}}, {{s[0-9]+}}, 0x8001{{$}}
; SI: s_endpgm
define amdgpu_kernel void @no_s_mulk_i32_k0(ptr addrspace(1) %out, i32 %b) {
%mul = mul i32 %b, 32769 ; 1 << 15 + 1
store i32 %mul, ptr addrspace(1) %out
ret void
}
@lds = addrspace(3) global [512 x i32] undef, align 4
; SI-LABEL: {{^}}commute_s_mulk_i32:
; SI: s_mulk_i32 s{{[0-9]+}}, 0x800{{$}}
define amdgpu_kernel void @commute_s_mulk_i32(ptr addrspace(1) %out, i32 %b) #0 {
%size = call i32 @llvm.amdgcn.groupstaticsize()
%add = mul i32 %size, %b
call void asm sideeffect "; foo $0, $1", "v,s"(ptr addrspace(3) @lds, i32 %add)
ret void
}
declare i32 @llvm.amdgcn.groupstaticsize() #1
attributes #0 = { nounwind }
attributes #1 = { nounwind readnone }
|
