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
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
|
; RUN: llc -march=amdgcn -verify-machineinstrs < %s | FileCheck -check-prefix=SI-SAFE -check-prefix=SI -check-prefix=FUNC %s
; RUN: llc -march=amdgcn -enable-no-nans-fp-math -verify-machineinstrs < %s | FileCheck -check-prefix=SI-NONAN -check-prefix=SI -check-prefix=FUNC %s
; RUN: llc -march=amdgcn -mcpu=tonga -verify-machineinstrs < %s | FileCheck -check-prefix=SI -check-prefix=FUNC %s
declare float @llvm.fabs.f32(float) #1
declare float @llvm.floor.f32(float) #1
; FUNC-LABEL: {{^}}cvt_flr_i32_f32_0:
; SI-SAFE-NOT: v_cvt_flr_i32_f32
; SI-NOT: add
; SI-NONAN: v_cvt_flr_i32_f32_e32 v{{[0-9]+}}, s{{[0-9]+}}
; SI: s_endpgm
define amdgpu_kernel void @cvt_flr_i32_f32_0(ptr addrspace(1) %out, float %x) #0 {
%floor = call float @llvm.floor.f32(float %x) #1
%cvt = fptosi float %floor to i32
store i32 %cvt, ptr addrspace(1) %out
ret void
}
; FUNC-LABEL: {{^}}cvt_flr_i32_f32_1:
; SI: v_add_f32_e64 [[TMP:v[0-9]+]], s{{[0-9]+}}, 1.0
; SI-SAFE-NOT: v_cvt_flr_i32_f32
; SI-NONAN: v_cvt_flr_i32_f32_e32 v{{[0-9]+}}, [[TMP]]
; SI: s_endpgm
define amdgpu_kernel void @cvt_flr_i32_f32_1(ptr addrspace(1) %out, float %x) #0 {
%fadd = fadd float %x, 1.0
%floor = call float @llvm.floor.f32(float %fadd) #1
%cvt = fptosi float %floor to i32
store i32 %cvt, ptr addrspace(1) %out
ret void
}
; FUNC-LABEL: {{^}}cvt_flr_i32_f32_fabs:
; SI-NOT: add
; SI-SAFE-NOT: v_cvt_flr_i32_f32
; SI-NONAN: v_cvt_flr_i32_f32_e64 v{{[0-9]+}}, |s{{[0-9]+}}|
; SI: s_endpgm
define amdgpu_kernel void @cvt_flr_i32_f32_fabs(ptr addrspace(1) %out, float %x) #0 {
%x.fabs = call float @llvm.fabs.f32(float %x) #1
%floor = call float @llvm.floor.f32(float %x.fabs) #1
%cvt = fptosi float %floor to i32
store i32 %cvt, ptr addrspace(1) %out
ret void
}
; FUNC-LABEL: {{^}}cvt_flr_i32_f32_fneg:
; SI-NOT: add
; SI-SAFE-NOT: v_cvt_flr_i32_f32
; SI-NONAN: v_cvt_flr_i32_f32_e64 v{{[0-9]+}}, -s{{[0-9]+}}
; SI: s_endpgm
define amdgpu_kernel void @cvt_flr_i32_f32_fneg(ptr addrspace(1) %out, float %x) #0 {
%x.fneg = fsub float -0.000000e+00, %x
%floor = call float @llvm.floor.f32(float %x.fneg) #1
%cvt = fptosi float %floor to i32
store i32 %cvt, ptr addrspace(1) %out
ret void
}
; FUNC-LABEL: {{^}}cvt_flr_i32_f32_fabs_fneg:
; SI-NOT: add
; SI-SAFE-NOT: v_cvt_flr_i32_f32
; SI-NONAN: v_cvt_flr_i32_f32_e64 v{{[0-9]+}}, -|s{{[0-9]+}}|
; SI: s_endpgm
define amdgpu_kernel void @cvt_flr_i32_f32_fabs_fneg(ptr addrspace(1) %out, float %x) #0 {
%x.fabs = call float @llvm.fabs.f32(float %x) #1
%x.fabs.fneg = fsub float -0.000000e+00, %x.fabs
%floor = call float @llvm.floor.f32(float %x.fabs.fneg) #1
%cvt = fptosi float %floor to i32
store i32 %cvt, ptr addrspace(1) %out
ret void
}
; FUNC-LABEL: {{^}}no_cvt_flr_i32_f32_0:
; SI-NOT: v_cvt_flr_i32_f32
; SI: v_floor_f32
; SI: v_cvt_u32_f32_e32
; SI: s_endpgm
define amdgpu_kernel void @no_cvt_flr_i32_f32_0(ptr addrspace(1) %out, float %x) #0 {
%floor = call float @llvm.floor.f32(float %x) #1
%cvt = fptoui float %floor to i32
store i32 %cvt, ptr addrspace(1) %out
ret void
}
attributes #0 = { nounwind }
attributes #1 = { nounwind readnone }
|
