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
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
|
; RUN: llc -march=amdgcn -verify-machineinstrs -enable-misched=0 < %s | FileCheck -check-prefix=GCN -check-prefix=SI -check-prefix=FUNC %s
; RUN: llc -march=amdgcn -mcpu=tonga -mattr=-flat-for-global,-xnack -enable-misched=0 -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=VI -check-prefix=FUNC %s
; RUN: llc -march=r600 -mcpu=redwood < %s | FileCheck -check-prefix=R600 -check-prefix=FUNC %s
; DAGCombiner will transform:
; (fabsf (f32 bitcast (i32 a))) => (f32 bitcast (and (i32 a), 0x7FFFFFFF))
; unless isFabsFree returns true
; FUNC-LABEL: {{^}}s_fabsf_fn_free:
; R600-NOT: AND
; R600: |PV.{{[XYZW]}}|
; GCN: s_bitset0_b32 s{{[0-9]+}}, 31
define amdgpu_kernel void @s_fabsf_fn_free(ptr addrspace(1) %out, i32 %in) {
%bc= bitcast i32 %in to float
%fabs = call float @fabsf(float %bc)
store float %fabs, ptr addrspace(1) %out
ret void
}
; FUNC-LABEL: {{^}}s_fabsf_free:
; R600-NOT: AND
; R600: |PV.{{[XYZW]}}|
; GCN: s_bitset0_b32 s{{[0-9]+}}, 31
define amdgpu_kernel void @s_fabsf_free(ptr addrspace(1) %out, i32 %in) {
%bc= bitcast i32 %in to float
%fabs = call float @llvm.fabs.f32(float %bc)
store float %fabs, ptr addrspace(1) %out
ret void
}
; FUNC-LABEL: {{^}}s_fabsf_f32:
; R600: |{{(PV|T[0-9])\.[XYZW]}}|
; GCN: s_bitset0_b32 s{{[0-9]+}}, 31
define amdgpu_kernel void @s_fabsf_f32(ptr addrspace(1) %out, float %in) {
%fabs = call float @llvm.fabs.f32(float %in)
store float %fabs, ptr addrspace(1) %out
ret void
}
; FUNC-LABEL: {{^}}fabs_v2f32:
; R600: |{{(PV|T[0-9])\.[XYZW]}}|
; R600: |{{(PV|T[0-9])\.[XYZW]}}|
; GCN: s_and_b32 s{{[0-9]+}}, s{{[0-9]+}}, 0x7fffffff
; GCN: s_and_b32 s{{[0-9]+}}, s{{[0-9]+}}, 0x7fffffff
define amdgpu_kernel void @fabs_v2f32(ptr addrspace(1) %out, <2 x float> %in) {
%fabs = call <2 x float> @llvm.fabs.v2f32(<2 x float> %in)
store <2 x float> %fabs, ptr addrspace(1) %out
ret void
}
; FUNC-LABEL: {{^}}fabsf_v4f32:
; R600: |{{(PV|T[0-9])\.[XYZW]}}|
; R600: |{{(PV|T[0-9])\.[XYZW]}}|
; R600: |{{(PV|T[0-9])\.[XYZW]}}|
; R600: |{{(PV|T[0-9])\.[XYZW]}}|
; GCN: s_bitset0_b32
; GCN: s_bitset0_b32
; GCN: s_bitset0_b32
; GCN: s_bitset0_b32
define amdgpu_kernel void @fabsf_v4f32(ptr addrspace(1) %out, <4 x float> %in) {
%fabs = call <4 x float> @llvm.fabs.v4f32(<4 x float> %in)
store <4 x float> %fabs, ptr addrspace(1) %out
ret void
}
; GCN-LABEL: {{^}}fabsf_fn_fold:
; SI: s_load_dwordx4 s[[[#LOAD:]]:[[#END:]]], s[{{[0-9]+:[0-9]+}}], 0x9
; VI: s_load_dwordx4 s[[[#LOAD:]]:[[#END:]]], s[{{[0-9]+:[0-9]+}}], 0x24
; GCN-NOT: and
; GCN: v_mov_b32_e32 [[V_MUL_VI:v[0-9]+]], s[[#LOAD + 3]]
; GCN: v_mul_f32_e64 v{{[0-9]+}}, |s[[#LOAD + 2]]|, [[V_MUL_VI]]
define amdgpu_kernel void @fabsf_fn_fold(ptr addrspace(1) %out, float %in0, float %in1) {
%fabs = call float @fabsf(float %in0)
%fmul = fmul float %fabs, %in1
store float %fmul, ptr addrspace(1) %out
ret void
}
; FUNC-LABEL: {{^}}fabs_fold:
; SI: s_load_dwordx4 s[[[#LOAD:]]:[[#END:]]], s[{{[0-9]+:[0-9]+}}], 0x9
; VI: s_load_dwordx4 s[[[#LOAD:]]:[[#END:]]], s[{{[0-9]+:[0-9]+}}], 0x24
; GCN-NOT: and
; GCN: v_mov_b32_e32 [[V_MUL_VI:v[0-9]+]], s[[#LOAD + 3]]
; GCN: v_mul_f32_e64 v{{[0-9]+}}, |s[[#LOAD + 2]]|, [[V_MUL_VI]]
define amdgpu_kernel void @fabs_fold(ptr addrspace(1) %out, float %in0, float %in1) {
%fabs = call float @llvm.fabs.f32(float %in0)
%fmul = fmul float %fabs, %in1
store float %fmul, ptr addrspace(1) %out
ret void
}
; Make sure we turn some integer operations back into fabsf
; FUNC-LABEL: {{^}}bitpreserve_fabsf_f32:
; GCN: v_add_f32_e64 v{{[0-9]+}}, |s{{[0-9]+}}|, 1.0
define amdgpu_kernel void @bitpreserve_fabsf_f32(ptr addrspace(1) %out, float %in) {
%in.bc = bitcast float %in to i32
%int.abs = and i32 %in.bc, 2147483647
%bc = bitcast i32 %int.abs to float
%fadd = fadd float %bc, 1.0
store float %fadd, ptr addrspace(1) %out
ret void
}
declare float @fabsf(float) readnone
declare float @llvm.fabs.f32(float) readnone
declare <2 x float> @llvm.fabs.v2f32(<2 x float>) readnone
declare <4 x float> @llvm.fabs.v4f32(<4 x float>) readnone
|