fma.f64.ll
9.72 KB
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
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
; RUN: llc -amdgpu-scalarize-global-loads=false -march=amdgcn -verify-machineinstrs < %s | FileCheck -check-prefix=SI -check-prefix=FUNC %s
; RUN: llc -amdgpu-scalarize-global-loads=false -march=amdgcn -mcpu=tonga -mattr=-flat-for-global -verify-machineinstrs < %s | FileCheck -check-prefix=SI -check-prefix=FUNC %s
declare double @llvm.fma.f64(double, double, double) nounwind readnone
declare <2 x double> @llvm.fma.v2f64(<2 x double>, <2 x double>, <2 x double>) nounwind readnone
declare <4 x double> @llvm.fma.v4f64(<4 x double>, <4 x double>, <4 x double>) nounwind readnone
declare double @llvm.fabs.f64(double) nounwind readnone
; FUNC-LABEL: {{^}}fma_f64:
; SI: v_fma_f64 {{v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\]}}
define amdgpu_kernel void @fma_f64(double addrspace(1)* %out, double addrspace(1)* %in1,
double addrspace(1)* %in2, double addrspace(1)* %in3) {
%r0 = load double, double addrspace(1)* %in1
%r1 = load double, double addrspace(1)* %in2
%r2 = load double, double addrspace(1)* %in3
%r3 = tail call double @llvm.fma.f64(double %r0, double %r1, double %r2)
store double %r3, double addrspace(1)* %out
ret void
}
; FUNC-LABEL: {{^}}fma_v2f64:
; SI: v_fma_f64
; SI: v_fma_f64
define amdgpu_kernel void @fma_v2f64(<2 x double> addrspace(1)* %out, <2 x double> addrspace(1)* %in1,
<2 x double> addrspace(1)* %in2, <2 x double> addrspace(1)* %in3) {
%r0 = load <2 x double>, <2 x double> addrspace(1)* %in1
%r1 = load <2 x double>, <2 x double> addrspace(1)* %in2
%r2 = load <2 x double>, <2 x double> addrspace(1)* %in3
%r3 = tail call <2 x double> @llvm.fma.v2f64(<2 x double> %r0, <2 x double> %r1, <2 x double> %r2)
store <2 x double> %r3, <2 x double> addrspace(1)* %out
ret void
}
; FUNC-LABEL: {{^}}fma_v4f64:
; SI: v_fma_f64
; SI: v_fma_f64
; SI: v_fma_f64
; SI: v_fma_f64
define amdgpu_kernel void @fma_v4f64(<4 x double> addrspace(1)* %out, <4 x double> addrspace(1)* %in1,
<4 x double> addrspace(1)* %in2, <4 x double> addrspace(1)* %in3) {
%r0 = load <4 x double>, <4 x double> addrspace(1)* %in1
%r1 = load <4 x double>, <4 x double> addrspace(1)* %in2
%r2 = load <4 x double>, <4 x double> addrspace(1)* %in3
%r3 = tail call <4 x double> @llvm.fma.v4f64(<4 x double> %r0, <4 x double> %r1, <4 x double> %r2)
store <4 x double> %r3, <4 x double> addrspace(1)* %out
ret void
}
; FUNC-LABEL: {{^}}fma_f64_abs_src0:
; SI: v_fma_f64 {{v\[[0-9]+:[0-9]+\], |v\[[0-9]+:[0-9]+\]|, v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\]}}
define amdgpu_kernel void @fma_f64_abs_src0(double addrspace(1)* %out, double addrspace(1)* %in1,
double addrspace(1)* %in2, double addrspace(1)* %in3) {
%r0 = load double, double addrspace(1)* %in1
%r1 = load double, double addrspace(1)* %in2
%r2 = load double, double addrspace(1)* %in3
%fabs = call double @llvm.fabs.f64(double %r0)
%r3 = tail call double @llvm.fma.f64(double %fabs, double %r1, double %r2)
store double %r3, double addrspace(1)* %out
ret void
}
; FUNC-LABEL: {{^}}fma_f64_abs_src1:
; SI: v_fma_f64 {{v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\], \|v\[[0-9]+:[0-9]+\]\|, v\[[0-9]+:[0-9]+\]}}
define amdgpu_kernel void @fma_f64_abs_src1(double addrspace(1)* %out, double addrspace(1)* %in1,
double addrspace(1)* %in2, double addrspace(1)* %in3) {
%r0 = load double, double addrspace(1)* %in1
%r1 = load double, double addrspace(1)* %in2
%r2 = load double, double addrspace(1)* %in3
%fabs = call double @llvm.fabs.f64(double %r1)
%r3 = tail call double @llvm.fma.f64(double %r0, double %fabs, double %r2)
store double %r3, double addrspace(1)* %out
ret void
}
; FUNC-LABEL: {{^}}fma_f64_abs_src2:
; SI: v_fma_f64 {{v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\], \|v\[[0-9]+:[0-9]+\]\|}}
define amdgpu_kernel void @fma_f64_abs_src2(double addrspace(1)* %out, double addrspace(1)* %in1,
double addrspace(1)* %in2, double addrspace(1)* %in3) {
%r0 = load double, double addrspace(1)* %in1
%r1 = load double, double addrspace(1)* %in2
%r2 = load double, double addrspace(1)* %in3
%fabs = call double @llvm.fabs.f64(double %r2)
%r3 = tail call double @llvm.fma.f64(double %r0, double %r1, double %fabs)
store double %r3, double addrspace(1)* %out
ret void
}
; FUNC-LABEL: {{^}}fma_f64_neg_src0:
; SI: v_fma_f64 {{v\[[0-9]+:[0-9]+\], -v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\]}}
define amdgpu_kernel void @fma_f64_neg_src0(double addrspace(1)* %out, double addrspace(1)* %in1,
double addrspace(1)* %in2, double addrspace(1)* %in3) {
%r0 = load double, double addrspace(1)* %in1
%r1 = load double, double addrspace(1)* %in2
%r2 = load double, double addrspace(1)* %in3
%fsub = fsub double -0.000000e+00, %r0
%r3 = tail call double @llvm.fma.f64(double %fsub, double %r1, double %r2)
store double %r3, double addrspace(1)* %out
ret void
}
; FUNC-LABEL: {{^}}fma_f64_neg_src1:
; SI: v_fma_f64 {{v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\], -v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\]}}
define amdgpu_kernel void @fma_f64_neg_src1(double addrspace(1)* %out, double addrspace(1)* %in1,
double addrspace(1)* %in2, double addrspace(1)* %in3) {
%r0 = load double, double addrspace(1)* %in1
%r1 = load double, double addrspace(1)* %in2
%r2 = load double, double addrspace(1)* %in3
%fsub = fsub double -0.000000e+00, %r1
%r3 = tail call double @llvm.fma.f64(double %r0, double %fsub, double %r2)
store double %r3, double addrspace(1)* %out
ret void
}
; FUNC-LABEL: {{^}}fma_f64_neg_src2:
; SI: v_fma_f64 {{v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\], -v\[[0-9]+:[0-9]+\]}}
define amdgpu_kernel void @fma_f64_neg_src2(double addrspace(1)* %out, double addrspace(1)* %in1,
double addrspace(1)* %in2, double addrspace(1)* %in3) {
%r0 = load double, double addrspace(1)* %in1
%r1 = load double, double addrspace(1)* %in2
%r2 = load double, double addrspace(1)* %in3
%fsub = fsub double -0.000000e+00, %r2
%r3 = tail call double @llvm.fma.f64(double %r0, double %r1, double %fsub)
store double %r3, double addrspace(1)* %out
ret void
}
; FUNC-LABEL: {{^}}fma_f64_abs_neg_src0:
; SI: v_fma_f64 {{v\[[0-9]+:[0-9]+\], -\|v\[[0-9]+:[0-9]+\]\|, v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\]}}
define amdgpu_kernel void @fma_f64_abs_neg_src0(double addrspace(1)* %out, double addrspace(1)* %in1,
double addrspace(1)* %in2, double addrspace(1)* %in3) {
%r0 = load double, double addrspace(1)* %in1
%r1 = load double, double addrspace(1)* %in2
%r2 = load double, double addrspace(1)* %in3
%fabs = call double @llvm.fabs.f64(double %r0)
%fsub = fsub double -0.000000e+00, %fabs
%r3 = tail call double @llvm.fma.f64(double %fsub, double %r1, double %r2)
store double %r3, double addrspace(1)* %out
ret void
}
; FUNC-LABEL: {{^}}fma_f64_abs_neg_src1:
; SI: v_fma_f64 {{v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\], -\|v\[[0-9]+:[0-9]+\]\|, v\[[0-9]+:[0-9]+\]}}
define amdgpu_kernel void @fma_f64_abs_neg_src1(double addrspace(1)* %out, double addrspace(1)* %in1,
double addrspace(1)* %in2, double addrspace(1)* %in3) {
%r0 = load double, double addrspace(1)* %in1
%r1 = load double, double addrspace(1)* %in2
%r2 = load double, double addrspace(1)* %in3
%fabs = call double @llvm.fabs.f64(double %r1)
%fsub = fsub double -0.000000e+00, %fabs
%r3 = tail call double @llvm.fma.f64(double %r0, double %fsub, double %r2)
store double %r3, double addrspace(1)* %out
ret void
}
; FUNC-LABEL: {{^}}fma_f64_abs_neg_src2:
; SI: v_fma_f64 {{v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\], -\|v\[[0-9]+:[0-9]+\]\|}}
define amdgpu_kernel void @fma_f64_abs_neg_src2(double addrspace(1)* %out, double addrspace(1)* %in1,
double addrspace(1)* %in2, double addrspace(1)* %in3) {
%r0 = load double, double addrspace(1)* %in1
%r1 = load double, double addrspace(1)* %in2
%r2 = load double, double addrspace(1)* %in3
%fabs = call double @llvm.fabs.f64(double %r2)
%fsub = fsub double -0.000000e+00, %fabs
%r3 = tail call double @llvm.fma.f64(double %r0, double %r1, double %fsub)
store double %r3, double addrspace(1)* %out
ret void
}
; FUNC-LABEL: {{^}}fma_f64_lit_src0:
; SI: v_fma_f64 {{v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\], 2.0, v\[[0-9]+:[0-9]+\]}}
define amdgpu_kernel void @fma_f64_lit_src0(double addrspace(1)* %out,
double addrspace(1)* %in2, double addrspace(1)* %in3) {
%r1 = load double, double addrspace(1)* %in2
%r2 = load double, double addrspace(1)* %in3
%r3 = tail call double @llvm.fma.f64(double +2.0, double %r1, double %r2)
store double %r3, double addrspace(1)* %out
ret void
}
; FUNC-LABEL: {{^}}fma_f64_lit_src1:
; SI: v_fma_f64 {{v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\], 2.0, v\[[0-9]+:[0-9]+\]}}
define amdgpu_kernel void @fma_f64_lit_src1(double addrspace(1)* %out, double addrspace(1)* %in1,
double addrspace(1)* %in3) {
%r0 = load double, double addrspace(1)* %in1
%r2 = load double, double addrspace(1)* %in3
%r3 = tail call double @llvm.fma.f64(double %r0, double +2.0, double %r2)
store double %r3, double addrspace(1)* %out
ret void
}
; FUNC-LABEL: {{^}}fma_f64_lit_src2:
; SI: v_fma_f64 {{v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\], v\[[0-9]+:[0-9]+\], 2.0}}
define amdgpu_kernel void @fma_f64_lit_src2(double addrspace(1)* %out, double addrspace(1)* %in1,
double addrspace(1)* %in2) {
%r0 = load double, double addrspace(1)* %in1
%r1 = load double, double addrspace(1)* %in2
%r3 = tail call double @llvm.fma.f64(double %r0, double %r1, double +2.0)
store double %r3, double addrspace(1)* %out
ret void
}