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
|
// RUN: %clang_cc1 -finclude-default-header -fnative-half-type -triple \
// RUN: dxil-pc-shadermodel6.3-compute %s -emit-llvm -o - | \
// RUN: FileCheck %s --check-prefixes=CHECK,CHECK-DXIL
// RUN: %clang_cc1 -finclude-default-header -fnative-half-type -triple \
// RUN: spirv-pc-vulkan-compute %s -emit-llvm -o - | \
// RUN: FileCheck %s --check-prefixes=CHECK,CHECK-SPIRV
// Test basic lowering to runtime function call.
// CHECK-LABEL: define {{.*}}test_default_parameter_type
float test_default_parameter_type(half2 p1, half2 p2, float p3) {
// CHECK-SPIRV: %[[MUL:.*]] = call reassoc nnan ninf nsz arcp afn half @llvm.spv.fdot.v2f16(<2 x half> %{{.*}}, <2 x half> %{{.*}})
// CHECK-SPIRV: %[[CONV:.*]] = fpext reassoc nnan ninf nsz arcp afn half %[[MUL]] to float
// CHECK-SPIRV: %[[C:.*]] = load float, ptr %c.addr.i, align 4
// CHECK-SPIRV: %[[RES:.*]] = fadd reassoc nnan ninf nsz arcp afn float %[[CONV]], %[[C]]
// CHECK-DXIL: %[[AX:.*]] = extractelement <2 x half> %{{.*}}, i32 0
// CHECK-DXIL: %[[AY:.*]] = extractelement <2 x half> %{{.*}}, i32 1
// CHECK-DXIL: %[[BX:.*]] = extractelement <2 x half> %{{.*}}, i32 0
// CHECK-DXIL: %[[BY:.*]] = extractelement <2 x half> %{{.*}}, i32 1
// CHECK-DXIL: %[[RES:.*]] = call {{.*}} float @llvm.dx.dot2add(float %{{.*}}, half %[[AX]], half %[[AY]], half %[[BX]], half %[[BY]])
// CHECK: ret float %[[RES]]
return dot2add(p1, p2, p3);
}
// CHECK-LABEL: define {{.*}}test_float_arg2_type
float test_float_arg2_type(half2 p1, float2 p2, float p3) {
// CHECK: %conv = fptrunc reassoc nnan ninf nsz arcp afn <2 x float> %{{.*}} to <2 x half>
// CHECK-SPIRV: %[[MUL:.*]] = call reassoc nnan ninf nsz arcp afn half @llvm.spv.fdot.v2f16(<2 x half> %{{.*}}, <2 x half> %{{.*}})
// CHECK-SPIRV: %[[CONV:.*]] = fpext reassoc nnan ninf nsz arcp afn half %[[MUL]] to float
// CHECK-SPIRV: %[[C:.*]] = load float, ptr %c.addr.i, align 4
// CHECK-SPIRV: %[[RES:.*]] = fadd reassoc nnan ninf nsz arcp afn float %[[CONV]], %[[C]]
// CHECK-DXIL: %[[AX:.*]] = extractelement <2 x half> %{{.*}}, i32 0
// CHECK-DXIL: %[[AY:.*]] = extractelement <2 x half> %{{.*}}, i32 1
// CHECK-DXIL: %[[BX:.*]] = extractelement <2 x half> %{{.*}}, i32 0
// CHECK-DXIL: %[[BY:.*]] = extractelement <2 x half> %{{.*}}, i32 1
// CHECK-DXIL: %[[RES:.*]] = call {{.*}} float @llvm.dx.dot2add(float %{{.*}}, half %[[AX]], half %[[AY]], half %[[BX]], half %[[BY]])
// CHECK: ret float %[[RES]]
return dot2add(p1, p2, p3);
}
// CHECK-LABEL: define {{.*}}test_float_arg1_type
float test_float_arg1_type(float2 p1, half2 p2, float p3) {
// CHECK: %conv = fptrunc reassoc nnan ninf nsz arcp afn <2 x float> %{{.*}} to <2 x half>
// CHECK-SPIRV: %[[MUL:.*]] = call reassoc nnan ninf nsz arcp afn half @llvm.spv.fdot.v2f16(<2 x half> %{{.*}}, <2 x half> %{{.*}})
// CHECK-SPIRV: %[[CONV:.*]] = fpext reassoc nnan ninf nsz arcp afn half %[[MUL]] to float
// CHECK-SPIRV: %[[C:.*]] = load float, ptr %c.addr.i, align 4
// CHECK-SPIRV: %[[RES:.*]] = fadd reassoc nnan ninf nsz arcp afn float %[[CONV]], %[[C]]
// CHECK-DXIL: %[[AX:.*]] = extractelement <2 x half> %{{.*}}, i32 0
// CHECK-DXIL: %[[AY:.*]] = extractelement <2 x half> %{{.*}}, i32 1
// CHECK-DXIL: %[[BX:.*]] = extractelement <2 x half> %{{.*}}, i32 0
// CHECK-DXIL: %[[BY:.*]] = extractelement <2 x half> %{{.*}}, i32 1
// CHECK-DXIL: %[[RES:.*]] = call {{.*}} float @llvm.dx.dot2add(float %{{.*}}, half %[[AX]], half %[[AY]], half %[[BX]], half %[[BY]])
// CHECK: ret float %[[RES]]
return dot2add(p1, p2, p3);
}
// CHECK-LABEL: define {{.*}}test_double_arg3_type
float test_double_arg3_type(half2 p1, half2 p2, double p3) {
// CHECK: %conv = fptrunc reassoc nnan ninf nsz arcp afn double %{{.*}} to float
// CHECK-SPIRV: %[[MUL:.*]] = call reassoc nnan ninf nsz arcp afn half @llvm.spv.fdot.v2f16(<2 x half> %{{.*}}, <2 x half> %{{.*}})
// CHECK-SPIRV: %[[CONV:.*]] = fpext reassoc nnan ninf nsz arcp afn half %[[MUL]] to float
// CHECK-SPIRV: %[[C:.*]] = load float, ptr %c.addr.i, align 4
// CHECK-SPIRV: %[[RES:.*]] = fadd reassoc nnan ninf nsz arcp afn float %[[CONV]], %[[C]]
// CHECK-DXIL: %[[AX:.*]] = extractelement <2 x half> %{{.*}}, i32 0
// CHECK-DXIL: %[[AY:.*]] = extractelement <2 x half> %{{.*}}, i32 1
// CHECK-DXIL: %[[BX:.*]] = extractelement <2 x half> %{{.*}}, i32 0
// CHECK-DXIL: %[[BY:.*]] = extractelement <2 x half> %{{.*}}, i32 1
// CHECK-DXIL: %[[RES:.*]] = call {{.*}} float @llvm.dx.dot2add(float %{{.*}}, half %[[AX]], half %[[AY]], half %[[BX]], half %[[BY]])
// CHECK: ret float %[[RES]]
return dot2add(p1, p2, p3);
}
// CHECK-LABEL: define {{.*}}test_float_arg1_arg2_type
float test_float_arg1_arg2_type(float2 p1, float2 p2, float p3) {
// CHECK: %conv = fptrunc reassoc nnan ninf nsz arcp afn <2 x float> %{{.*}} to <2 x half>
// CHECK: %conv1 = fptrunc reassoc nnan ninf nsz arcp afn <2 x float> %{{.*}} to <2 x half>
// CHECK-SPIRV: %[[MUL:.*]] = call reassoc nnan ninf nsz arcp afn half @llvm.spv.fdot.v2f16(<2 x half> %{{.*}}, <2 x half> %{{.*}})
// CHECK-SPIRV: %[[CONV:.*]] = fpext reassoc nnan ninf nsz arcp afn half %[[MUL]] to float
// CHECK-SPIRV: %[[C:.*]] = load float, ptr %c.addr.i, align 4
// CHECK-SPIRV: %[[RES:.*]] = fadd reassoc nnan ninf nsz arcp afn float %[[CONV]], %[[C]]
// CHECK-DXIL: %[[AX:.*]] = extractelement <2 x half> %{{.*}}, i32 0
// CHECK-DXIL: %[[AY:.*]] = extractelement <2 x half> %{{.*}}, i32 1
// CHECK-DXIL: %[[BX:.*]] = extractelement <2 x half> %{{.*}}, i32 0
// CHECK-DXIL: %[[BY:.*]] = extractelement <2 x half> %{{.*}}, i32 1
// CHECK-DXIL: %[[RES:.*]] = call {{.*}} float @llvm.dx.dot2add(float %{{.*}}, half %[[AX]], half %[[AY]], half %[[BX]], half %[[BY]])
// CHECK: ret float %[[RES]]
return dot2add(p1, p2, p3);
}
// CHECK-LABEL: define {{.*}}test_double_arg1_arg2_type
float test_double_arg1_arg2_type(double2 p1, double2 p2, float p3) {
// CHECK: %conv = fptrunc reassoc nnan ninf nsz arcp afn <2 x double> %{{.*}} to <2 x half>
// CHECK: %conv1 = fptrunc reassoc nnan ninf nsz arcp afn <2 x double> %{{.*}} to <2 x half>
// CHECK-SPIRV: %[[MUL:.*]] = call reassoc nnan ninf nsz arcp afn half @llvm.spv.fdot.v2f16(<2 x half> %{{.*}}, <2 x half> %{{.*}})
// CHECK-SPIRV: %[[CONV:.*]] = fpext reassoc nnan ninf nsz arcp afn half %[[MUL]] to float
// CHECK-SPIRV: %[[C:.*]] = load float, ptr %c.addr.i, align 4
// CHECK-SPIRV: %[[RES:.*]] = fadd reassoc nnan ninf nsz arcp afn float %[[CONV]], %[[C]]
// CHECK-DXIL: %[[AX:.*]] = extractelement <2 x half> %{{.*}}, i32 0
// CHECK-DXIL: %[[AY:.*]] = extractelement <2 x half> %{{.*}}, i32 1
// CHECK-DXIL: %[[BX:.*]] = extractelement <2 x half> %{{.*}}, i32 0
// CHECK-DXIL: %[[BY:.*]] = extractelement <2 x half> %{{.*}}, i32 1
// CHECK-DXIL: %[[RES:.*]] = call {{.*}} float @llvm.dx.dot2add(float %{{.*}}, half %[[AX]], half %[[AY]], half %[[BX]], half %[[BY]])
// CHECK: ret float %[[RES]]
return dot2add(p1, p2, p3);
}
// CHECK-LABEL: define {{.*}}test_int16_arg1_arg2_type
float test_int16_arg1_arg2_type(int16_t2 p1, int16_t2 p2, float p3) {
// CHECK: %conv = sitofp <2 x i16> %{{.*}} to <2 x half>
// CHECK: %conv1 = sitofp <2 x i16> %{{.*}} to <2 x half>
// CHECK-SPIRV: %[[MUL:.*]] = call reassoc nnan ninf nsz arcp afn half @llvm.spv.fdot.v2f16(<2 x half> %{{.*}}, <2 x half> %{{.*}})
// CHECK-SPIRV: %[[CONV:.*]] = fpext reassoc nnan ninf nsz arcp afn half %[[MUL]] to float
// CHECK-SPIRV: %[[C:.*]] = load float, ptr %c.addr.i, align 4
// CHECK-SPIRV: %[[RES:.*]] = fadd reassoc nnan ninf nsz arcp afn float %[[CONV]], %[[C]]
// CHECK-DXIL: %[[AX:.*]] = extractelement <2 x half> %{{.*}}, i32 0
// CHECK-DXIL: %[[AY:.*]] = extractelement <2 x half> %{{.*}}, i32 1
// CHECK-DXIL: %[[BX:.*]] = extractelement <2 x half> %{{.*}}, i32 0
// CHECK-DXIL: %[[BY:.*]] = extractelement <2 x half> %{{.*}}, i32 1
// CHECK-DXIL: %[[RES:.*]] = call {{.*}} float @llvm.dx.dot2add(float %{{.*}}, half %[[AX]], half %[[AY]], half %[[BX]], half %[[BY]])
// CHECK: ret float %[[RES]]
return dot2add(p1, p2, p3);
}
// CHECK-LABEL: define {{.*}}test_int32_arg1_arg2_type
float test_int32_arg1_arg2_type(int32_t2 p1, int32_t2 p2, float p3) {
// CHECK: %conv = sitofp <2 x i32> %{{.*}} to <2 x half>
// CHECK: %conv1 = sitofp <2 x i32> %{{.*}} to <2 x half>
// CHECK-SPIRV: %[[MUL:.*]] = call reassoc nnan ninf nsz arcp afn half @llvm.spv.fdot.v2f16(<2 x half> %{{.*}}, <2 x half> %{{.*}})
// CHECK-SPIRV: %[[CONV:.*]] = fpext reassoc nnan ninf nsz arcp afn half %[[MUL]] to float
// CHECK-SPIRV: %[[C:.*]] = load float, ptr %c.addr.i, align 4
// CHECK-SPIRV: %[[RES:.*]] = fadd reassoc nnan ninf nsz arcp afn float %[[CONV]], %[[C]]
// CHECK-DXIL: %[[AX:.*]] = extractelement <2 x half> %{{.*}}, i32 0
// CHECK-DXIL: %[[AY:.*]] = extractelement <2 x half> %{{.*}}, i32 1
// CHECK-DXIL: %[[BX:.*]] = extractelement <2 x half> %{{.*}}, i32 0
// CHECK-DXIL: %[[BY:.*]] = extractelement <2 x half> %{{.*}}, i32 1
// CHECK-DXIL: %[[RES:.*]] = call {{.*}} float @llvm.dx.dot2add(float %{{.*}}, half %[[AX]], half %[[AY]], half %[[BX]], half %[[BY]])
// CHECK: ret float %[[RES]]
return dot2add(p1, p2, p3);
}
// CHECK-LABEL: define {{.*}}test_int64_arg1_arg2_type
float test_int64_arg1_arg2_type(int64_t2 p1, int64_t2 p2, float p3) {
// CHECK: %conv = sitofp <2 x i64> %{{.*}} to <2 x half>
// CHECK: %conv1 = sitofp <2 x i64> %{{.*}} to <2 x half>
// CHECK-SPIRV: %[[MUL:.*]] = call reassoc nnan ninf nsz arcp afn half @llvm.spv.fdot.v2f16(<2 x half> %{{.*}}, <2 x half> %{{.*}})
// CHECK-SPIRV: %[[CONV:.*]] = fpext reassoc nnan ninf nsz arcp afn half %[[MUL]] to float
// CHECK-SPIRV: %[[C:.*]] = load float, ptr %c.addr.i, align 4
// CHECK-SPIRV: %[[RES:.*]] = fadd reassoc nnan ninf nsz arcp afn float %[[CONV]], %[[C]]
// CHECK-DXIL: %[[AX:.*]] = extractelement <2 x half> %{{.*}}, i32 0
// CHECK-DXIL: %[[AY:.*]] = extractelement <2 x half> %{{.*}}, i32 1
// CHECK-DXIL: %[[BX:.*]] = extractelement <2 x half> %{{.*}}, i32 0
// CHECK-DXIL: %[[BY:.*]] = extractelement <2 x half> %{{.*}}, i32 1
// CHECK-DXIL: %[[RES:.*]] = call {{.*}} float @llvm.dx.dot2add(float %{{.*}}, half %[[AX]], half %[[AY]], half %[[BX]], half %[[BY]])
// CHECK: ret float %[[RES]]
return dot2add(p1, p2, p3);
}
// CHECK-LABEL: define {{.*}}test_bool_arg1_arg2_type
float test_bool_arg1_arg2_type(bool2 p1, bool2 p2, float p3) {
// CHECK: %loadedv = trunc <2 x i32> %{{.*}} to <2 x i1>
// CHECK: %conv = uitofp <2 x i1> %loadedv to <2 x half>
// CHECK: %loadedv1 = trunc <2 x i32> %{{.*}} to <2 x i1>
// CHECK: %conv2 = uitofp <2 x i1> %loadedv1 to <2 x half>
// CHECK-SPIRV: %[[MUL:.*]] = call reassoc nnan ninf nsz arcp afn half @llvm.spv.fdot.v2f16(<2 x half> %{{.*}}, <2 x half> %{{.*}})
// CHECK-SPIRV: %[[CONV:.*]] = fpext reassoc nnan ninf nsz arcp afn half %[[MUL]] to float
// CHECK-SPIRV: %[[C:.*]] = load float, ptr %c.addr.i, align 4
// CHECK-SPIRV: %[[RES:.*]] = fadd reassoc nnan ninf nsz arcp afn float %[[CONV]], %[[C]]
// CHECK-DXIL: %[[AX:.*]] = extractelement <2 x half> %{{.*}}, i32 0
// CHECK-DXIL: %[[AY:.*]] = extractelement <2 x half> %{{.*}}, i32 1
// CHECK-DXIL: %[[BX:.*]] = extractelement <2 x half> %{{.*}}, i32 0
// CHECK-DXIL: %[[BY:.*]] = extractelement <2 x half> %{{.*}}, i32 1
// CHECK-DXIL: %[[RES:.*]] = call {{.*}} float @llvm.dx.dot2add(float %{{.*}}, half %[[AX]], half %[[AY]], half %[[BX]], half %[[BY]])
// CHECK: ret float %[[RES]]
return dot2add(p1, p2, p3);
}
|
