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
201
202
203
204
205
206
207
|
// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
// REQUIRES: riscv-registered-target
// RUN: %clang_cc1 -triple riscv64 -target-feature +v -target-feature +zfh \
// RUN: -target-feature +experimental-zvfh -disable-O0-optnone \
// RUN: -emit-llvm %s -o - | opt -S -passes=mem2reg | \
// RUN: FileCheck --check-prefix=CHECK-RV64 %s
#include <riscv_vector.h>
// CHECK-RV64-LABEL: @test_vfwredusum_vs_f16mf4_f32m1_tu(
// CHECK-RV64-NEXT: entry:
// CHECK-RV64-NEXT: [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwredusum.nxv2f32.nxv1f16.i64(<vscale x 2 x float> [[MASKEDOFF:%.*]], <vscale x 1 x half> [[VECTOR:%.*]], <vscale x 2 x float> [[SCALAR:%.*]], i64 [[VL:%.*]])
// CHECK-RV64-NEXT: ret <vscale x 2 x float> [[TMP0]]
//
vfloat32m1_t test_vfwredusum_vs_f16mf4_f32m1_tu(vfloat32m1_t maskedoff, vfloat16mf4_t vector, vfloat32m1_t scalar, size_t vl) {
return __riscv_vfwredusum_tu(maskedoff, vector, scalar, vl);
}
// CHECK-RV64-LABEL: @test_vfwredusum_vs_f16mf2_f32m1_tu(
// CHECK-RV64-NEXT: entry:
// CHECK-RV64-NEXT: [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwredusum.nxv2f32.nxv2f16.i64(<vscale x 2 x float> [[MASKEDOFF:%.*]], <vscale x 2 x half> [[VECTOR:%.*]], <vscale x 2 x float> [[SCALAR:%.*]], i64 [[VL:%.*]])
// CHECK-RV64-NEXT: ret <vscale x 2 x float> [[TMP0]]
//
vfloat32m1_t test_vfwredusum_vs_f16mf2_f32m1_tu(vfloat32m1_t maskedoff, vfloat16mf2_t vector, vfloat32m1_t scalar, size_t vl) {
return __riscv_vfwredusum_tu(maskedoff, vector, scalar, vl);
}
// CHECK-RV64-LABEL: @test_vfwredusum_vs_f16m1_f32m1_tu(
// CHECK-RV64-NEXT: entry:
// CHECK-RV64-NEXT: [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwredusum.nxv2f32.nxv4f16.i64(<vscale x 2 x float> [[MASKEDOFF:%.*]], <vscale x 4 x half> [[VECTOR:%.*]], <vscale x 2 x float> [[SCALAR:%.*]], i64 [[VL:%.*]])
// CHECK-RV64-NEXT: ret <vscale x 2 x float> [[TMP0]]
//
vfloat32m1_t test_vfwredusum_vs_f16m1_f32m1_tu(vfloat32m1_t maskedoff, vfloat16m1_t vector, vfloat32m1_t scalar, size_t vl) {
return __riscv_vfwredusum_tu(maskedoff, vector, scalar, vl);
}
// CHECK-RV64-LABEL: @test_vfwredusum_vs_f16m2_f32m1_tu(
// CHECK-RV64-NEXT: entry:
// CHECK-RV64-NEXT: [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwredusum.nxv2f32.nxv8f16.i64(<vscale x 2 x float> [[MASKEDOFF:%.*]], <vscale x 8 x half> [[VECTOR:%.*]], <vscale x 2 x float> [[SCALAR:%.*]], i64 [[VL:%.*]])
// CHECK-RV64-NEXT: ret <vscale x 2 x float> [[TMP0]]
//
vfloat32m1_t test_vfwredusum_vs_f16m2_f32m1_tu(vfloat32m1_t maskedoff, vfloat16m2_t vector, vfloat32m1_t scalar, size_t vl) {
return __riscv_vfwredusum_tu(maskedoff, vector, scalar, vl);
}
// CHECK-RV64-LABEL: @test_vfwredusum_vs_f16m4_f32m1_tu(
// CHECK-RV64-NEXT: entry:
// CHECK-RV64-NEXT: [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwredusum.nxv2f32.nxv16f16.i64(<vscale x 2 x float> [[MASKEDOFF:%.*]], <vscale x 16 x half> [[VECTOR:%.*]], <vscale x 2 x float> [[SCALAR:%.*]], i64 [[VL:%.*]])
// CHECK-RV64-NEXT: ret <vscale x 2 x float> [[TMP0]]
//
vfloat32m1_t test_vfwredusum_vs_f16m4_f32m1_tu(vfloat32m1_t maskedoff, vfloat16m4_t vector, vfloat32m1_t scalar, size_t vl) {
return __riscv_vfwredusum_tu(maskedoff, vector, scalar, vl);
}
// CHECK-RV64-LABEL: @test_vfwredusum_vs_f16m8_f32m1_tu(
// CHECK-RV64-NEXT: entry:
// CHECK-RV64-NEXT: [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwredusum.nxv2f32.nxv32f16.i64(<vscale x 2 x float> [[MASKEDOFF:%.*]], <vscale x 32 x half> [[VECTOR:%.*]], <vscale x 2 x float> [[SCALAR:%.*]], i64 [[VL:%.*]])
// CHECK-RV64-NEXT: ret <vscale x 2 x float> [[TMP0]]
//
vfloat32m1_t test_vfwredusum_vs_f16m8_f32m1_tu(vfloat32m1_t maskedoff, vfloat16m8_t vector, vfloat32m1_t scalar, size_t vl) {
return __riscv_vfwredusum_tu(maskedoff, vector, scalar, vl);
}
// CHECK-RV64-LABEL: @test_vfwredusum_vs_f32mf2_f64m1_tu(
// CHECK-RV64-NEXT: entry:
// CHECK-RV64-NEXT: [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfwredusum.nxv1f64.nxv1f32.i64(<vscale x 1 x double> [[MASKEDOFF:%.*]], <vscale x 1 x float> [[VECTOR:%.*]], <vscale x 1 x double> [[SCALAR:%.*]], i64 [[VL:%.*]])
// CHECK-RV64-NEXT: ret <vscale x 1 x double> [[TMP0]]
//
vfloat64m1_t test_vfwredusum_vs_f32mf2_f64m1_tu(vfloat64m1_t maskedoff, vfloat32mf2_t vector, vfloat64m1_t scalar, size_t vl) {
return __riscv_vfwredusum_tu(maskedoff, vector, scalar, vl);
}
// CHECK-RV64-LABEL: @test_vfwredusum_vs_f32m1_f64m1_tu(
// CHECK-RV64-NEXT: entry:
// CHECK-RV64-NEXT: [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfwredusum.nxv1f64.nxv2f32.i64(<vscale x 1 x double> [[MASKEDOFF:%.*]], <vscale x 2 x float> [[VECTOR:%.*]], <vscale x 1 x double> [[SCALAR:%.*]], i64 [[VL:%.*]])
// CHECK-RV64-NEXT: ret <vscale x 1 x double> [[TMP0]]
//
vfloat64m1_t test_vfwredusum_vs_f32m1_f64m1_tu(vfloat64m1_t maskedoff, vfloat32m1_t vector, vfloat64m1_t scalar, size_t vl) {
return __riscv_vfwredusum_tu(maskedoff, vector, scalar, vl);
}
// CHECK-RV64-LABEL: @test_vfwredusum_vs_f32m2_f64m1_tu(
// CHECK-RV64-NEXT: entry:
// CHECK-RV64-NEXT: [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfwredusum.nxv1f64.nxv4f32.i64(<vscale x 1 x double> [[MASKEDOFF:%.*]], <vscale x 4 x float> [[VECTOR:%.*]], <vscale x 1 x double> [[SCALAR:%.*]], i64 [[VL:%.*]])
// CHECK-RV64-NEXT: ret <vscale x 1 x double> [[TMP0]]
//
vfloat64m1_t test_vfwredusum_vs_f32m2_f64m1_tu(vfloat64m1_t maskedoff, vfloat32m2_t vector, vfloat64m1_t scalar, size_t vl) {
return __riscv_vfwredusum_tu(maskedoff, vector, scalar, vl);
}
// CHECK-RV64-LABEL: @test_vfwredusum_vs_f32m4_f64m1_tu(
// CHECK-RV64-NEXT: entry:
// CHECK-RV64-NEXT: [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfwredusum.nxv1f64.nxv8f32.i64(<vscale x 1 x double> [[MASKEDOFF:%.*]], <vscale x 8 x float> [[VECTOR:%.*]], <vscale x 1 x double> [[SCALAR:%.*]], i64 [[VL:%.*]])
// CHECK-RV64-NEXT: ret <vscale x 1 x double> [[TMP0]]
//
vfloat64m1_t test_vfwredusum_vs_f32m4_f64m1_tu(vfloat64m1_t maskedoff, vfloat32m4_t vector, vfloat64m1_t scalar, size_t vl) {
return __riscv_vfwredusum_tu(maskedoff, vector, scalar, vl);
}
// CHECK-RV64-LABEL: @test_vfwredusum_vs_f32m8_f64m1_tu(
// CHECK-RV64-NEXT: entry:
// CHECK-RV64-NEXT: [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfwredusum.nxv1f64.nxv16f32.i64(<vscale x 1 x double> [[MASKEDOFF:%.*]], <vscale x 16 x float> [[VECTOR:%.*]], <vscale x 1 x double> [[SCALAR:%.*]], i64 [[VL:%.*]])
// CHECK-RV64-NEXT: ret <vscale x 1 x double> [[TMP0]]
//
vfloat64m1_t test_vfwredusum_vs_f32m8_f64m1_tu(vfloat64m1_t maskedoff, vfloat32m8_t vector, vfloat64m1_t scalar, size_t vl) {
return __riscv_vfwredusum_tu(maskedoff, vector, scalar, vl);
}
// CHECK-RV64-LABEL: @test_vfwredusum_vs_f16mf4_f32m1_tum(
// CHECK-RV64-NEXT: entry:
// CHECK-RV64-NEXT: [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwredusum.mask.nxv2f32.nxv1f16.i64(<vscale x 2 x float> [[MASKEDOFF:%.*]], <vscale x 1 x half> [[VECTOR:%.*]], <vscale x 2 x float> [[SCALAR:%.*]], <vscale x 1 x i1> [[MASK:%.*]], i64 [[VL:%.*]])
// CHECK-RV64-NEXT: ret <vscale x 2 x float> [[TMP0]]
//
vfloat32m1_t test_vfwredusum_vs_f16mf4_f32m1_tum(vbool64_t mask, vfloat32m1_t maskedoff, vfloat16mf4_t vector, vfloat32m1_t scalar, size_t vl) {
return __riscv_vfwredusum_tum(mask, maskedoff, vector, scalar, vl);
}
// CHECK-RV64-LABEL: @test_vfwredusum_vs_f16mf2_f32m1_tum(
// CHECK-RV64-NEXT: entry:
// CHECK-RV64-NEXT: [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwredusum.mask.nxv2f32.nxv2f16.i64(<vscale x 2 x float> [[MASKEDOFF:%.*]], <vscale x 2 x half> [[VECTOR:%.*]], <vscale x 2 x float> [[SCALAR:%.*]], <vscale x 2 x i1> [[MASK:%.*]], i64 [[VL:%.*]])
// CHECK-RV64-NEXT: ret <vscale x 2 x float> [[TMP0]]
//
vfloat32m1_t test_vfwredusum_vs_f16mf2_f32m1_tum(vbool32_t mask, vfloat32m1_t maskedoff, vfloat16mf2_t vector, vfloat32m1_t scalar, size_t vl) {
return __riscv_vfwredusum_tum(mask, maskedoff, vector, scalar, vl);
}
// CHECK-RV64-LABEL: @test_vfwredusum_vs_f16m1_f32m1_tum(
// CHECK-RV64-NEXT: entry:
// CHECK-RV64-NEXT: [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwredusum.mask.nxv2f32.nxv4f16.i64(<vscale x 2 x float> [[MASKEDOFF:%.*]], <vscale x 4 x half> [[VECTOR:%.*]], <vscale x 2 x float> [[SCALAR:%.*]], <vscale x 4 x i1> [[MASK:%.*]], i64 [[VL:%.*]])
// CHECK-RV64-NEXT: ret <vscale x 2 x float> [[TMP0]]
//
vfloat32m1_t test_vfwredusum_vs_f16m1_f32m1_tum(vbool16_t mask, vfloat32m1_t maskedoff, vfloat16m1_t vector, vfloat32m1_t scalar, size_t vl) {
return __riscv_vfwredusum_tum(mask, maskedoff, vector, scalar, vl);
}
// CHECK-RV64-LABEL: @test_vfwredusum_vs_f16m2_f32m1_tum(
// CHECK-RV64-NEXT: entry:
// CHECK-RV64-NEXT: [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwredusum.mask.nxv2f32.nxv8f16.i64(<vscale x 2 x float> [[MASKEDOFF:%.*]], <vscale x 8 x half> [[VECTOR:%.*]], <vscale x 2 x float> [[SCALAR:%.*]], <vscale x 8 x i1> [[MASK:%.*]], i64 [[VL:%.*]])
// CHECK-RV64-NEXT: ret <vscale x 2 x float> [[TMP0]]
//
vfloat32m1_t test_vfwredusum_vs_f16m2_f32m1_tum(vbool8_t mask, vfloat32m1_t maskedoff, vfloat16m2_t vector, vfloat32m1_t scalar, size_t vl) {
return __riscv_vfwredusum_tum(mask, maskedoff, vector, scalar, vl);
}
// CHECK-RV64-LABEL: @test_vfwredusum_vs_f16m4_f32m1_tum(
// CHECK-RV64-NEXT: entry:
// CHECK-RV64-NEXT: [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwredusum.mask.nxv2f32.nxv16f16.i64(<vscale x 2 x float> [[MASKEDOFF:%.*]], <vscale x 16 x half> [[VECTOR:%.*]], <vscale x 2 x float> [[SCALAR:%.*]], <vscale x 16 x i1> [[MASK:%.*]], i64 [[VL:%.*]])
// CHECK-RV64-NEXT: ret <vscale x 2 x float> [[TMP0]]
//
vfloat32m1_t test_vfwredusum_vs_f16m4_f32m1_tum(vbool4_t mask, vfloat32m1_t maskedoff, vfloat16m4_t vector, vfloat32m1_t scalar, size_t vl) {
return __riscv_vfwredusum_tum(mask, maskedoff, vector, scalar, vl);
}
// CHECK-RV64-LABEL: @test_vfwredusum_vs_f16m8_f32m1_tum(
// CHECK-RV64-NEXT: entry:
// CHECK-RV64-NEXT: [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwredusum.mask.nxv2f32.nxv32f16.i64(<vscale x 2 x float> [[MASKEDOFF:%.*]], <vscale x 32 x half> [[VECTOR:%.*]], <vscale x 2 x float> [[SCALAR:%.*]], <vscale x 32 x i1> [[MASK:%.*]], i64 [[VL:%.*]])
// CHECK-RV64-NEXT: ret <vscale x 2 x float> [[TMP0]]
//
vfloat32m1_t test_vfwredusum_vs_f16m8_f32m1_tum(vbool2_t mask, vfloat32m1_t maskedoff, vfloat16m8_t vector, vfloat32m1_t scalar, size_t vl) {
return __riscv_vfwredusum_tum(mask, maskedoff, vector, scalar, vl);
}
// CHECK-RV64-LABEL: @test_vfwredusum_vs_f32mf2_f64m1_tum(
// CHECK-RV64-NEXT: entry:
// CHECK-RV64-NEXT: [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfwredusum.mask.nxv1f64.nxv1f32.i64(<vscale x 1 x double> [[MASKEDOFF:%.*]], <vscale x 1 x float> [[VECTOR:%.*]], <vscale x 1 x double> [[SCALAR:%.*]], <vscale x 1 x i1> [[MASK:%.*]], i64 [[VL:%.*]])
// CHECK-RV64-NEXT: ret <vscale x 1 x double> [[TMP0]]
//
vfloat64m1_t test_vfwredusum_vs_f32mf2_f64m1_tum(vbool64_t mask, vfloat64m1_t maskedoff, vfloat32mf2_t vector, vfloat64m1_t scalar, size_t vl) {
return __riscv_vfwredusum_tum(mask, maskedoff, vector, scalar, vl);
}
// CHECK-RV64-LABEL: @test_vfwredusum_vs_f32m1_f64m1_tum(
// CHECK-RV64-NEXT: entry:
// CHECK-RV64-NEXT: [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfwredusum.mask.nxv1f64.nxv2f32.i64(<vscale x 1 x double> [[MASKEDOFF:%.*]], <vscale x 2 x float> [[VECTOR:%.*]], <vscale x 1 x double> [[SCALAR:%.*]], <vscale x 2 x i1> [[MASK:%.*]], i64 [[VL:%.*]])
// CHECK-RV64-NEXT: ret <vscale x 1 x double> [[TMP0]]
//
vfloat64m1_t test_vfwredusum_vs_f32m1_f64m1_tum(vbool32_t mask, vfloat64m1_t maskedoff, vfloat32m1_t vector, vfloat64m1_t scalar, size_t vl) {
return __riscv_vfwredusum_tum(mask, maskedoff, vector, scalar, vl);
}
// CHECK-RV64-LABEL: @test_vfwredusum_vs_f32m2_f64m1_tum(
// CHECK-RV64-NEXT: entry:
// CHECK-RV64-NEXT: [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfwredusum.mask.nxv1f64.nxv4f32.i64(<vscale x 1 x double> [[MASKEDOFF:%.*]], <vscale x 4 x float> [[VECTOR:%.*]], <vscale x 1 x double> [[SCALAR:%.*]], <vscale x 4 x i1> [[MASK:%.*]], i64 [[VL:%.*]])
// CHECK-RV64-NEXT: ret <vscale x 1 x double> [[TMP0]]
//
vfloat64m1_t test_vfwredusum_vs_f32m2_f64m1_tum(vbool16_t mask, vfloat64m1_t maskedoff, vfloat32m2_t vector, vfloat64m1_t scalar, size_t vl) {
return __riscv_vfwredusum_tum(mask, maskedoff, vector, scalar, vl);
}
// CHECK-RV64-LABEL: @test_vfwredusum_vs_f32m4_f64m1_tum(
// CHECK-RV64-NEXT: entry:
// CHECK-RV64-NEXT: [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfwredusum.mask.nxv1f64.nxv8f32.i64(<vscale x 1 x double> [[MASKEDOFF:%.*]], <vscale x 8 x float> [[VECTOR:%.*]], <vscale x 1 x double> [[SCALAR:%.*]], <vscale x 8 x i1> [[MASK:%.*]], i64 [[VL:%.*]])
// CHECK-RV64-NEXT: ret <vscale x 1 x double> [[TMP0]]
//
vfloat64m1_t test_vfwredusum_vs_f32m4_f64m1_tum(vbool8_t mask, vfloat64m1_t maskedoff, vfloat32m4_t vector, vfloat64m1_t scalar, size_t vl) {
return __riscv_vfwredusum_tum(mask, maskedoff, vector, scalar, vl);
}
// CHECK-RV64-LABEL: @test_vfwredusum_vs_f32m8_f64m1_tum(
// CHECK-RV64-NEXT: entry:
// CHECK-RV64-NEXT: [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfwredusum.mask.nxv1f64.nxv16f32.i64(<vscale x 1 x double> [[MASKEDOFF:%.*]], <vscale x 16 x float> [[VECTOR:%.*]], <vscale x 1 x double> [[SCALAR:%.*]], <vscale x 16 x i1> [[MASK:%.*]], i64 [[VL:%.*]])
// CHECK-RV64-NEXT: ret <vscale x 1 x double> [[TMP0]]
//
vfloat64m1_t test_vfwredusum_vs_f32m8_f64m1_tum(vbool4_t mask, vfloat64m1_t maskedoff, vfloat32m8_t vector, vfloat64m1_t scalar, size_t vl) {
return __riscv_vfwredusum_tum(mask, maskedoff, vector, scalar, vl);
}
|
