| 1 |
/* { dg-do compile } */ |
| 2 |
/* { dg-options "-march=rv32gcv -mabi=ilp32d -O3" } */ |
| 3 |
|
| 4 |
#include "riscv_vector.h" |
| 5 |
|
| 6 |
void f (int8_t* base1,int8_t* base2,int8_t* out,int n) |
| 7 |
{ |
| 8 |
vint8mf4_t v = __riscv_vle8_v_i8mf4 (base1, 32); |
| 9 |
for (int i = 0; i < n; i++){ |
| 10 |
v = __riscv_vmv_v_v_i8mf4 (v, 32); |
| 11 |
v = __riscv_vle8_v_i8mf4_tu (v, base2, 32); |
| 12 |
} |
| 13 |
__riscv_vse8_v_i8mf4 (out, v, 32); |
| 14 |
} |
| 15 |
|
| 16 |
void f2 (int8_t* base1,int8_t* base2,int8_t* out,int n) |
| 17 |
{ |
| 18 |
vint8mf4_t v = __riscv_vle8_v_i8mf4 (base1, 32); |
| 19 |
for (int i = 0; i < n; i++){ |
| 20 |
v = __riscv_vmv_v_x_i8mf4_tu (v, 3, 32); |
| 21 |
v = __riscv_vle8_v_i8mf4_tu (v, base2, 32); |
| 22 |
} |
| 23 |
__riscv_vse8_v_i8mf4 (out, v, 32); |
| 24 |
} |
| 25 |
|
| 26 |
void f3 (int8_t* base1,int8_t* base2,int8_t* out,int n) |
| 27 |
{ |
| 28 |
vint8mf4_t v = __riscv_vle8_v_i8mf4 (base1, 32); |
| 29 |
vint8mf4_t v2 = __riscv_vle8_v_i8mf4 (base1 + 100, 32); |
| 30 |
vbool32_t m = __riscv_vlm_v_b32 (base1 + 200, 32); |
| 31 |
for (int i = 0; i < n; i++){ |
| 32 |
v = __riscv_vmerge_vvm_i8mf4 (v, v2, m, 32); |
| 33 |
v = __riscv_vle8_v_i8mf4_tu (v, base2, 32); |
| 34 |
} |
| 35 |
__riscv_vse8_v_i8mf4 (out, v, 32); |
| 36 |
} |
| 37 |
|
| 38 |
void f4 (int8_t* base1,int8_t* base2,int8_t* out,int n) |
| 39 |
{ |
| 40 |
vint8mf4_t v = __riscv_vle8_v_i8mf4 (base1, 32); |
| 41 |
vbool32_t m = __riscv_vlm_v_b32 (base1 + 200, 32); |
| 42 |
for (int i = 0; i < n; i++){ |
| 43 |
v = __riscv_vmerge_vxm_i8mf4 (v, 3, m, 32); |
| 44 |
v = __riscv_vle8_v_i8mf4_tu (v, base2, 32); |
| 45 |
} |
| 46 |
__riscv_vse8_v_i8mf4 (out, v, 32); |
| 47 |
} |
| 48 |
|
| 49 |
void f5 (int8_t* base1,int8_t* base2,int8_t* out,int n) |
| 50 |
{ |
| 51 |
vint8mf4_t v = __riscv_vle8_v_i8mf4 (base1, 32); |
| 52 |
vbool32_t m = __riscv_vlm_v_b32 (base1 + 200, 32); |
| 53 |
for (int i = 0; i < n; i++){ |
| 54 |
v = __riscv_vmerge_vxm_i8mf4 (v, 100, m, 32); |
| 55 |
v = __riscv_vle8_v_i8mf4_tu (v, base2, 32); |
| 56 |
} |
| 57 |
__riscv_vse8_v_i8mf4 (out, v, 32); |
| 58 |
} |
| 59 |
|
| 60 |
void f6 (int8_t* base1,int8_t* base2,int8_t* out,int n) |
| 61 |
{ |
| 62 |
vuint8mf4_t v = __riscv_vle8_v_u8mf4 (base1 + 100, 32); |
| 63 |
for (int i = 0; i < n; i++){ |
| 64 |
v = __riscv_vluxei8_v_u8mf4 (base2, v, 32); |
| 65 |
v = __riscv_vle8_v_u8mf4_tu (v, base2 + 200, 32); |
| 66 |
} |
| 67 |
__riscv_vse8_v_u8mf4 (out, v, 32); |
| 68 |
} |
| 69 |
|
| 70 |
void f7 (int8_t* base1,int8_t* base2,int8_t* out,int n) |
| 71 |
{ |
| 72 |
vint8mf4_t v = __riscv_vle8_v_i8mf4 (base1, 32); |
| 73 |
for (int i = 0; i < n; i++){ |
| 74 |
v = __riscv_vadd_vv_i8mf4 (v, v, 32); |
| 75 |
v = __riscv_vle8_v_i8mf4_tu (v, base2, 32); |
| 76 |
} |
| 77 |
__riscv_vse8_v_i8mf4 (out, v, 32); |
| 78 |
} |
| 79 |
|
| 80 |
void f8 (int8_t* base1,int8_t* base2,int8_t* out,int n) |
| 81 |
{ |
| 82 |
vint8mf4_t v = __riscv_vle8_v_i8mf4 (base1, 32); |
| 83 |
for (int i = 0; i < n; i++){ |
| 84 |
v = __riscv_vadd_vv_i8mf4 (v, v, 32); |
| 85 |
v = __riscv_vle8_v_i8mf4_tu (v, base2, 32); |
| 86 |
} |
| 87 |
__riscv_vse8_v_i8mf4 (out, v, 32); |
| 88 |
} |
| 89 |
|
| 90 |
void f9 (int8_t* base1,int8_t* base2,int8_t* out,int n) |
| 91 |
{ |
| 92 |
vint8mf4_t v = __riscv_vle8_v_i8mf4 (base1, 32); |
| 93 |
for (int i = 0; i < n; i++){ |
| 94 |
v = __riscv_vsll_vx_i8mf4 (v, 101, 32); |
| 95 |
v = __riscv_vle8_v_i8mf4_tu (v, base2, 32); |
| 96 |
} |
| 97 |
__riscv_vse8_v_i8mf4 (out, v, 32); |
| 98 |
} |
| 99 |
|
| 100 |
void f10 (int8_t* base1,int8_t* base2,int8_t* out,int n) |
| 101 |
{ |
| 102 |
vint8mf4_t v = __riscv_vle8_v_i8mf4 (base1, 32); |
| 103 |
for (int i = 0; i < n; i++){ |
| 104 |
v = __riscv_vor_vx_i8mf4 (v, 101, 32); |
| 105 |
v = __riscv_vle8_v_i8mf4_tu (v, base2, 32); |
| 106 |
} |
| 107 |
__riscv_vse8_v_i8mf4 (out, v, 32); |
| 108 |
} |
| 109 |
|
| 110 |
void f11 (int8_t* base1,int8_t* base2,int8_t* out,int n) |
| 111 |
{ |
| 112 |
vint8mf4_t v = __riscv_vle8_v_i8mf4 (base1, 32); |
| 113 |
vbool32_t m = __riscv_vlm_v_b32 (base1 + 200, 32); |
| 114 |
for (int i = 0; i < n; i++){ |
| 115 |
v = __riscv_vadc_vvm_i8mf4 (v, v, m, 32); |
| 116 |
v = __riscv_vle8_v_i8mf4_tu (v, base2, 32); |
| 117 |
} |
| 118 |
__riscv_vse8_v_i8mf4 (out, v, 32); |
| 119 |
} |
| 120 |
|
| 121 |
void f12 (int8_t* base1,int8_t* base2,int8_t* out,int n) |
| 122 |
{ |
| 123 |
vint8mf4_t v = __riscv_vle8_v_i8mf4 (base1, 32); |
| 124 |
vbool32_t m = __riscv_vlm_v_b32 (base1 + 200, 32); |
| 125 |
for (int i = 0; i < n; i++){ |
| 126 |
v = __riscv_vsbc_vvm_i8mf4 (v, v, m, 32); |
| 127 |
v = __riscv_vle8_v_i8mf4_tu (v, base2, 32); |
| 128 |
} |
| 129 |
__riscv_vse8_v_i8mf4 (out, v, 32); |
| 130 |
} |
| 131 |
|
| 132 |
void f13 (int8_t* base1,int8_t* base2,int8_t* out,int n) |
| 133 |
{ |
| 134 |
vint8mf4_t v = __riscv_vle8_v_i8mf4 (base1, 32); |
| 135 |
vbool32_t m = __riscv_vlm_v_b32 (base1 + 200, 32); |
| 136 |
for (int i = 0; i < n; i++){ |
| 137 |
v = __riscv_vadc_vxm_i8mf4 (v, 100, m, 32); |
| 138 |
v = __riscv_vle8_v_i8mf4_tu (v, base2, 32); |
| 139 |
} |
| 140 |
__riscv_vse8_v_i8mf4 (out, v, 32); |
| 141 |
} |
| 142 |
|
| 143 |
void f14 (int8_t* base1,int8_t* base2,int8_t* out,int n) |
| 144 |
{ |
| 145 |
vint8mf4_t v = __riscv_vle8_v_i8mf4 (base1, 32); |
| 146 |
vbool32_t m = __riscv_vlm_v_b32 (base1 + 200, 32); |
| 147 |
for (int i = 0; i < n; i++){ |
| 148 |
v = __riscv_vsbc_vxm_i8mf4 (v, 100, m, 32); |
| 149 |
v = __riscv_vle8_v_i8mf4_tu (v, base2, 32); |
| 150 |
} |
| 151 |
__riscv_vse8_v_i8mf4 (out, v, 32); |
| 152 |
} |
| 153 |
|
| 154 |
void f15 (int8_t* base1,int8_t* base2,int8_t* out,int n) |
| 155 |
{ |
| 156 |
vint8mf4_t v = __riscv_vle8_v_i8mf4 (base1, 32); |
| 157 |
for (int i = 0; i < n; i++){ |
| 158 |
v = __riscv_vneg_v_i8mf4 (v, 32); |
| 159 |
v = __riscv_vle8_v_i8mf4_tu (v, base2, 32); |
| 160 |
} |
| 161 |
__riscv_vse8_v_i8mf4 (out, v, 32); |
| 162 |
} |
| 163 |
|
| 164 |
void f16 (int8_t* base1,int8_t* base2,int8_t* out,int n) |
| 165 |
{ |
| 166 |
vint8mf4_t v = __riscv_vle8_v_i8mf4 (base1, 32); |
| 167 |
for (int i = 0; i < n; i++){ |
| 168 |
v = __riscv_vsadd_vv_i8mf4 (v, v, 32); |
| 169 |
v = __riscv_vle8_v_i8mf4_tu (v, base2, 32); |
| 170 |
} |
| 171 |
__riscv_vse8_v_i8mf4 (out, v, 32); |
| 172 |
} |
| 173 |
|
| 174 |
void f17 (int8_t* base1,int8_t* base2,int8_t* out,int n) |
| 175 |
{ |
| 176 |
vint8mf4_t v = __riscv_vle8_v_i8mf4 (base1, 32); |
| 177 |
for (int i = 0; i < n; i++){ |
| 178 |
v = __riscv_vsadd_vx_i8mf4 (v, 100, 32); |
| 179 |
v = __riscv_vle8_v_i8mf4_tu (v, base2, 32); |
| 180 |
} |
| 181 |
__riscv_vse8_v_i8mf4 (out, v, 32); |
| 182 |
} |
| 183 |
|
| 184 |
void f18 (int8_t* base1,int8_t* base2,int8_t* out,int n) |
| 185 |
{ |
| 186 |
vint8mf4_t v = __riscv_vle8_v_i8mf4 (base1, 32); |
| 187 |
for (int i = 0; i < n; i++){ |
| 188 |
v = __riscv_vaadd_vv_i8mf4 (v, v, 0, 32); |
| 189 |
v = __riscv_vle8_v_i8mf4_tu (v, base2, 32); |
| 190 |
} |
| 191 |
__riscv_vse8_v_i8mf4 (out, v, 32); |
| 192 |
} |
| 193 |
|
| 194 |
void f19 (int8_t* base1,int8_t* base2,int8_t* out,int n) |
| 195 |
{ |
| 196 |
vint8mf4_t v = __riscv_vle8_v_i8mf4 (base1, 32); |
| 197 |
for (int i = 0; i < n; i++){ |
| 198 |
v = __riscv_vaadd_vx_i8mf4 (v, 100, 0, 32); |
| 199 |
v = __riscv_vle8_v_i8mf4_tu (v, base2, 32); |
| 200 |
} |
| 201 |
__riscv_vse8_v_i8mf4 (out, v, 32); |
| 202 |
} |
| 203 |
|
| 204 |
/* { dg-final { scan-assembler-not {vmv1r} } } */ |