| 1 |
#include <arm_neon.h> |
| 2 |
#include "arm-neon-ref.h" |
| 3 |
#include "compute-ref-data.h" |
| 4 |
|
| 5 |
/* Expected results. */ |
| 6 |
/* Chunk 0. */ |
| 7 |
VECT_VAR_DECL(expected0,int,8,8) [] = { 0xf0, 0xf0, 0xf0, 0xf0, |
| 8 |
0xf0, 0xf0, 0xf0, 0xf0 }; |
| 9 |
VECT_VAR_DECL(expected0,int,16,4) [] = { 0xfff0, 0xfff0, 0xfff0, 0xfff0 }; |
| 10 |
VECT_VAR_DECL(expected0,int,32,2) [] = { 0xfffffff0, 0xfffffff0 }; |
| 11 |
VECT_VAR_DECL(expected0,int,64,1) [] = { 0xfffffffffffffff0 }; |
| 12 |
VECT_VAR_DECL(expected0,uint,8,8) [] = { 0xf0, 0xf0, 0xf0, 0xf0, |
| 13 |
0xf0, 0xf0, 0xf0, 0xf0 }; |
| 14 |
VECT_VAR_DECL(expected0,uint,16,4) [] = { 0xfff0, 0xfff0, 0xfff0, 0xfff0 }; |
| 15 |
VECT_VAR_DECL(expected0,uint,32,2) [] = { 0xfffffff0, 0xfffffff0 }; |
| 16 |
VECT_VAR_DECL(expected0,uint,64,1) [] = { 0xfffffffffffffff0 }; |
| 17 |
VECT_VAR_DECL(expected0,poly,8,8) [] = { 0xf0, 0xf0, 0xf0, 0xf0, |
| 18 |
0xf0, 0xf0, 0xf0, 0xf0 }; |
| 19 |
VECT_VAR_DECL(expected0,poly,16,4) [] = { 0xfff0, 0xfff0, 0xfff0, 0xfff0 }; |
| 20 |
#if MFLOAT8_SUPPORTED |
| 21 |
VECT_VAR_DECL(expected0,hmfloat,8,8) [] = { 0xf0, 0xf0, 0xf0, 0xf0, |
| 22 |
0xf0, 0xf0, 0xf0, 0xf0 }; |
| 23 |
#endif |
| 24 |
VECT_VAR_DECL(expected0,hfloat,16,4) [] = { 0xcc00, 0xcc00, 0xcc00, 0xcc00 }; |
| 25 |
VECT_VAR_DECL(expected0,hfloat,32,2) [] = { 0xc1800000, 0xc1800000 }; |
| 26 |
VECT_VAR_DECL(expected0,int,8,16) [] = { 0xf0, 0xf0, 0xf0, 0xf0, |
| 27 |
0xf0, 0xf0, 0xf0, 0xf0, |
| 28 |
0xf0, 0xf0, 0xf0, 0xf0, |
| 29 |
0xf0, 0xf0, 0xf0, 0xf0 }; |
| 30 |
VECT_VAR_DECL(expected0,int,16,8) [] = { 0xfff0, 0xfff0, 0xfff0, 0xfff0, |
| 31 |
0xfff0, 0xfff0, 0xfff0, 0xfff0 }; |
| 32 |
VECT_VAR_DECL(expected0,int,32,4) [] = { 0xfffffff0, 0xfffffff0, |
| 33 |
0xfffffff0, 0xfffffff0 }; |
| 34 |
VECT_VAR_DECL(expected0,int,64,2) [] = { 0xfffffffffffffff0, |
| 35 |
0xfffffffffffffff0 }; |
| 36 |
VECT_VAR_DECL(expected0,uint,8,16) [] = { 0xf0, 0xf0, 0xf0, 0xf0, |
| 37 |
0xf0, 0xf0, 0xf0, 0xf0, |
| 38 |
0xf0, 0xf0, 0xf0, 0xf0, |
| 39 |
0xf0, 0xf0, 0xf0, 0xf0 }; |
| 40 |
VECT_VAR_DECL(expected0,uint,16,8) [] = { 0xfff0, 0xfff0, 0xfff0, 0xfff0, |
| 41 |
0xfff0, 0xfff0, 0xfff0, 0xfff0 }; |
| 42 |
VECT_VAR_DECL(expected0,uint,32,4) [] = { 0xfffffff0, 0xfffffff0, |
| 43 |
0xfffffff0, 0xfffffff0 }; |
| 44 |
VECT_VAR_DECL(expected0,uint,64,2) [] = { 0xfffffffffffffff0, |
| 45 |
0xfffffffffffffff0 }; |
| 46 |
VECT_VAR_DECL(expected0,poly,8,16) [] = { 0xf0, 0xf0, 0xf0, 0xf0, |
| 47 |
0xf0, 0xf0, 0xf0, 0xf0, |
| 48 |
0xf0, 0xf0, 0xf0, 0xf0, |
| 49 |
0xf0, 0xf0, 0xf0, 0xf0 }; |
| 50 |
VECT_VAR_DECL(expected0,poly,16,8) [] = { 0xfff0, 0xfff0, 0xfff0, 0xfff0, |
| 51 |
0xfff0, 0xfff0, 0xfff0, 0xfff0 }; |
| 52 |
#if MFLOAT8_SUPPORTED |
| 53 |
VECT_VAR_DECL(expected0,hmfloat,8,16) [] = { 0xf0, 0xf0, 0xf0, 0xf0, |
| 54 |
0xf0, 0xf0, 0xf0, 0xf0, |
| 55 |
0xf0, 0xf0, 0xf0, 0xf0, |
| 56 |
0xf0, 0xf0, 0xf0, 0xf0 }; |
| 57 |
#endif |
| 58 |
VECT_VAR_DECL(expected0,hfloat,16,8) [] = { 0xcc00, 0xcc00, 0xcc00, 0xcc00, |
| 59 |
0xcc00, 0xcc00, 0xcc00, 0xcc00 }; |
| 60 |
VECT_VAR_DECL(expected0,hfloat,32,4) [] = { 0xc1800000, 0xc1800000, |
| 61 |
0xc1800000, 0xc1800000 }; |
| 62 |
|
| 63 |
/* Chunk 1. */ |
| 64 |
VECT_VAR_DECL(expected1,int,8,8) [] = { 0xf1, 0xf1, 0xf1, 0xf1, |
| 65 |
0xf1, 0xf1, 0xf1, 0xf1 }; |
| 66 |
VECT_VAR_DECL(expected1,int,16,4) [] = { 0xfff1, 0xfff1, 0xfff1, 0xfff1 }; |
| 67 |
VECT_VAR_DECL(expected1,int,32,2) [] = { 0xfffffff1, 0xfffffff1 }; |
| 68 |
VECT_VAR_DECL(expected1,int,64,1) [] = { 0xfffffffffffffff1 }; |
| 69 |
VECT_VAR_DECL(expected1,uint,8,8) [] = { 0xf1, 0xf1, 0xf1, 0xf1, |
| 70 |
0xf1, 0xf1, 0xf1, 0xf1 }; |
| 71 |
VECT_VAR_DECL(expected1,uint,16,4) [] = { 0xfff1, 0xfff1, 0xfff1, 0xfff1 }; |
| 72 |
VECT_VAR_DECL(expected1,uint,32,2) [] = { 0xfffffff1, 0xfffffff1 }; |
| 73 |
VECT_VAR_DECL(expected1,uint,64,1) [] = { 0xfffffffffffffff1 }; |
| 74 |
VECT_VAR_DECL(expected1,poly,8,8) [] = { 0xf1, 0xf1, 0xf1, 0xf1, |
| 75 |
0xf1, 0xf1, 0xf1, 0xf1 }; |
| 76 |
VECT_VAR_DECL(expected1,poly,16,4) [] = { 0xfff1, 0xfff1, 0xfff1, 0xfff1 }; |
| 77 |
#if MFLOAT8_SUPPORTED |
| 78 |
VECT_VAR_DECL(expected1,hmfloat,8,8) [] = { 0xf1, 0xf1, 0xf1, 0xf1, |
| 79 |
0xf1, 0xf1, 0xf1, 0xf1 }; |
| 80 |
#endif |
| 81 |
VECT_VAR_DECL(expected1,hfloat,16,4) [] = { 0xcb80, 0xcb80, 0xcb80, 0xcb80 }; |
| 82 |
VECT_VAR_DECL(expected1,hfloat,32,2) [] = { 0xc1700000, 0xc1700000 }; |
| 83 |
VECT_VAR_DECL(expected1,int,8,16) [] = { 0xf1, 0xf1, 0xf1, 0xf1, |
| 84 |
0xf1, 0xf1, 0xf1, 0xf1, |
| 85 |
0xf1, 0xf1, 0xf1, 0xf1, |
| 86 |
0xf1, 0xf1, 0xf1, 0xf1 }; |
| 87 |
VECT_VAR_DECL(expected1,int,16,8) [] = { 0xfff1, 0xfff1, 0xfff1, 0xfff1, |
| 88 |
0xfff1, 0xfff1, 0xfff1, 0xfff1 }; |
| 89 |
VECT_VAR_DECL(expected1,int,32,4) [] = { 0xfffffff1, 0xfffffff1, |
| 90 |
0xfffffff1, 0xfffffff1 }; |
| 91 |
VECT_VAR_DECL(expected1,int,64,2) [] = { 0xfffffffffffffff1, |
| 92 |
0xfffffffffffffff1 }; |
| 93 |
VECT_VAR_DECL(expected1,uint,8,16) [] = { 0xf1, 0xf1, 0xf1, 0xf1, |
| 94 |
0xf1, 0xf1, 0xf1, 0xf1, |
| 95 |
0xf1, 0xf1, 0xf1, 0xf1, |
| 96 |
0xf1, 0xf1, 0xf1, 0xf1 }; |
| 97 |
VECT_VAR_DECL(expected1,uint,16,8) [] = { 0xfff1, 0xfff1, 0xfff1, 0xfff1, |
| 98 |
0xfff1, 0xfff1, 0xfff1, 0xfff1 }; |
| 99 |
VECT_VAR_DECL(expected1,uint,32,4) [] = { 0xfffffff1, 0xfffffff1, |
| 100 |
0xfffffff1, 0xfffffff1 }; |
| 101 |
VECT_VAR_DECL(expected1,uint,64,2) [] = { 0xfffffffffffffff1, |
| 102 |
0xfffffffffffffff1 }; |
| 103 |
VECT_VAR_DECL(expected1,poly,8,16) [] = { 0xf1, 0xf1, 0xf1, 0xf1, |
| 104 |
0xf1, 0xf1, 0xf1, 0xf1, |
| 105 |
0xf1, 0xf1, 0xf1, 0xf1, |
| 106 |
0xf1, 0xf1, 0xf1, 0xf1 }; |
| 107 |
VECT_VAR_DECL(expected1,poly,16,8) [] = { 0xfff1, 0xfff1, 0xfff1, 0xfff1, |
| 108 |
0xfff1, 0xfff1, 0xfff1, 0xfff1 }; |
| 109 |
#if MFLOAT8_SUPPORTED |
| 110 |
VECT_VAR_DECL(expected1,hmfloat,8,16) [] = { 0xf1, 0xf1, 0xf1, 0xf1, |
| 111 |
0xf1, 0xf1, 0xf1, 0xf1, |
| 112 |
0xf1, 0xf1, 0xf1, 0xf1, |
| 113 |
0xf1, 0xf1, 0xf1, 0xf1 }; |
| 114 |
#endif |
| 115 |
VECT_VAR_DECL(expected1,hfloat,16,8) [] = { 0xcb80, 0xcb80, 0xcb80, 0xcb80, |
| 116 |
0xcb80, 0xcb80, 0xcb80, 0xcb80 }; |
| 117 |
VECT_VAR_DECL(expected1,hfloat,32,4) [] = { 0xc1700000, 0xc1700000, |
| 118 |
0xc1700000, 0xc1700000 }; |
| 119 |
|
| 120 |
/* Chunk 2. */ |
| 121 |
VECT_VAR_DECL(expected2,int,8,8) [] = { 0xf2, 0xf2, 0xf2, 0xf2, |
| 122 |
0xf2, 0xf2, 0xf2, 0xf2 }; |
| 123 |
VECT_VAR_DECL(expected2,int,16,4) [] = { 0xfff2, 0xfff2, 0xfff2, 0xfff2 }; |
| 124 |
VECT_VAR_DECL(expected2,int,32,2) [] = { 0xfffffff2, 0xfffffff2 }; |
| 125 |
VECT_VAR_DECL(expected2,int,64,1) [] = { 0xfffffffffffffff2 }; |
| 126 |
VECT_VAR_DECL(expected2,uint,8,8) [] = { 0xf2, 0xf2, 0xf2, 0xf2, |
| 127 |
0xf2, 0xf2, 0xf2, 0xf2 }; |
| 128 |
VECT_VAR_DECL(expected2,uint,16,4) [] = { 0xfff2, 0xfff2, 0xfff2, 0xfff2 }; |
| 129 |
VECT_VAR_DECL(expected2,uint,32,2) [] = { 0xfffffff2, 0xfffffff2 }; |
| 130 |
VECT_VAR_DECL(expected2,uint,64,1) [] = { 0xfffffffffffffff2 }; |
| 131 |
VECT_VAR_DECL(expected2,poly,8,8) [] = { 0xf2, 0xf2, 0xf2, 0xf2, |
| 132 |
0xf2, 0xf2, 0xf2, 0xf2 }; |
| 133 |
VECT_VAR_DECL(expected2,poly,16,4) [] = { 0xfff2, 0xfff2, 0xfff2, 0xfff2 }; |
| 134 |
#if MFLOAT8_SUPPORTED |
| 135 |
VECT_VAR_DECL(expected2,hmfloat,8,8) [] = { 0xf2, 0xf2, 0xf2, 0xf2, |
| 136 |
0xf2, 0xf2, 0xf2, 0xf2 }; |
| 137 |
#endif |
| 138 |
VECT_VAR_DECL(expected2,hfloat,16,4) [] = { 0xcb00, 0xcb00, 0xcb00, 0xcb00 }; |
| 139 |
VECT_VAR_DECL(expected2,hfloat,32,2) [] = { 0xc1600000, 0xc1600000 }; |
| 140 |
VECT_VAR_DECL(expected2,int,8,16) [] = { 0xf2, 0xf2, 0xf2, 0xf2, |
| 141 |
0xf2, 0xf2, 0xf2, 0xf2, |
| 142 |
0xf2, 0xf2, 0xf2, 0xf2, |
| 143 |
0xf2, 0xf2, 0xf2, 0xf2 }; |
| 144 |
VECT_VAR_DECL(expected2,int,16,8) [] = { 0xfff2, 0xfff2, 0xfff2, 0xfff2, |
| 145 |
0xfff2, 0xfff2, 0xfff2, 0xfff2 }; |
| 146 |
VECT_VAR_DECL(expected2,int,32,4) [] = { 0xfffffff2, 0xfffffff2, |
| 147 |
0xfffffff2, 0xfffffff2 }; |
| 148 |
VECT_VAR_DECL(expected2,int,64,2) [] = { 0xfffffffffffffff2, |
| 149 |
0xfffffffffffffff2 }; |
| 150 |
VECT_VAR_DECL(expected2,uint,8,16) [] = { 0xf2, 0xf2, 0xf2, 0xf2, |
| 151 |
0xf2, 0xf2, 0xf2, 0xf2, |
| 152 |
0xf2, 0xf2, 0xf2, 0xf2, |
| 153 |
0xf2, 0xf2, 0xf2, 0xf2 }; |
| 154 |
VECT_VAR_DECL(expected2,uint,16,8) [] = { 0xfff2, 0xfff2, 0xfff2, 0xfff2, |
| 155 |
0xfff2, 0xfff2, 0xfff2, 0xfff2 }; |
| 156 |
VECT_VAR_DECL(expected2,uint,32,4) [] = { 0xfffffff2, 0xfffffff2, |
| 157 |
0xfffffff2, 0xfffffff2 }; |
| 158 |
VECT_VAR_DECL(expected2,uint,64,2) [] = { 0xfffffffffffffff2, |
| 159 |
0xfffffffffffffff2 }; |
| 160 |
VECT_VAR_DECL(expected2,poly,8,16) [] = { 0xf2, 0xf2, 0xf2, 0xf2, |
| 161 |
0xf2, 0xf2, 0xf2, 0xf2, |
| 162 |
0xf2, 0xf2, 0xf2, 0xf2, |
| 163 |
0xf2, 0xf2, 0xf2, 0xf2 }; |
| 164 |
VECT_VAR_DECL(expected2,poly,16,8) [] = { 0xfff2, 0xfff2, 0xfff2, 0xfff2, |
| 165 |
0xfff2, 0xfff2, 0xfff2, 0xfff2 }; |
| 166 |
#if MFLOAT8_SUPPORTED |
| 167 |
VECT_VAR_DECL(expected2,hmfloat,8,16) [] = { 0xf2, 0xf2, 0xf2, 0xf2, |
| 168 |
0xf2, 0xf2, 0xf2, 0xf2, |
| 169 |
0xf2, 0xf2, 0xf2, 0xf2, |
| 170 |
0xf2, 0xf2, 0xf2, 0xf2 }; |
| 171 |
#endif |
| 172 |
VECT_VAR_DECL(expected2,hfloat,16,8) [] = { 0xcb00, 0xcb00, 0xcb00, 0xcb00, |
| 173 |
0xcb00, 0xcb00, 0xcb00, 0xcb00 }; |
| 174 |
VECT_VAR_DECL(expected2,hfloat,32,4) [] = { 0xc1600000, 0xc1600000, |
| 175 |
0xc1600000, 0xc1600000 }; |
| 176 |
|
| 177 |
#define TEST_MSG "VLD1_DUP/VLD1_DUPQ" |
| 178 |
void exec_vld1_dup (void) |
| 179 |
{ |
| 180 |
int i; |
| 181 |
|
| 182 |
/* Fill vector with buffer item #i. */ |
| 183 |
#define TEST_VLD1_DUP(VAR, BUF, Q, T1, T2, W, N) \ |
| 184 |
VECT_VAR(VAR, T1, W, N) = \ |
| 185 |
vld1##Q##_dup_##T2##W(&VECT_VAR(BUF, T1, W, N)[i]); \ |
| 186 |
vst1##Q##_##T2##W(VECT_VAR(result, T1, W, N), VECT_VAR(VAR, T1, W, N)) |
| 187 |
|
| 188 |
DECL_VARIABLE_ALL_VARIANTS(vector); |
| 189 |
|
| 190 |
/* Try to read different places from the input buffer. */ |
| 191 |
for (i=0; i<3; i++) { |
| 192 |
clean_results (); |
| 193 |
|
| 194 |
TEST_MACRO_ALL_VARIANTS_2_5(TEST_VLD1_DUP, vector, buffer_dup); |
| 195 |
|
| 196 |
#if MFLOAT8_SUPPORTED |
| 197 |
TEST_VLD1_DUP(vector, buffer_dup, , mfloat, mf, 8, 8); |
| 198 |
TEST_VLD1_DUP(vector, buffer_dup, q, mfloat, mf, 8, 16); |
| 199 |
#endif |
| 200 |
#if defined (__ARM_FP16_FORMAT_IEEE) || defined (__ARM_FP16_FORMAT_ALTERNATIVE) |
| 201 |
TEST_VLD1_DUP(vector, buffer_dup, , float, f, 16, 4); |
| 202 |
TEST_VLD1_DUP(vector, buffer_dup, q, float, f, 16, 8); |
| 203 |
#endif |
| 204 |
TEST_VLD1_DUP(vector, buffer_dup, , float, f, 32, 2); |
| 205 |
TEST_VLD1_DUP(vector, buffer_dup, q, float, f, 32, 4); |
| 206 |
|
| 207 |
switch (i) { |
| 208 |
case 0: |
| 209 |
CHECK_RESULTS_NAMED (TEST_MSG, expected0, ""); |
| 210 |
break; |
| 211 |
case 1: |
| 212 |
CHECK_RESULTS_NAMED (TEST_MSG, expected1, ""); |
| 213 |
break; |
| 214 |
case 2: |
| 215 |
CHECK_RESULTS_NAMED (TEST_MSG, expected2, ""); |
| 216 |
break; |
| 217 |
default: |
| 218 |
abort(); |
| 219 |
} |
| 220 |
} |
| 221 |
} |
| 222 |
|
| 223 |
int main (void) |
| 224 |
{ |
| 225 |
exec_vld1_dup (); |
| 226 |
return 0; |
| 227 |
} |