| 1 |
#include <arm_neon.h> |
| 2 |
#include "arm-neon-ref.h" |
| 3 |
#include "compute-ref-data.h" |
| 4 |
|
| 5 |
/* Expected results splitted in several chunks. */ |
| 6 |
/* Chunk 0. */ |
| 7 |
VECT_VAR_DECL(expected0,int,8,8) [] = { 0xf0, 0xf1, 0xf2, 0xf3, |
| 8 |
0xf4, 0xf5, 0xf6, 0xf7 }; |
| 9 |
VECT_VAR_DECL(expected0,int,16,4) [] = { 0xfff0, 0xfff1, |
| 10 |
0xfff2, 0xfff3 }; |
| 11 |
VECT_VAR_DECL(expected0,int,32,2) [] = { 0xfffffff0, 0xfffffff1 }; |
| 12 |
VECT_VAR_DECL(expected0,uint,8,8) [] = { 0xf0, 0xf1, 0xf2, 0xf3, |
| 13 |
0xf4, 0xf5, 0xf6, 0xf7 }; |
| 14 |
VECT_VAR_DECL(expected0,uint,16,4) [] = { 0xfff0, 0xfff1, |
| 15 |
0xfff2, 0xfff3 }; |
| 16 |
VECT_VAR_DECL(expected0,uint,32,2) [] = { 0xfffffff0, |
| 17 |
0xfffffff1 }; |
| 18 |
VECT_VAR_DECL(expected0,poly,8,8) [] = { 0xf0, 0xf1, 0xf2, 0xf3, |
| 19 |
0xf4, 0xf5, 0xf6, 0xf7 }; |
| 20 |
VECT_VAR_DECL(expected0,poly,16,4) [] = { 0xfff0, 0xfff1, |
| 21 |
0xfff2, 0xfff3 }; |
| 22 |
#if MFLOAT8_SUPPORTED |
| 23 |
VECT_VAR_DECL(expected0,hmfloat,8,8) [] = { 0xf0, 0xf1, 0xf2, 0xf3, |
| 24 |
0xf4, 0xf5, 0xf6, 0xf7 }; |
| 25 |
#endif |
| 26 |
#if defined (FP16_SUPPORTED) |
| 27 |
VECT_VAR_DECL (expected0, hfloat, 16, 4) [] = { 0xcc00, 0xcb80, |
| 28 |
0xcb00, 0xca80 }; |
| 29 |
#endif |
| 30 |
VECT_VAR_DECL(expected0,hfloat,32,2) [] = { 0xc1800000, 0xc1700000 }; |
| 31 |
VECT_VAR_DECL(expected0,int,8,16) [] = { 0xf0, 0xf1, 0xf2, 0xf3, |
| 32 |
0xf4, 0xf5, 0xf6, 0xf7, |
| 33 |
0xf8, 0xf9, 0xfa, 0xfb, |
| 34 |
0xfc, 0xfd, 0xfe, 0xff }; |
| 35 |
VECT_VAR_DECL(expected0,int,16,8) [] = { 0xfff0, 0xfff1, |
| 36 |
0xfff2, 0xfff3, |
| 37 |
0xfff4, 0xfff5, |
| 38 |
0xfff6, 0xfff7 }; |
| 39 |
VECT_VAR_DECL(expected0,int,32,4) [] = { 0xfffffff0, 0xfffffff1, |
| 40 |
0xfffffff2, 0xfffffff3 }; |
| 41 |
VECT_VAR_DECL(expected0,uint,8,16) [] = { 0xf0, 0xf1, 0xf2, 0xf3, |
| 42 |
0xf4, 0xf5, 0xf6, 0xf7, |
| 43 |
0xf8, 0xf9, 0xfa, 0xfb, |
| 44 |
0xfc, 0xfd, 0xfe, 0xff }; |
| 45 |
VECT_VAR_DECL(expected0,uint,16,8) [] = { 0xfff0, 0xfff1, |
| 46 |
0xfff2, 0xfff3, |
| 47 |
0xfff4, 0xfff5, |
| 48 |
0xfff6, 0xfff7 }; |
| 49 |
VECT_VAR_DECL(expected0,uint,32,4) [] = { 0xfffffff0, 0xfffffff1, |
| 50 |
0xfffffff2, 0xfffffff3 }; |
| 51 |
VECT_VAR_DECL(expected0,poly,8,16) [] = { 0xf0, 0xf1, 0xf2, 0xf3, |
| 52 |
0xf4, 0xf5, 0xf6, 0xf7, |
| 53 |
0xf8, 0xf9, 0xfa, 0xfb, |
| 54 |
0xfc, 0xfd, 0xfe, 0xff }; |
| 55 |
VECT_VAR_DECL(expected0,poly,16,8) [] = { 0xfff0, 0xfff1, |
| 56 |
0xfff2, 0xfff3, |
| 57 |
0xfff4, 0xfff5, |
| 58 |
0xfff6, 0xfff7 }; |
| 59 |
#if MFLOAT8_SUPPORTED |
| 60 |
VECT_VAR_DECL(expected0,hmfloat,8,16) [] = { 0xf0, 0xf1, 0xf2, 0xf3, |
| 61 |
0xf4, 0xf5, 0xf6, 0xf7, |
| 62 |
0xf8, 0xf9, 0xfa, 0xfb, |
| 63 |
0xfc, 0xfd, 0xfe, 0xff }; |
| 64 |
#endif |
| 65 |
#if defined (FP16_SUPPORTED) |
| 66 |
VECT_VAR_DECL (expected0, hfloat, 16, 8) [] = { 0xcc00, 0xcb80, |
| 67 |
0xcb00, 0xca80, |
| 68 |
0xca00, 0xc980, |
| 69 |
0xc900, 0xc880 }; |
| 70 |
#endif |
| 71 |
VECT_VAR_DECL(expected0,hfloat,32,4) [] = { 0xc1800000, 0xc1700000, |
| 72 |
0xc1600000, 0xc1500000 }; |
| 73 |
|
| 74 |
/* Chunk 1. */ |
| 75 |
VECT_VAR_DECL(expected1,int,8,8) [] = { 0x11, 0x11, 0x11, 0x11, |
| 76 |
0x11, 0x11, 0x11, 0x11 }; |
| 77 |
VECT_VAR_DECL(expected1,int,16,4) [] = { 0x22, 0x22, 0x22, 0x22 }; |
| 78 |
VECT_VAR_DECL(expected1,int,32,2) [] = { 0x33, 0x33 }; |
| 79 |
VECT_VAR_DECL(expected1,uint,8,8) [] = { 0x55, 0x55, 0x55, 0x55, |
| 80 |
0x55, 0x55, 0x55, 0x55 }; |
| 81 |
VECT_VAR_DECL(expected1,uint,16,4) [] = { 0x66, 0x66, 0x66, 0x66 }; |
| 82 |
VECT_VAR_DECL(expected1,uint,32,2) [] = { 0x77, 0x77 }; |
| 83 |
VECT_VAR_DECL(expected1,poly,8,8) [] = { 0x55, 0x55, 0x55, 0x55, |
| 84 |
0x55, 0x55, 0x55, 0x55 }; |
| 85 |
VECT_VAR_DECL(expected1,poly,16,4) [] = { 0x66, 0x66, 0x66, 0x66 }; |
| 86 |
#if MFLOAT8_SUPPORTED |
| 87 |
VECT_VAR_DECL(expected1,hmfloat,8,8) [] = { 0xaa, 0xaa, 0xaa, 0xaa, |
| 88 |
0xaa, 0xaa, 0xaa, 0xaa }; |
| 89 |
#endif |
| 90 |
#if defined (FP16_SUPPORTED) |
| 91 |
VECT_VAR_DECL (expected1, hfloat, 16, 4) [] = { 0x4b4d, 0x4b4d, |
| 92 |
0x4b4d, 0x4b4d }; |
| 93 |
#endif |
| 94 |
VECT_VAR_DECL(expected1,hfloat,32,2) [] = { 0x42066666, 0x42066666 }; |
| 95 |
VECT_VAR_DECL(expected1,int,8,16) [] = { 0x11, 0x11, 0x11, 0x11, |
| 96 |
0x11, 0x11, 0x11, 0x11, |
| 97 |
0x11, 0x11, 0x11, 0x11, |
| 98 |
0x11, 0x11, 0x11, 0x11 }; |
| 99 |
VECT_VAR_DECL(expected1,int,16,8) [] = { 0x22, 0x22, 0x22, 0x22, |
| 100 |
0x22, 0x22, 0x22, 0x22 }; |
| 101 |
VECT_VAR_DECL(expected1,int,32,4) [] = { 0x33, 0x33, 0x33, 0x33 }; |
| 102 |
VECT_VAR_DECL(expected1,uint,8,16) [] = { 0x55, 0x55, 0x55, 0x55, |
| 103 |
0x55, 0x55, 0x55, 0x55, |
| 104 |
0x55, 0x55, 0x55, 0x55, |
| 105 |
0x55, 0x55, 0x55, 0x55 }; |
| 106 |
VECT_VAR_DECL(expected1,uint,16,8) [] = { 0x66, 0x66, 0x66, 0x66, |
| 107 |
0x66, 0x66, 0x66, 0x66 }; |
| 108 |
VECT_VAR_DECL(expected1,uint,32,4) [] = { 0x77, 0x77, 0x77, 0x77 }; |
| 109 |
VECT_VAR_DECL(expected1,poly,8,16) [] = { 0x55, 0x55, 0x55, 0x55, |
| 110 |
0x55, 0x55, 0x55, 0x55, |
| 111 |
0x55, 0x55, 0x55, 0x55, |
| 112 |
0x55, 0x55, 0x55, 0x55 }; |
| 113 |
VECT_VAR_DECL(expected1,poly,16,8) [] = { 0x66, 0x66, 0x66, 0x66, |
| 114 |
0x66, 0x66, 0x66, 0x66 }; |
| 115 |
#if MFLOAT8_SUPPORTED |
| 116 |
VECT_VAR_DECL(expected1,hmfloat,8,16) [] = { 0xbc, 0xbc, 0xbc, 0xbc, |
| 117 |
0xbc, 0xbc, 0xbc, 0xbc, |
| 118 |
0xbc, 0xbc, 0xbc, 0xbc, |
| 119 |
0xbc, 0xbc, 0xbc, 0xbc }; |
| 120 |
#endif |
| 121 |
#if defined (FP16_SUPPORTED) |
| 122 |
VECT_VAR_DECL (expected1, hfloat, 16, 8) [] = { 0x4b4d, 0x4b4d, |
| 123 |
0x4b4d, 0x4b4d, |
| 124 |
0x4b4d, 0x4b4d, |
| 125 |
0x4b4d, 0x4b4d }; |
| 126 |
#endif |
| 127 |
VECT_VAR_DECL(expected1,hfloat,32,4) [] = { 0x42073333, 0x42073333, |
| 128 |
0x42073333, 0x42073333 }; |
| 129 |
|
| 130 |
#define INSN_NAME vuzp |
| 131 |
#define TEST_MSG "VUZP/VUZPQ" |
| 132 |
|
| 133 |
#include "vshuffle.inc" |