Details | Last modification | View Log | RSS feed
| Rev | Author | Line No. | Line |
|---|---|---|---|
| 6148 | serge | 1 | /* |
| 2 | * Copyright (c) 2008 Siarhei Siamashka |
||
| 3 | * |
||
| 4 | * This file is part of FFmpeg |
||
| 5 | * |
||
| 6 | * FFmpeg is free software; you can redistribute it and/or |
||
| 7 | * modify it under the terms of the GNU Lesser General Public |
||
| 8 | * License as published by the Free Software Foundation; either |
||
| 9 | * version 2.1 of the License, or (at your option) any later version. |
||
| 10 | * |
||
| 11 | * FFmpeg is distributed in the hope that it will be useful, |
||
| 12 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
||
| 13 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
||
| 14 | * Lesser General Public License for more details. |
||
| 15 | * |
||
| 16 | * You should have received a copy of the GNU Lesser General Public |
||
| 17 | * License along with FFmpeg; if not, write to the Free Software |
||
| 18 | * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
||
| 19 | */ |
||
| 20 | |||
| 21 | #include "config.h" |
||
| 22 | #include "asm.S" |
||
| 23 | |||
| 24 | /** |
||
| 25 | * Assume that len is a positive number and is multiple of 8 |
||
| 26 | */ |
||
| 27 | @ void ff_vector_fmul_vfp(float *dst, const float *src0, const float *src1, int len) |
||
| 28 | function ff_vector_fmul_vfp, export=1 |
||
| 29 | vpush {d8-d15} |
||
| 30 | fmrx r12, fpscr |
||
| 31 | orr r12, r12, #(3 << 16) /* set vector size to 4 */ |
||
| 32 | fmxr fpscr, r12 |
||
| 33 | |||
| 34 | vldmia r1!, {s0-s3} |
||
| 35 | vldmia r2!, {s8-s11} |
||
| 36 | vldmia r1!, {s4-s7} |
||
| 37 | vldmia r2!, {s12-s15} |
||
| 38 | vmul.f32 s8, s0, s8 |
||
| 39 | 1: |
||
| 40 | subs r3, r3, #16 |
||
| 41 | vmul.f32 s12, s4, s12 |
||
| 42 | itttt ge |
||
| 43 | vldmiage r1!, {s16-s19} |
||
| 44 | vldmiage r2!, {s24-s27} |
||
| 45 | vldmiage r1!, {s20-s23} |
||
| 46 | vldmiage r2!, {s28-s31} |
||
| 47 | it ge |
||
| 48 | vmulge.f32 s24, s16, s24 |
||
| 49 | vstmia r0!, {s8-s11} |
||
| 50 | vstmia r0!, {s12-s15} |
||
| 51 | it ge |
||
| 52 | vmulge.f32 s28, s20, s28 |
||
| 53 | itttt gt |
||
| 54 | vldmiagt r1!, {s0-s3} |
||
| 55 | vldmiagt r2!, {s8-s11} |
||
| 56 | vldmiagt r1!, {s4-s7} |
||
| 57 | vldmiagt r2!, {s12-s15} |
||
| 58 | ittt ge |
||
| 59 | vmulge.f32 s8, s0, s8 |
||
| 60 | vstmiage r0!, {s24-s27} |
||
| 61 | vstmiage r0!, {s28-s31} |
||
| 62 | bgt 1b |
||
| 63 | |||
| 64 | bic r12, r12, #(7 << 16) /* set vector size back to 1 */ |
||
| 65 | fmxr fpscr, r12 |
||
| 66 | vpop {d8-d15} |
||
| 67 | bx lr |
||
| 68 | endfunc |
||
| 69 | |||
| 70 | /** |
||
| 71 | * ARM VFP optimized implementation of 'vector_fmul_reverse_c' function. |
||
| 72 | * Assume that len is a positive number and is multiple of 8 |
||
| 73 | */ |
||
| 74 | @ void ff_vector_fmul_reverse_vfp(float *dst, const float *src0, |
||
| 75 | @ const float *src1, int len) |
||
| 76 | function ff_vector_fmul_reverse_vfp, export=1 |
||
| 77 | vpush {d8-d15} |
||
| 78 | add r2, r2, r3, lsl #2 |
||
| 79 | vldmdb r2!, {s0-s3} |
||
| 80 | vldmia r1!, {s8-s11} |
||
| 81 | vldmdb r2!, {s4-s7} |
||
| 82 | vldmia r1!, {s12-s15} |
||
| 83 | vmul.f32 s8, s3, s8 |
||
| 84 | vmul.f32 s9, s2, s9 |
||
| 85 | vmul.f32 s10, s1, s10 |
||
| 86 | vmul.f32 s11, s0, s11 |
||
| 87 | 1: |
||
| 88 | subs r3, r3, #16 |
||
| 89 | it ge |
||
| 90 | vldmdbge r2!, {s16-s19} |
||
| 91 | vmul.f32 s12, s7, s12 |
||
| 92 | it ge |
||
| 93 | vldmiage r1!, {s24-s27} |
||
| 94 | vmul.f32 s13, s6, s13 |
||
| 95 | it ge |
||
| 96 | vldmdbge r2!, {s20-s23} |
||
| 97 | vmul.f32 s14, s5, s14 |
||
| 98 | it ge |
||
| 99 | vldmiage r1!, {s28-s31} |
||
| 100 | vmul.f32 s15, s4, s15 |
||
| 101 | it ge |
||
| 102 | vmulge.f32 s24, s19, s24 |
||
| 103 | it gt |
||
| 104 | vldmdbgt r2!, {s0-s3} |
||
| 105 | it ge |
||
| 106 | vmulge.f32 s25, s18, s25 |
||
| 107 | vstmia r0!, {s8-s13} |
||
| 108 | it ge |
||
| 109 | vmulge.f32 s26, s17, s26 |
||
| 110 | it gt |
||
| 111 | vldmiagt r1!, {s8-s11} |
||
| 112 | itt ge |
||
| 113 | vmulge.f32 s27, s16, s27 |
||
| 114 | vmulge.f32 s28, s23, s28 |
||
| 115 | it gt |
||
| 116 | vldmdbgt r2!, {s4-s7} |
||
| 117 | it ge |
||
| 118 | vmulge.f32 s29, s22, s29 |
||
| 119 | vstmia r0!, {s14-s15} |
||
| 120 | ittt ge |
||
| 121 | vmulge.f32 s30, s21, s30 |
||
| 122 | vmulge.f32 s31, s20, s31 |
||
| 123 | vmulge.f32 s8, s3, s8 |
||
| 124 | it gt |
||
| 125 | vldmiagt r1!, {s12-s15} |
||
| 126 | itttt ge |
||
| 127 | vmulge.f32 s9, s2, s9 |
||
| 128 | vmulge.f32 s10, s1, s10 |
||
| 129 | vstmiage r0!, {s24-s27} |
||
| 130 | vmulge.f32 s11, s0, s11 |
||
| 131 | it ge |
||
| 132 | vstmiage r0!, {s28-s31} |
||
| 133 | bgt 1b |
||
| 134 | |||
| 135 | vpop {d8-d15} |
||
| 136 | bx lr |
||
| 137 | endfunc |