mpegvideoencdsp.c 5.13 KB
Newer Older
1
/*
2
 * This file is part of FFmpeg.
3
 *
4
 * FFmpeg is free software; you can redistribute it and/or
5 6 7 8
 * modify it under the terms of the GNU Lesser General Public
 * License as published by the Free Software Foundation; either
 * version 2.1 of the License, or (at your option) any later version.
 *
9
 * FFmpeg is distributed in the hope that it will be useful,
10 11 12 13 14
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * Lesser General Public License for more details.
 *
 * You should have received a copy of the GNU Lesser General Public
15
 * License along with FFmpeg; if not, write to the Free Software
16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33
 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 */

#include "config.h"
#include <stdint.h>
#if HAVE_ALTIVEC_H
#include <altivec.h>
#endif

#include "libavutil/attributes.h"
#include "libavutil/cpu.h"
#include "libavutil/ppc/cpu.h"
#include "libavutil/ppc/types_altivec.h"
#include "libavutil/ppc/util_altivec.h"
#include "libavcodec/mpegvideoencdsp.h"

#if HAVE_ALTIVEC

34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57
#if HAVE_VSX
static int pix_norm1_altivec(uint8_t *pix, int line_size)
{
    int i, s = 0;
    const vector unsigned int zero =
        (const vector unsigned int) vec_splat_u32(0);
    vector unsigned int sv = (vector unsigned int) vec_splat_u32(0);
    vector signed int sum;

    for (i = 0; i < 16; i++) {
        /* Read the potentially unaligned pixels. */
        //vector unsigned char pixl = vec_ld(0,  pix);
        //vector unsigned char pixr = vec_ld(15, pix);
        //vector unsigned char pixv = vec_perm(pixl, pixr, perm);
        vector unsigned char pixv = vec_vsx_ld(0,  pix);

        /* Square the values, and add them to our sum. */
        sv = vec_msum(pixv, pixv, sv);

        pix += line_size;
    }
    /* Sum up the four partial sums, and put the result into s. */
    sum = vec_sums((vector signed int) sv, (vector signed int) zero);
    sum = vec_splat(sum, 3);
58
    vec_ste(sum, 0, &s);
59 60 61
    return s;
}
#else
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
static int pix_norm1_altivec(uint8_t *pix, int line_size)
{
    int i, s = 0;
    const vector unsigned int zero =
        (const vector unsigned int) vec_splat_u32(0);
    vector unsigned char perm = vec_lvsl(0, pix);
    vector unsigned int sv = (vector unsigned int) vec_splat_u32(0);
    vector signed int sum;

    for (i = 0; i < 16; i++) {
        /* Read the potentially unaligned pixels. */
        vector unsigned char pixl = vec_ld(0,  pix);
        vector unsigned char pixr = vec_ld(15, pix);
        vector unsigned char pixv = vec_perm(pixl, pixr, perm);

        /* Square the values, and add them to our sum. */
        sv = vec_msum(pixv, pixv, sv);

        pix += line_size;
    }
    /* Sum up the four partial sums, and put the result into s. */
    sum = vec_sums((vector signed int) sv, (vector signed int) zero);
    sum = vec_splat(sum, 3);
    vec_ste(sum, 0, &s);

    return s;
}
89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111
#endif /* HAVE_VSX */

#if HAVE_VSX
static int pix_sum_altivec(uint8_t *pix, int line_size)
{
    int i, s;
    const vector unsigned int zero =
        (const vector unsigned int) vec_splat_u32(0);
    vector unsigned int sad = (vector unsigned int) vec_splat_u32(0);
    vector signed int sumdiffs;

    for (i = 0; i < 16; i++) {
        /* Read the potentially unaligned 16 pixels into t1. */
        //vector unsigned char pixl = vec_ld(0,  pix);
        //vector unsigned char pixr = vec_ld(15, pix);
        //vector unsigned char t1   = vec_perm(pixl, pixr, perm);
        vector unsigned char t1   = vec_vsx_ld(0,  pix);

        /* Add each 4 pixel group together and put 4 results into sad. */
        sad = vec_sum4s(t1, sad);

        pix += line_size;
    }
112

113 114 115
    /* Sum up the four partial sums, and put the result into s. */
    sumdiffs = vec_sums((vector signed int) sad, (vector signed int) zero);
    sumdiffs = vec_splat(sumdiffs, 3);
116
    vec_ste(sumdiffs, 0, &s);
117 118 119
    return s;
}
#else
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
static int pix_sum_altivec(uint8_t *pix, int line_size)
{
    int i, s;
    const vector unsigned int zero =
        (const vector unsigned int) vec_splat_u32(0);
    vector unsigned char perm = vec_lvsl(0, pix);
    vector unsigned int sad = (vector unsigned int) vec_splat_u32(0);
    vector signed int sumdiffs;

    for (i = 0; i < 16; i++) {
        /* Read the potentially unaligned 16 pixels into t1. */
        vector unsigned char pixl = vec_ld(0,  pix);
        vector unsigned char pixr = vec_ld(15, pix);
        vector unsigned char t1   = vec_perm(pixl, pixr, perm);

        /* Add each 4 pixel group together and put 4 results into sad. */
        sad = vec_sum4s(t1, sad);

        pix += line_size;
    }

    /* Sum up the four partial sums, and put the result into s. */
    sumdiffs = vec_sums((vector signed int) sad, (vector signed int) zero);
    sumdiffs = vec_splat(sumdiffs, 3);
    vec_ste(sumdiffs, 0, &s);

    return s;
}

149 150
#endif /* HAVE_VSX */

151 152 153 154 155 156 157 158 159 160 161 162 163
#endif /* HAVE_ALTIVEC */

av_cold void ff_mpegvideoencdsp_init_ppc(MpegvideoEncDSPContext *c,
                                         AVCodecContext *avctx)
{
#if HAVE_ALTIVEC
    if (!PPC_ALTIVEC(av_get_cpu_flags()))
        return;

    c->pix_norm1 = pix_norm1_altivec;
    c->pix_sum   = pix_sum_altivec;
#endif /* HAVE_ALTIVEC */
}