From c166148409fe8f0dbccef2fe684286a40ba1e37d Mon Sep 17 00:00:00 2001 From: Diego Biurrun Date: Mon, 30 Dec 2013 19:56:07 +0100 Subject: dsputil: Move pix_sum, pix_norm1, shrink function pointers to mpegvideoenc --- libavcodec/ppc/Makefile | 1 + libavcodec/ppc/dsputil_altivec.c | 60 ----------------------- libavcodec/ppc/mpegvideoencdsp.c | 103 +++++++++++++++++++++++++++++++++++++++ 3 files changed, 104 insertions(+), 60 deletions(-) create mode 100644 libavcodec/ppc/mpegvideoencdsp.c (limited to 'libavcodec/ppc') diff --git a/libavcodec/ppc/Makefile b/libavcodec/ppc/Makefile index ee0c18c09e..256c081dc1 100644 --- a/libavcodec/ppc/Makefile +++ b/libavcodec/ppc/Makefile @@ -13,6 +13,7 @@ OBJS-$(CONFIG_IDCTDSP) += ppc/idctdsp.o OBJS-$(CONFIG_MPEGAUDIODSP) += ppc/mpegaudiodsp_altivec.o OBJS-$(CONFIG_MPEGVIDEO) += ppc/mpegvideo_altivec.o \ ppc/mpegvideodsp.o +OBJS-$(CONFIG_MPEGVIDEOENC) += ppc/mpegvideoencdsp.o OBJS-$(CONFIG_VIDEODSP) += ppc/videodsp_ppc.o OBJS-$(CONFIG_VP3DSP) += ppc/vp3dsp_altivec.o diff --git a/libavcodec/ppc/dsputil_altivec.c b/libavcodec/ppc/dsputil_altivec.c index c3f90e91c4..efb4a18535 100644 --- a/libavcodec/ppc/dsputil_altivec.c +++ b/libavcodec/ppc/dsputil_altivec.c @@ -308,34 +308,6 @@ static int sad8_altivec(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, return s; } -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; -} - /* Sum of Squared Errors for an 8x8 block, AltiVec-enhanced. * It's the sad8_altivec code above w/ squaring added. */ static int sse8_altivec(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, @@ -430,35 +402,6 @@ static int sse16_altivec(MpegEncContext *v, uint8_t *pix1, uint8_t *pix2, return s; } -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; -} - static void get_pixels_altivec(int16_t *restrict block, const uint8_t *pixels, int line_size) { @@ -911,9 +854,6 @@ av_cold void ff_dsputil_init_altivec(DSPContext *c, AVCodecContext *avctx, c->sse[0] = sse16_altivec; c->sse[1] = sse8_altivec; - c->pix_norm1 = pix_norm1_altivec; - c->pix_sum = pix_sum_altivec; - c->diff_pixels = diff_pixels_altivec; if (!high_bit_depth) { diff --git a/libavcodec/ppc/mpegvideoencdsp.c b/libavcodec/ppc/mpegvideoencdsp.c new file mode 100644 index 0000000000..b5348e6abd --- /dev/null +++ b/libavcodec/ppc/mpegvideoencdsp.c @@ -0,0 +1,103 @@ +/* + * This file is part of Libav. + * + * Libav is free software; you can redistribute it and/or + * 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. + * + * Libav is distributed in the hope that it will be useful, + * 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 + * License along with Libav; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include "config.h" +#include +#if HAVE_ALTIVEC_H +#include +#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 + +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; +} + +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; +} + +#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 */ +} -- cgit v1.2.3