diff options
author | Diego Biurrun <diego@biurrun.de> | 2013-12-23 19:48:43 +0100 |
---|---|---|
committer | Diego Biurrun <diego@biurrun.de> | 2014-05-29 06:41:15 -0700 |
commit | 65d5d5865845f057cc6530a8d0f34db952d9009c (patch) | |
tree | 87f2ff44f58d550ae4076e741a4c6b3564d98bc0 /libavcodec/x86 | |
parent | 2ea2612df508abdd1f97c6a6ef56275a52c5c41e (diff) |
dsputil: Move SVQ1 encoding specific bits into svq1enc
Diffstat (limited to 'libavcodec/x86')
-rw-r--r-- | libavcodec/x86/Makefile | 1 | ||||
-rw-r--r-- | libavcodec/x86/dsputilenc_mmx.c | 36 | ||||
-rw-r--r-- | libavcodec/x86/svq1enc_mmx.c | 73 |
3 files changed, 74 insertions, 36 deletions
diff --git a/libavcodec/x86/Makefile b/libavcodec/x86/Makefile index fef98a5e40..8830a22a8f 100644 --- a/libavcodec/x86/Makefile +++ b/libavcodec/x86/Makefile @@ -51,6 +51,7 @@ MMX-OBJS-$(CONFIG_HPELDSP) += x86/fpel_mmx.o \ x86/hpeldsp_mmx.o MMX-OBJS-$(CONFIG_HUFFYUVDSP) += x86/huffyuvdsp_mmx.o +MMX-OBJS-$(CONFIG_SVQ1_ENCODER) += x86/svq1enc_mmx.o MMX-OBJS-$(CONFIG_VC1_DECODER) += x86/vc1dsp_mmx.o YASM-OBJS += x86/deinterlace.o \ diff --git a/libavcodec/x86/dsputilenc_mmx.c b/libavcodec/x86/dsputilenc_mmx.c index 79066a74e5..81c9d137f2 100644 --- a/libavcodec/x86/dsputilenc_mmx.c +++ b/libavcodec/x86/dsputilenc_mmx.c @@ -805,40 +805,6 @@ DCT_SAD_FUNC(ssse3) #undef HSUM #undef DCT_SAD -static int ssd_int8_vs_int16_mmx(const int8_t *pix1, const int16_t *pix2, - int size) -{ - int sum; - x86_reg i = size; - - __asm__ volatile ( - "pxor %%mm4, %%mm4 \n" - "1: \n" - "sub $8, %0 \n" - "movq (%2, %0), %%mm2 \n" - "movq (%3, %0, 2), %%mm0 \n" - "movq 8(%3, %0, 2), %%mm1 \n" - "punpckhbw %%mm2, %%mm3 \n" - "punpcklbw %%mm2, %%mm2 \n" - "psraw $8, %%mm3 \n" - "psraw $8, %%mm2 \n" - "psubw %%mm3, %%mm1 \n" - "psubw %%mm2, %%mm0 \n" - "pmaddwd %%mm1, %%mm1 \n" - "pmaddwd %%mm0, %%mm0 \n" - "paddd %%mm1, %%mm4 \n" - "paddd %%mm0, %%mm4 \n" - "jg 1b \n" - "movq %%mm4, %%mm3 \n" - "psrlq $32, %%mm3 \n" - "paddd %%mm3, %%mm4 \n" - "movd %%mm4, %1 \n" - : "+r" (i), "=r" (sum) - : "r" (pix1), "r" (pix2)); - - return sum; -} - #define PHADDD(a, t) \ "movq " #a ", " #t " \n\t" \ "psrlq $32, " #a " \n\t" \ @@ -958,8 +924,6 @@ av_cold void ff_dsputilenc_init_mmx(DSPContext *c, AVCodecContext *avctx, c->try_8x8basis = try_8x8basis_mmx; } c->add_8x8basis = add_8x8basis_mmx; - - c->ssd_int8_vs_int16 = ssd_int8_vs_int16_mmx; } if (INLINE_AMD3DNOW(cpu_flags)) { diff --git a/libavcodec/x86/svq1enc_mmx.c b/libavcodec/x86/svq1enc_mmx.c new file mode 100644 index 0000000000..02b0a84b8c --- /dev/null +++ b/libavcodec/x86/svq1enc_mmx.c @@ -0,0 +1,73 @@ +/* + * 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 "libavutil/attributes.h" +#include "libavutil/cpu.h" +#include "libavutil/x86/asm.h" +#include "libavutil/x86/cpu.h" +#include "libavcodec/svq1enc.h" + +#if HAVE_INLINE_ASM + +static int ssd_int8_vs_int16_mmx(const int8_t *pix1, const int16_t *pix2, + int size) +{ + int sum; + x86_reg i = size; + + __asm__ volatile ( + "pxor %%mm4, %%mm4 \n" + "1: \n" + "sub $8, %0 \n" + "movq (%2, %0), %%mm2 \n" + "movq (%3, %0, 2), %%mm0 \n" + "movq 8(%3, %0, 2), %%mm1 \n" + "punpckhbw %%mm2, %%mm3 \n" + "punpcklbw %%mm2, %%mm2 \n" + "psraw $8, %%mm3 \n" + "psraw $8, %%mm2 \n" + "psubw %%mm3, %%mm1 \n" + "psubw %%mm2, %%mm0 \n" + "pmaddwd %%mm1, %%mm1 \n" + "pmaddwd %%mm0, %%mm0 \n" + "paddd %%mm1, %%mm4 \n" + "paddd %%mm0, %%mm4 \n" + "jg 1b \n" + "movq %%mm4, %%mm3 \n" + "psrlq $32, %%mm3 \n" + "paddd %%mm3, %%mm4 \n" + "movd %%mm4, %1 \n" + : "+r" (i), "=r" (sum) + : "r" (pix1), "r" (pix2)); + + return sum; +} + +#endif /* HAVE_INLINE_ASM */ + +av_cold void ff_svq1enc_init_x86(SVQ1EncContext *c) +{ +#if HAVE_INLINE_ASM + int cpu_flags = av_get_cpu_flags(); + + if (INLINE_MMX(cpu_flags)) { + c->ssd_int8_vs_int16 = ssd_int8_vs_int16_mmx; + } +#endif /* HAVE_INLINE_ASM */ +} |