comparison i386/dsputil_mmx.c @ 1984:ef919e9ef73e libavcodec

separate out put_signed_pixels_clamped() into its own function and implement an optimized MMX version of the function

1983:1205bf58c420

 * MMX optimization by Nick Kurshev <[e-mail]>
 */

#include "../dsputil.h"
#include "../simple_idct.h"

//#undef NDEBUG
//#include <assert.h>

extern const uint8_t ff_h263_loop_filter_strength[32];

	    "movq	%%mm2, (%0, %1)\n\t"
	    "movq	%%mm4, (%0, %1, 2)\n\t"
	    "movq	%%mm6, (%0, %2)\n\t"
	    ::"r" (pix), "r" (line_size), "r" (line_size*3), "r"(p)
	    :"memory");
}

void add_pixels_clamped_mmx(const DCTELEM *block, uint8_t *pixels, int line_size)
{
    const DCTELEM *p;

#ifdef CONFIG_ENCODERS
        c->get_pixels = get_pixels_mmx;
        c->diff_pixels = diff_pixels_mmx;
#endif //CONFIG_ENCODERS
        c->put_pixels_clamped = put_pixels_clamped_mmx;
        c->add_pixels_clamped = add_pixels_clamped_mmx;
        c->clear_blocks = clear_blocks_mmx;
#ifdef CONFIG_ENCODERS
        c->pix_sum = pix_sum16_mmx;
#endif //CONFIG_ENCODERS

1984:ef919e9ef73e

 * MMX optimization by Nick Kurshev <[e-mail]>
 */

#include "../dsputil.h"
#include "../simple_idct.h"
#include "mmx.h"

//#undef NDEBUG
//#include <assert.h>

extern const uint8_t ff_h263_loop_filter_strength[32];

	    "movq	%%mm2, (%0, %1)\n\t"
	    "movq	%%mm4, (%0, %1, 2)\n\t"
	    "movq	%%mm6, (%0, %2)\n\t"
	    ::"r" (pix), "r" (line_size), "r" (line_size*3), "r"(p)
	    :"memory");
}

void put_signed_pixels_clamped_mmx(const DCTELEM *block, uint8_t *pixels, int line_size)
{
    int i;
    unsigned char __align8 vector128[8] =
      { 0x80, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80, 0x80 };

    movq_m2r(*vector128, mm1);
    for (i = 0; i < 8; i++) {
        movq_m2r(*(block), mm0);
        packsswb_m2r(*(block + 4), mm0);
        block += 8;
        paddb_r2r(mm1, mm0);
        movq_r2m(mm0, *pixels);
        pixels += line_size;
    }
}

void add_pixels_clamped_mmx(const DCTELEM *block, uint8_t *pixels, int line_size)
{
    const DCTELEM *p;

#ifdef CONFIG_ENCODERS
        c->get_pixels = get_pixels_mmx;
        c->diff_pixels = diff_pixels_mmx;
#endif //CONFIG_ENCODERS
        c->put_pixels_clamped = put_pixels_clamped_mmx;
        c->put_signed_pixels_clamped = put_signed_pixels_clamped_mmx;
        c->add_pixels_clamped = add_pixels_clamped_mmx;
        c->clear_blocks = clear_blocks_mmx;
#ifdef CONFIG_ENCODERS
        c->pix_sum = pix_sum16_mmx;
#endif //CONFIG_ENCODERS