Commit 2571e8fe authored by James Almer's avatar James Almer Committed by Michael Niedermayer

alpha/pixblockdsp: move code out of dsputil

Signed-off-by: 's avatarJames Almer <jamrial@gmail.com>
Signed-off-by: 's avatarMichael Niedermayer <michaelni@gmx.at>
parent 81538470
......@@ -8,3 +8,4 @@ OBJS-$(CONFIG_IDCTDSP) += alpha/idctdsp_alpha.o \
alpha/idctdsp_alpha_asm.o \
alpha/simple_idct_alpha.o
OBJS-$(CONFIG_MPEGVIDEO) += alpha/mpegvideo_alpha.o
OBJS-$(CONFIG_PIXBLOCKDSP) += alpha/pixblockdsp_alpha.o
......@@ -98,13 +98,8 @@ void add_pixels_clamped_mvi(const int16_t *block, uint8_t *pixels,
av_cold void ff_dsputil_init_alpha(DSPContext *c, AVCodecContext *avctx)
{
const int high_bit_depth = avctx->bits_per_raw_sample > 8;
/* amask clears all bits that correspond to present features. */
if (amask(AMASK_MVI) == 0) {
if (!high_bit_depth)
c->get_pixels = get_pixels_mvi;
c->diff_pixels = diff_pixels_mvi;
c->sad[0] = pix_abs16x16_mvi_asm;
c->sad[1] = pix_abs8x8_mvi;
c->pix_abs[0][0] = pix_abs16x16_mvi_asm;
......
......@@ -22,10 +22,6 @@
#include <stddef.h>
#include <stdint.h>
void get_pixels_mvi(int16_t *restrict block,
const uint8_t *restrict pixels, int line_size);
void diff_pixels_mvi(int16_t *block, const uint8_t *s1, const uint8_t *s2,
int stride);
int pix_abs8x8_mvi(void *v, uint8_t *pix1, uint8_t *pix2, int line_size, int h);
int pix_abs16x16_mvi_asm(void *v, uint8_t *pix1, uint8_t *pix2, int line_size, int h);
int pix_abs16x16_x2_mvi(void *v, uint8_t *pix1, uint8_t *pix2, int line_size, int h);
......
......@@ -22,51 +22,6 @@
#include "dsputil_alpha.h"
#include "asm.h"
void get_pixels_mvi(int16_t *restrict block,
const uint8_t *restrict pixels, int line_size)
{
int h = 8;
do {
uint64_t p;
p = ldq(pixels);
stq(unpkbw(p), block);
stq(unpkbw(p >> 32), block + 4);
pixels += line_size;
block += 8;
} while (--h);
}
void diff_pixels_mvi(int16_t *block, const uint8_t *s1, const uint8_t *s2,
int stride) {
int h = 8;
uint64_t mask = 0x4040;
mask |= mask << 16;
mask |= mask << 32;
do {
uint64_t x, y, c, d, a;
uint64_t signs;
x = ldq(s1);
y = ldq(s2);
c = cmpbge(x, y);
d = x - y;
a = zap(mask, c); /* We use 0x4040404040404040 here... */
d += 4 * a; /* ...so we can use s4addq here. */
signs = zap(-1, c);
stq(unpkbw(d) | (unpkbw(signs) << 8), block);
stq(unpkbw(d >> 32) | (unpkbw(signs >> 32) << 8), block + 4);
s1 += stride;
s2 += stride;
block += 8;
} while (--h);
}
static inline uint64_t avg2(uint64_t a, uint64_t b)
{
return (a | b) - (((a ^ b) & BYTE_VEC(0xfe)) >> 1);
......
/*
* SIMD-optimized pixel operations
*
* This file is part of FFmpeg.
*
* FFmpeg 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.
*
* FFmpeg 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 FFmpeg; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
#include "libavutil/attributes.h"
#include "libavcodec/dsputil.h"
#include "libavcodec/pixblockdsp.h"
#include "asm.h"
static void get_pixels_mvi(int16_t *restrict block,
const uint8_t *restrict pixels, int line_size)
{
int h = 8;
do {
uint64_t p;
p = ldq(pixels);
stq(unpkbw(p), block);
stq(unpkbw(p >> 32), block + 4);
pixels += line_size;
block += 8;
} while (--h);
}
static void diff_pixels_mvi(int16_t *block, const uint8_t *s1, const uint8_t *s2,
int stride) {
int h = 8;
uint64_t mask = 0x4040;
mask |= mask << 16;
mask |= mask << 32;
do {
uint64_t x, y, c, d, a;
uint64_t signs;
x = ldq(s1);
y = ldq(s2);
c = cmpbge(x, y);
d = x - y;
a = zap(mask, c); /* We use 0x4040404040404040 here... */
d += 4 * a; /* ...so we can use s4addq here. */
signs = zap(-1, c);
stq(unpkbw(d) | (unpkbw(signs) << 8), block);
stq(unpkbw(d >> 32) | (unpkbw(signs >> 32) << 8), block + 4);
s1 += stride;
s2 += stride;
block += 8;
} while (--h);
}
av_cold void ff_pixblockdsp_init_alpha(PixblockDSPContext *c, AVCodecContext *avctx,
unsigned high_bit_depth)
{
if (amask(AMASK_MVI) == 0) {
if (!high_bit_depth)
c->get_pixels = get_pixels_mvi;
c->diff_pixels = diff_pixels_mvi;
}
}
......@@ -71,6 +71,8 @@ av_cold void ff_pixblockdsp_init(PixblockDSPContext *c, AVCodecContext *avctx)
break;
}
if (ARCH_ALPHA)
ff_pixblockdsp_init_alpha(c, avctx, high_bit_depth);
if (ARCH_ARM)
ff_pixblockdsp_init_arm(c, avctx, high_bit_depth);
if (ARCH_PPC)
......
......@@ -34,6 +34,8 @@ typedef struct PixblockDSPContext {
} PixblockDSPContext;
void ff_pixblockdsp_init(PixblockDSPContext *c, AVCodecContext *avctx);
void ff_pixblockdsp_init_alpha(PixblockDSPContext *c, AVCodecContext *avctx,
unsigned high_bit_depth);
void ff_pixblockdsp_init_arm(PixblockDSPContext *c, AVCodecContext *avctx,
unsigned high_bit_depth);
void ff_pixblockdsp_init_ppc(PixblockDSPContext *c, AVCodecContext *avctx,
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment