vpx/vp8/common/recon.c
John Koleszar 19638c2309 arm: move unrolled loops back to generic code
Some of the ARM functions differed from their generic counterparts
only by unrolling their loops. Since this change may be useful
on other platforms, or might even supercede the looped version
in the generic case, move it back to the generic file.

This code is left under #if ARCH_ARM for now, but it may be worth
considering a different (possibly new) conditional for these. If
it turns out that this should be runtime selectable, these
functions will have to move to the RTCD infrastructure. Don't want
to take that step at this time without more profile data.

Change-Id: I4612fdbc606fbebba4971a690fb743ad184ff15f
2010-10-26 09:51:35 -04:00

182 lines
4.1 KiB
C

/*
* Copyright (c) 2010 The WebM project authors. All Rights Reserved.
*
* Use of this source code is governed by a BSD-style license
* that can be found in the LICENSE file in the root of the source
* tree. An additional intellectual property rights grant can be found
* in the file PATENTS. All contributing project authors may
* be found in the AUTHORS file in the root of the source tree.
*/
#include "vpx_ports/config.h"
#include "recon.h"
#include "blockd.h"
void vp8_recon_b_c
(
unsigned char *pred_ptr,
short *diff_ptr,
unsigned char *dst_ptr,
int stride
)
{
int r, c;
for (r = 0; r < 4; r++)
{
for (c = 0; c < 4; c++)
{
int a = diff_ptr[c] + pred_ptr[c] ;
if (a < 0)
a = 0;
if (a > 255)
a = 255;
dst_ptr[c] = (unsigned char) a ;
}
dst_ptr += stride;
diff_ptr += 16;
pred_ptr += 16;
}
}
void vp8_recon4b_c
(
unsigned char *pred_ptr,
short *diff_ptr,
unsigned char *dst_ptr,
int stride
)
{
int r, c;
for (r = 0; r < 4; r++)
{
for (c = 0; c < 16; c++)
{
int a = diff_ptr[c] + pred_ptr[c] ;
if (a < 0)
a = 0;
if (a > 255)
a = 255;
dst_ptr[c] = (unsigned char) a ;
}
dst_ptr += stride;
diff_ptr += 16;
pred_ptr += 16;
}
}
void vp8_recon2b_c
(
unsigned char *pred_ptr,
short *diff_ptr,
unsigned char *dst_ptr,
int stride
)
{
int r, c;
for (r = 0; r < 4; r++)
{
for (c = 0; c < 8; c++)
{
int a = diff_ptr[c] + pred_ptr[c] ;
if (a < 0)
a = 0;
if (a > 255)
a = 255;
dst_ptr[c] = (unsigned char) a ;
}
dst_ptr += stride;
diff_ptr += 8;
pred_ptr += 8;
}
}
void vp8_recon16x16mby(const vp8_recon_rtcd_vtable_t *rtcd, MACROBLOCKD *x)
{
#if ARCH_ARM
BLOCKD *b = &x->block[0];
RECON_INVOKE(rtcd, recon4)(b->predictor, b->diff, *(b->base_dst) + b->dst, b->dst_stride);
//b = &x->block[4];
b += 4;
RECON_INVOKE(rtcd, recon4)(b->predictor, b->diff, *(b->base_dst) + b->dst, b->dst_stride);
//b = &x->block[8];
b += 4;
RECON_INVOKE(rtcd, recon4)(b->predictor, b->diff, *(b->base_dst) + b->dst, b->dst_stride);
//b = &x->block[12];
b += 4;
RECON_INVOKE(rtcd, recon4)(b->predictor, b->diff, *(b->base_dst) + b->dst, b->dst_stride);
#else
int i;
for (i = 0; i < 16; i += 4)
{
BLOCKD *b = &x->block[i];
RECON_INVOKE(rtcd, recon4)(b->predictor, b->diff, *(b->base_dst) + b->dst, b->dst_stride);
}
#endif
}
void vp8_recon16x16mb(const vp8_recon_rtcd_vtable_t *rtcd, MACROBLOCKD *x)
{
#if ARCH_ARM
BLOCKD *b = &x->block[0];
RECON_INVOKE(rtcd, recon4)(b->predictor, b->diff, *(b->base_dst) + b->dst, b->dst_stride);
b += 4;
RECON_INVOKE(rtcd, recon4)(b->predictor, b->diff, *(b->base_dst) + b->dst, b->dst_stride);
b += 4;
RECON_INVOKE(rtcd, recon4)(b->predictor, b->diff, *(b->base_dst) + b->dst, b->dst_stride);
b += 4;
RECON_INVOKE(rtcd, recon4)(b->predictor, b->diff, *(b->base_dst) + b->dst, b->dst_stride);
b += 4;
//b = &x->block[16];
RECON_INVOKE(rtcd, recon2)(b->predictor, b->diff, *(b->base_dst) + b->dst, b->dst_stride);
b++;
b++;
RECON_INVOKE(rtcd, recon2)(b->predictor, b->diff, *(b->base_dst) + b->dst, b->dst_stride);
b++;
b++;
RECON_INVOKE(rtcd, recon2)(b->predictor, b->diff, *(b->base_dst) + b->dst, b->dst_stride);
b++;
b++;
RECON_INVOKE(rtcd, recon2)(b->predictor, b->diff, *(b->base_dst) + b->dst, b->dst_stride);
#else
int i;
for (i = 0; i < 16; i += 4)
{
BLOCKD *b = &x->block[i];
RECON_INVOKE(rtcd, recon4)(b->predictor, b->diff, *(b->base_dst) + b->dst, b->dst_stride);
}
for (i = 16; i < 24; i += 2)
{
BLOCKD *b = &x->block[i];
RECON_INVOKE(rtcd, recon2)(b->predictor, b->diff, *(b->base_dst) + b->dst, b->dst_stride);
}
#endif
}