mirror of
https://github.com/webmproject/libwebp.git
synced 2025-07-13 06:24:27 +02:00
MIPS: dspr2: added optimization for EmitAlphaRGB
New dsp function: WebPDispatchAlpha() Change-Id: I48e539d22471279ec75185759bc68d18b127f716
This commit is contained in:
@ -10,6 +10,7 @@ commondir = $(includedir)/webp
|
||||
|
||||
COMMON_SOURCES =
|
||||
COMMON_SOURCES += alpha_processing.c
|
||||
COMMON_SOURCES += alpha_processing_mips_dsp_r2.c
|
||||
COMMON_SOURCES += cpu.c
|
||||
COMMON_SOURCES += dec.c
|
||||
COMMON_SOURCES += dec_clip_tables.c
|
||||
|
@ -284,25 +284,48 @@ static void ApplyAlphaMultiply_16b(uint8_t* rgba4444,
|
||||
#endif
|
||||
}
|
||||
|
||||
static int DispatchAlpha(const uint8_t* alpha, int alpha_stride,
|
||||
int width, int height,
|
||||
uint8_t* dst, int dst_stride) {
|
||||
uint32_t alpha_mask = 0xff;
|
||||
int i, j;
|
||||
|
||||
for (j = 0; j < height; ++j) {
|
||||
for (i = 0; i < width; ++i) {
|
||||
const uint32_t alpha_value = alpha[i];
|
||||
dst[4 * i] = alpha_value;
|
||||
alpha_mask &= alpha_value;
|
||||
}
|
||||
alpha += alpha_stride;
|
||||
dst += dst_stride;
|
||||
}
|
||||
|
||||
return (alpha_mask != 0xff);
|
||||
}
|
||||
|
||||
void (*WebPApplyAlphaMultiply)(uint8_t*, int, int, int, int);
|
||||
void (*WebPApplyAlphaMultiply4444)(uint8_t*, int, int, int);
|
||||
int (*WebPDispatchAlpha)(const uint8_t*, int, int, int, uint8_t*, int);
|
||||
|
||||
//------------------------------------------------------------------------------
|
||||
// Init function
|
||||
|
||||
extern void VP8FiltersInitMIPSdspR2(void);
|
||||
extern void WebPInitAlphaProcessingMIPSdspR2(void);
|
||||
|
||||
void WebPInitAlphaProcessing(void) {
|
||||
WebPMultARGBRow = MultARGBRow;
|
||||
WebPMultRow = MultRow;
|
||||
WebPApplyAlphaMultiply = ApplyAlphaMultiply;
|
||||
WebPApplyAlphaMultiply4444 = ApplyAlphaMultiply_16b;
|
||||
WebPDispatchAlpha = DispatchAlpha;
|
||||
|
||||
// If defined, use CPUInfo() to overwrite some pointers with faster versions.
|
||||
if (VP8GetCPUInfo != NULL) {
|
||||
#if defined(WEBP_USE_MIPS_DSP_R2)
|
||||
if (VP8GetCPUInfo(kMIPSdspR2)) {
|
||||
VP8FiltersInitMIPSdspR2();
|
||||
WebPInitAlphaProcessingMIPSdspR2();
|
||||
}
|
||||
#endif
|
||||
}
|
||||
|
93
src/dsp/alpha_processing_mips_dsp_r2.c
Normal file
93
src/dsp/alpha_processing_mips_dsp_r2.c
Normal file
@ -0,0 +1,93 @@
|
||||
// Copyright 2014 Google Inc. All Rights Reserved.
|
||||
//
|
||||
// Use of this source code is governed by a BSD-style license
|
||||
// that can be found in the COPYING 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.
|
||||
// -----------------------------------------------------------------------------
|
||||
//
|
||||
// Utilities for processing transparent channel.
|
||||
//
|
||||
// Author(s): Branimir Vasic (branimir.vasic@imgtec.com)
|
||||
// Djordje Pesut (djordje.pesut@imgtec.com)
|
||||
|
||||
#include "./dsp.h"
|
||||
|
||||
#if defined(WEBP_USE_MIPS_DSP_R2)
|
||||
|
||||
static int DispatchAlpha(const uint8_t* alpha, int alpha_stride,
|
||||
int width, int height,
|
||||
uint8_t* dst, int dst_stride) {
|
||||
uint32_t alpha_mask = 0xffffffff;
|
||||
int i, j, temp0;
|
||||
|
||||
for (j = 0; j < height; ++j) {
|
||||
uint8_t* pdst = dst;
|
||||
const uint8_t* palpha = alpha;
|
||||
for (i = 0; i < (width >> 2); ++i) {
|
||||
int temp1, temp2, temp3;
|
||||
|
||||
__asm__ volatile (
|
||||
"ulw %[temp0], 0(%[palpha]) \n\t"
|
||||
"addiu %[palpha], %[palpha], 4 \n\t"
|
||||
"addiu %[pdst], %[pdst], 16 \n\t"
|
||||
"srl %[temp1], %[temp0], 8 \n\t"
|
||||
"srl %[temp2], %[temp0], 16 \n\t"
|
||||
"srl %[temp3], %[temp0], 24 \n\t"
|
||||
"and %[alpha_mask], %[alpha_mask], %[temp0] \n\t"
|
||||
"sb %[temp0], -16(%[pdst]) \n\t"
|
||||
"sb %[temp1], -12(%[pdst]) \n\t"
|
||||
"sb %[temp2], -8(%[pdst]) \n\t"
|
||||
"sb %[temp3], -4(%[pdst]) \n\t"
|
||||
: [temp0]"=&r"(temp0), [temp1]"=&r"(temp1), [temp2]"=&r"(temp2),
|
||||
[temp3]"=&r"(temp3), [palpha]"+r"(palpha), [pdst]"+r"(pdst),
|
||||
[alpha_mask]"+r"(alpha_mask)
|
||||
:
|
||||
: "memory"
|
||||
);
|
||||
}
|
||||
|
||||
for (i = 0; i < (width & 3); ++i) {
|
||||
__asm__ volatile (
|
||||
"lbu %[temp0], 0(%[palpha]) \n\t"
|
||||
"addiu %[palpha], %[palpha], 1 \n\t"
|
||||
"sb %[temp0], 0(%[pdst]) \n\t"
|
||||
"and %[alpha_mask], %[alpha_mask], %[temp0] \n\t"
|
||||
"addiu %[pdst], %[pdst], 4 \n\t"
|
||||
: [temp0]"=&r"(temp0), [palpha]"+r"(palpha), [pdst]"+r"(pdst),
|
||||
[alpha_mask]"+r"(alpha_mask)
|
||||
:
|
||||
: "memory"
|
||||
);
|
||||
}
|
||||
alpha += alpha_stride;
|
||||
dst += dst_stride;
|
||||
}
|
||||
|
||||
__asm__ volatile (
|
||||
"ext %[temp0], %[alpha_mask], 0, 16 \n\t"
|
||||
"srl %[alpha_mask], %[alpha_mask], 16 \n\t"
|
||||
"and %[alpha_mask], %[alpha_mask], %[temp0] \n\t"
|
||||
"ext %[temp0], %[alpha_mask], 0, 8 \n\t"
|
||||
"srl %[alpha_mask], %[alpha_mask], 8 \n\t"
|
||||
"and %[alpha_mask], %[alpha_mask], %[temp0] \n\t"
|
||||
: [temp0]"=&r"(temp0), [alpha_mask]"+r"(alpha_mask)
|
||||
:
|
||||
);
|
||||
|
||||
return (alpha_mask != 0xff);
|
||||
}
|
||||
|
||||
#endif // WEBP_USE_MIPS_DSP_R2
|
||||
|
||||
//------------------------------------------------------------------------------
|
||||
// Init function
|
||||
|
||||
extern void WebPInitAlphaProcessingMIPSdspR2(void);
|
||||
|
||||
void WebPInitAlphaProcessingMIPSdspR2(void) {
|
||||
#if defined(WEBP_USE_MIPS_DSP_R2)
|
||||
WebPDispatchAlpha = DispatchAlpha;
|
||||
#endif
|
||||
}
|
@ -251,6 +251,13 @@ extern void (*WebPApplyAlphaMultiply)(
|
||||
extern void (*WebPApplyAlphaMultiply4444)(
|
||||
uint8_t* rgba4444, int w, int h, int stride);
|
||||
|
||||
|
||||
// Dispatch the values from alpha[] plane to the ARGB destination 'dst'.
|
||||
// Returns true if alpha[] plane has non-trivial values different from 0xff.
|
||||
extern int (*WebPDispatchAlpha)(const uint8_t* alpha, int alpha_stride,
|
||||
int width, int height,
|
||||
uint8_t* dst, int dst_stride);
|
||||
|
||||
// Pre-Multiply operation transforms x into x * A / 255 (where x=Y,R,G or B).
|
||||
// Un-Multiply operation transforms x into x * 255 / A.
|
||||
|
||||
|
Reference in New Issue
Block a user