view src/video/SDL_blit.h @ 689:5bb080d35049

Date: Tue, 19 Aug 2003 17:57:00 +0200 From: Stephane Marchesin Subject: Re: [SDL] [patch] MMX alpha blit patches with MMX detection I think everything is correct now. I've done as much testing as I could, but some real-world testing wouldn't hurt, I think. The patch is here : http://icps.u-strasbg.fr/~marchesin/sdl_mmxblit.patch If you do byte-by-byte comparison of the output between C and MMX functions, you'll notice that the results for 555 and 565 RGB alpha blits aren't exactly the same. This is because MMX functions for 555 and 565 RGB have an higher accuracy. If you want the exact same behaviour that's possible by masking the three lower alpha bits in the MMX functions. Just ask ! I removed one MMX function because after I fixed it to match its C equivalent, it revealed to be slower than the C version on a PIII (although a bit faster on an Athlon XP). I've also added MMX and PIII replacements for SDL_memcpy. Those provide some speed up in testvidinfo -benchmark (at least for me, under linux & X11).
author Sam Lantinga <slouken@libsdl.org>
date Fri, 22 Aug 2003 05:51:19 +0000
parents 417f8709e648
children b8d311d90021
line wrap: on
line source

/*
    SDL - Simple DirectMedia Layer
    Copyright (C) 1997, 1998, 1999, 2000, 2001, 2002  Sam Lantinga

    This library is free software; you can redistribute it and/or
    modify it under the terms of the GNU Library General Public
    License as published by the Free Software Foundation; either
    version 2 of the License, or (at your option) any later version.

    This library 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
    Library General Public License for more details.

    You should have received a copy of the GNU Library General Public
    License along with this library; if not, write to the Free
    Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA

    Sam Lantinga
    slouken@libsdl.org
*/

#ifdef SAVE_RCSID
static char rcsid =
 "@(#) $Id$";
#endif

#ifndef _SDL_blit_h
#define _SDL_blit_h

#include "SDL_endian.h"

/* The structure passed to the low level blit functions */
typedef struct {
	Uint8 *s_pixels;
	int s_width;
	int s_height;
	int s_skip;
	Uint8 *d_pixels;
	int d_width;
	int d_height;
	int d_skip;
	void *aux_data;
	SDL_PixelFormat *src;
	Uint8 *table;
	SDL_PixelFormat *dst;
} SDL_BlitInfo;

/* The type definition for the low level blit functions */
typedef void (*SDL_loblit)(SDL_BlitInfo *info);

/* This is the private info structure for software accelerated blits */
struct private_swaccel {
	SDL_loblit blit;
	void *aux_data;
};

/* Blit mapping definition */
typedef struct SDL_BlitMap {
	SDL_Surface *dst;
	int identity;
	Uint8 *table;
	SDL_blit hw_blit;
	SDL_blit sw_blit;
	struct private_hwaccel *hw_data;
	struct private_swaccel *sw_data;

	/* the version count matches the destination; mismatch indicates
	   an invalid mapping */
        unsigned int format_version;
} SDL_BlitMap;


/* Functions found in SDL_blit.c */
extern int SDL_CalculateBlit(SDL_Surface *surface);

/* Functions found in SDL_blit_{0,1,N,A}.c */
extern SDL_loblit SDL_CalculateBlit0(SDL_Surface *surface, int complex);
extern SDL_loblit SDL_CalculateBlit1(SDL_Surface *surface, int complex);
extern SDL_loblit SDL_CalculateBlitN(SDL_Surface *surface, int complex);
extern SDL_loblit SDL_CalculateAlphaBlit(SDL_Surface *surface, int complex);

/*
 * Useful macros for blitting routines
 */

#define FORMAT_EQUAL(A, B)						\
    ((A)->BitsPerPixel == (B)->BitsPerPixel				\
     && ((A)->Rmask == (B)->Rmask) && ((A)->Amask == (B)->Amask))

/* Load pixel of the specified format from a buffer and get its R-G-B values */
/* FIXME: rescale values to 0..255 here? */
#define RGB_FROM_PIXEL(pixel, fmt, r, g, b)				\
{									\
	r = (((pixel&fmt->Rmask)>>fmt->Rshift)<<fmt->Rloss); 		\
	g = (((pixel&fmt->Gmask)>>fmt->Gshift)<<fmt->Gloss); 		\
	b = (((pixel&fmt->Bmask)>>fmt->Bshift)<<fmt->Bloss); 		\
}
#define RGB_FROM_RGB565(pixel, r, g, b)					\
{									\
	r = (((pixel&0xF800)>>11)<<3);		 			\
	g = (((pixel&0x07E0)>>5)<<2); 					\
	b = ((pixel&0x001F)<<3); 					\
}
#define RGB_FROM_RGB555(pixel, r, g, b)					\
{									\
	r = (((pixel&0x7C00)>>10)<<3);		 			\
	g = (((pixel&0x03E0)>>5)<<3); 					\
	b = ((pixel&0x001F)<<3); 					\
}
#define RGB_FROM_RGB888(pixel, r, g, b)					\
{									\
	r = ((pixel&0xFF0000)>>16);		 			\
	g = ((pixel&0xFF00)>>8);		 			\
	b = (pixel&0xFF);			 			\
}
#define RETRIEVE_RGB_PIXEL(buf, bpp, pixel)				   \
do {									   \
	switch (bpp) {							   \
		case 2:							   \
			pixel = *((Uint16 *)(buf));			   \
		break;							   \
									   \
		case 3: {						   \
		        Uint8 *B = (Uint8 *)(buf);			   \
			if(SDL_BYTEORDER == SDL_LIL_ENDIAN) {		   \
			        pixel = B[0] + (B[1] << 8) + (B[2] << 16); \
			} else {					   \
			        pixel = (B[0] << 16) + (B[1] << 8) + B[2]; \
			}						   \
		}							   \
		break;							   \
									   \
		case 4:							   \
			pixel = *((Uint32 *)(buf));			   \
		break;							   \
									   \
		default:						   \
			pixel = 0; /* appease gcc */			   \
		break;							   \
	}								   \
} while(0)

#define DISEMBLE_RGB(buf, bpp, fmt, pixel, r, g, b)			   \
do {									   \
	switch (bpp) {							   \
		case 2:							   \
			pixel = *((Uint16 *)(buf));			   \
		break;							   \
									   \
		case 3: {						   \
		        Uint8 *B = (Uint8 *)buf;			   \
			if(SDL_BYTEORDER == SDL_LIL_ENDIAN) {		   \
			        pixel = B[0] + (B[1] << 8) + (B[2] << 16); \
			} else {					   \
			        pixel = (B[0] << 16) + (B[1] << 8) + B[2]; \
			}						   \
		}							   \
		break;							   \
									   \
		case 4:							   \
			pixel = *((Uint32 *)(buf));			   \
		break;							   \
									   \
	        default:						   \
		        pixel = 0;	/* prevent gcc from complaining */ \
		break;							   \
	}								   \
	RGB_FROM_PIXEL(pixel, fmt, r, g, b);				   \
} while(0)

/* Assemble R-G-B values into a specified pixel format and store them */
#define PIXEL_FROM_RGB(pixel, fmt, r, g, b)				\
{									\
	pixel = ((r>>fmt->Rloss)<<fmt->Rshift)|				\
		((g>>fmt->Gloss)<<fmt->Gshift)|				\
		((b>>fmt->Bloss)<<fmt->Bshift);				\
}
#define RGB565_FROM_RGB(pixel, r, g, b)					\
{									\
	pixel = ((r>>3)<<11)|((g>>2)<<5)|(b>>3);			\
}
#define RGB555_FROM_RGB(pixel, r, g, b)					\
{									\
	pixel = ((r>>3)<<10)|((g>>3)<<5)|(b>>3);			\
}
#define RGB888_FROM_RGB(pixel, r, g, b)					\
{									\
	pixel = (r<<16)|(g<<8)|b;					\
}
#define ASSEMBLE_RGB(buf, bpp, fmt, r, g, b) 				\
{									\
	switch (bpp) {							\
		case 2: {						\
			Uint16 pixel;					\
									\
			PIXEL_FROM_RGB(pixel, fmt, r, g, b);		\
			*((Uint16 *)(buf)) = pixel;			\
		}							\
		break;							\
									\
		case 3: {						\
                        if(SDL_BYTEORDER == SDL_LIL_ENDIAN) {		\
			        *((buf)+fmt->Rshift/8) = r;		\
				*((buf)+fmt->Gshift/8) = g;		\
				*((buf)+fmt->Bshift/8) = b;		\
			} else {					\
			        *((buf)+2-fmt->Rshift/8) = r;		\
				*((buf)+2-fmt->Gshift/8) = g;		\
				*((buf)+2-fmt->Bshift/8) = b;		\
			}						\
		}							\
		break;							\
									\
		case 4: {						\
			Uint32 pixel;					\
									\
			PIXEL_FROM_RGB(pixel, fmt, r, g, b);		\
			*((Uint32 *)(buf)) = pixel;			\
		}							\
		break;							\
	}								\
}
#define ASSEMBLE_RGB_AMASK(buf, bpp, fmt, r, g, b, Amask)		\
{									\
	switch (bpp) {							\
		case 2: {						\
			Uint16 *bufp;					\
			Uint16 pixel;					\
									\
			bufp = (Uint16 *)buf;				\
			PIXEL_FROM_RGB(pixel, fmt, r, g, b);		\
			*bufp = pixel | (*bufp & Amask);		\
		}							\
		break;							\
									\
		case 3: {						\
                        if(SDL_BYTEORDER == SDL_LIL_ENDIAN) {		\
			        *((buf)+fmt->Rshift/8) = r;		\
				*((buf)+fmt->Gshift/8) = g;		\
				*((buf)+fmt->Bshift/8) = b;		\
			} else {					\
			        *((buf)+2-fmt->Rshift/8) = r;		\
				*((buf)+2-fmt->Gshift/8) = g;		\
				*((buf)+2-fmt->Bshift/8) = b;		\
			}						\
		}							\
		break;							\
									\
		case 4: {						\
			Uint32 *bufp;					\
			Uint32 pixel;					\
									\
			bufp = (Uint32 *)buf;				\
			PIXEL_FROM_RGB(pixel, fmt, r, g, b);		\
			*bufp = pixel | (*bufp & Amask);		\
		}							\
		break;							\
	}								\
}

/* FIXME: Should we rescale alpha into 0..255 here? */
#define RGBA_FROM_PIXEL(pixel, fmt, r, g, b, a)				\
{									\
	r = ((pixel&fmt->Rmask)>>fmt->Rshift)<<fmt->Rloss; 		\
	g = ((pixel&fmt->Gmask)>>fmt->Gshift)<<fmt->Gloss; 		\
	b = ((pixel&fmt->Bmask)>>fmt->Bshift)<<fmt->Bloss; 		\
	a = ((pixel&fmt->Amask)>>fmt->Ashift)<<fmt->Aloss;	 	\
}
#define RGBA_FROM_8888(pixel, fmt, r, g, b, a)	\
{						\
	r = (pixel&fmt->Rmask)>>fmt->Rshift;	\
	g = (pixel&fmt->Gmask)>>fmt->Gshift;	\
	b = (pixel&fmt->Bmask)>>fmt->Bshift;	\
	a = (pixel&fmt->Amask)>>fmt->Ashift;	\
}
#define RGBA_FROM_RGBA8888(pixel, r, g, b, a)				\
{									\
	r = (pixel>>24);						\
	g = ((pixel>>16)&0xFF);						\
	b = ((pixel>>8)&0xFF);						\
	a = (pixel&0xFF);						\
}
#define RGBA_FROM_ARGB8888(pixel, r, g, b, a)				\
{									\
	r = ((pixel>>16)&0xFF);						\
	g = ((pixel>>8)&0xFF);						\
	b = (pixel&0xFF);						\
	a = (pixel>>24);						\
}
#define RGBA_FROM_ABGR8888(pixel, r, g, b, a)				\
{									\
	r = (pixel&0xFF);						\
	g = ((pixel>>8)&0xFF);						\
	b = ((pixel>>16)&0xFF);						\
	a = (pixel>>24);						\
}
#define DISEMBLE_RGBA(buf, bpp, fmt, pixel, r, g, b, a)			   \
do {									   \
	switch (bpp) {							   \
		case 2:							   \
			pixel = *((Uint16 *)(buf));			   \
		break;							   \
									   \
		case 3:	{/* FIXME: broken code (no alpha) */		   \
		        Uint8 *b = (Uint8 *)buf;			   \
			if(SDL_BYTEORDER == SDL_LIL_ENDIAN) {		   \
			        pixel = b[0] + (b[1] << 8) + (b[2] << 16); \
			} else {					   \
			        pixel = (b[0] << 16) + (b[1] << 8) + b[2]; \
			}						   \
		}							   \
		break;							   \
									   \
		case 4:							   \
			pixel = *((Uint32 *)(buf));			   \
		break;							   \
									   \
		default:						   \
		        pixel = 0; /* stop gcc complaints */		   \
		break;							   \
	}								   \
	RGBA_FROM_PIXEL(pixel, fmt, r, g, b, a);			   \
	pixel &= ~fmt->Amask;						   \
} while(0)

/* FIXME: this isn't correct, especially for Alpha (maximum != 255) */
#define PIXEL_FROM_RGBA(pixel, fmt, r, g, b, a)				\
{									\
	pixel = ((r>>fmt->Rloss)<<fmt->Rshift)|				\
		((g>>fmt->Gloss)<<fmt->Gshift)|				\
		((b>>fmt->Bloss)<<fmt->Bshift)|				\
		((a>>fmt->Aloss)<<fmt->Ashift);				\
}
#define ASSEMBLE_RGBA(buf, bpp, fmt, r, g, b, a)			\
{									\
	switch (bpp) {							\
		case 2: {						\
			Uint16 pixel;					\
									\
			PIXEL_FROM_RGBA(pixel, fmt, r, g, b, a);	\
			*((Uint16 *)(buf)) = pixel;			\
		}							\
		break;							\
									\
		case 3: { /* FIXME: broken code (no alpha) */		\
                        if(SDL_BYTEORDER == SDL_LIL_ENDIAN) {		\
			        *((buf)+fmt->Rshift/8) = r;		\
				*((buf)+fmt->Gshift/8) = g;		\
				*((buf)+fmt->Bshift/8) = b;		\
			} else {					\
			        *((buf)+2-fmt->Rshift/8) = r;		\
				*((buf)+2-fmt->Gshift/8) = g;		\
				*((buf)+2-fmt->Bshift/8) = b;		\
			}						\
		}							\
		break;							\
									\
		case 4: {						\
			Uint32 pixel;					\
									\
			PIXEL_FROM_RGBA(pixel, fmt, r, g, b, a);	\
			*((Uint32 *)(buf)) = pixel;			\
		}							\
		break;							\
	}								\
}

/* Blend the RGB values of two pixels based on a source alpha value */
#define ALPHA_BLEND(sR, sG, sB, A, dR, dG, dB)	\
do {						\
	dR = (((sR-dR)*(A))>>8)+dR;		\
	dG = (((sG-dG)*(A))>>8)+dG;		\
	dB = (((sB-dB)*(A))>>8)+dB;		\
} while(0)

/* This is a very useful loop for optimizing blitters */
#if defined(_MSC_VER) && (_MSC_VER == 1300)
/* There's a bug in the Visual C++ 7 optimizer when compiling this code */
#else
#define USE_DUFFS_LOOP
#endif
#ifdef USE_DUFFS_LOOP

/* 8-times unrolled loop */
#define DUFFS_LOOP8(pixel_copy_increment, width)			\
{ int n = (width+7)/8;							\
	switch (width & 7) {						\
	case 0: do {	pixel_copy_increment;				\
	case 7:		pixel_copy_increment;				\
	case 6:		pixel_copy_increment;				\
	case 5:		pixel_copy_increment;				\
	case 4:		pixel_copy_increment;				\
	case 3:		pixel_copy_increment;				\
	case 2:		pixel_copy_increment;				\
	case 1:		pixel_copy_increment;				\
		} while ( --n > 0 );					\
	}								\
}

/* 4-times unrolled loop */
#define DUFFS_LOOP4(pixel_copy_increment, width)			\
{ int n = (width+3)/4;							\
	switch (width & 3) {						\
	case 0: do {	pixel_copy_increment;				\
	case 3:		pixel_copy_increment;				\
	case 2:		pixel_copy_increment;				\
	case 1:		pixel_copy_increment;				\
		} while ( --n > 0 );					\
	}								\
}

/* 2 - times unrolled loop */
#define DUFFS_LOOP_DOUBLE2(pixel_copy_increment,			\
				double_pixel_copy_increment, width)	\
{ int n, w = width;							\
	if( w & 1 ) {							\
	    pixel_copy_increment;					\
	    w--;							\
	}								\
	if ( w > 0 )	{						\
	    n = ( w + 2) / 4;						\
	    switch( w & 2 ) {						\
	    case 0: do {	double_pixel_copy_increment;		\
	    case 2:		double_pixel_copy_increment;		\
		    } while ( --n > 0 );					\
	    }								\
	}								\
}

/* 2 - times unrolled loop 4 pixels */
#define DUFFS_LOOP_QUATRO2(pixel_copy_increment,			\
				double_pixel_copy_increment,		\
				quatro_pixel_copy_increment, width)	\
{ int n, w = width;								\
        if(w & 1) {							\
	  pixel_copy_increment;						\
	  w--;								\
	}								\
	if(w & 2) {							\
	  double_pixel_copy_increment;					\
	  w -= 2;							\
	}								\
	if ( w > 0 ) {							\
	    n = ( w + 7 ) / 8;						\
	    switch( w & 4 ) {						\
	    case 0: do {	quatro_pixel_copy_increment;		\
	    case 4:		quatro_pixel_copy_increment;		\
		    } while ( --n > 0 );					\
	    }								\
	}								\
}

/* Use the 8-times version of the loop by default */
#define DUFFS_LOOP(pixel_copy_increment, width)				\
	DUFFS_LOOP8(pixel_copy_increment, width)

#else

/* Don't use Duff's device to unroll loops */
#define DUFFS_LOOP_DOUBLE2(pixel_copy_increment,			\
			 double_pixel_copy_increment, width)		\
{ int n = width;								\
    if( n & 1 ) {							\
	pixel_copy_increment;						\
	n--;								\
    }									\
    n=n>>1;								\
    for(; n > 0; --n) {   						\
	double_pixel_copy_increment;					\
    }									\
}

/* Don't use Duff's device to unroll loops */
#define DUFFS_LOOP_QUATRO2(pixel_copy_increment,			\
				double_pixel_copy_increment,		\
				quatro_pixel_copy_increment, width)	\
{ int n = width;								\
        if(n & 1) {							\
	  pixel_copy_increment;						\
	  n--;								\
	}								\
	if(n & 2) {							\
	  double_pixel_copy_increment;					\
	  n -= 2;							\
	}								\
	n=n>>2;								\
	for(; n > 0; --n) {   						\
	  quatro_pixel_copy_increment;					\
        }								\
}

/* Don't use Duff's device to unroll loops */
#define DUFFS_LOOP(pixel_copy_increment, width)				\
{ int n;								\
	for ( n=width; n > 0; --n ) {					\
		pixel_copy_increment;					\
	}								\
}
#define DUFFS_LOOP8(pixel_copy_increment, width)			\
	DUFFS_LOOP(pixel_copy_increment, width)
#define DUFFS_LOOP4(pixel_copy_increment, width)			\
	DUFFS_LOOP(pixel_copy_increment, width)

#endif /* USE_DUFFS_LOOP */

/* Prevent Visual C++ 6.0 from printing out stupid warnings */
#if defined(_MSC_VER) && (_MSC_VER >= 600)
#pragma warning(disable: 4550)
#endif

#endif /* _SDL_blit_h */