src/audio/SDL_mixer_MMX_VC.c
author Ryan C. Gordon <icculus@icculus.org>
Thu, 31 Aug 2006 21:00:10 +0000
changeset 1994 6abc7e6f9817
parent 1895 c121d94672cb
child 2280 1863fa2050e8
permissions -rw-r--r--
Added int32 adn float32 support to SDL_LoadWAV_RW().
     1 /*
     2     SDL - Simple DirectMedia Layer
     3     Copyright (C) 1997-2006 Sam Lantinga
     4 
     5     This library is free software; you can redistribute it and/or
     6     modify it under the terms of the GNU Lesser General Public
     7     License as published by the Free Software Foundation; either
     8     version 2.1 of the License, or (at your option) any later version.
     9 
    10     This library is distributed in the hope that it will be useful,
    11     but WITHOUT ANY WARRANTY; without even the implied warranty of
    12     MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
    13     Lesser General Public License for more details.
    14 
    15     You should have received a copy of the GNU Lesser General Public
    16     License along with this library; if not, write to the Free Software
    17     Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
    18 
    19     Sam Lantinga
    20     slouken@libsdl.org
    21 */
    22 #include "SDL_config.h"
    23 
    24 #include "SDL_mixer_MMX_VC.h"
    25 
    26 #if ((defined(_MSC_VER) && defined(_M_IX86)) || defined(__WATCOMC__)) && defined(SDL_ASSEMBLY_ROUTINES)
    27 // MMX assembler version of SDL_MixAudio for signed little endian 16 bit samples and signed 8 bit samples
    28 // Copyright 2002 Stephane Marchesin (stephane.marchesin@wanadoo.fr)
    29 // Converted to Intel ASM notation by Cth
    30 // This code is licensed under the LGPL (see COPYING for details)
    31 // 
    32 // Assumes buffer size in bytes is a multiple of 16
    33 // Assumes SDL_MIX_MAXVOLUME = 128
    34 
    35 
    36 ////////////////////////////////////////////////
    37 // Mixing for 16 bit signed buffers
    38 ////////////////////////////////////////////////
    39 
    40 void
    41 SDL_MixAudio_MMX_S16_VC(char *dst, char *src, unsigned int nSize, int volume)
    42 {
    43     /* *INDENT-OFF* */
    44 	__asm
    45 	{
    46 
    47 		push	edi
    48 		push	esi
    49 		push	ebx
    50 		
    51 		mov		edi, dst		// edi = dst
    52 		mov		esi, src		// esi = src
    53 		mov		eax, volume		// eax = volume
    54 		mov		ebx, nSize		// ebx = size
    55 		shr		ebx, 4			// process 16 bytes per iteration = 8 samples
    56 		jz		endS16
    57 		
    58 		pxor	mm0, mm0
    59 		movd	mm0, eax		//%%eax,%%mm0
    60 		movq	mm1, mm0		//%%mm0,%%mm1
    61 		psllq	mm0, 16			//$16,%%mm0
    62 		por		mm0, mm1		//%%mm1,%%mm0
    63 		psllq	mm0, 16			//$16,%%mm0
    64 		por		mm0, mm1		//%%mm1,%%mm0
    65 		psllq	mm0, 16			//$16,%%mm0
    66 		por		mm0, mm1		//%%mm1,%%mm0			// mm0 = vol|vol|vol|vol
    67 
    68 		#ifndef __WATCOMC__
    69 		align	16
    70 		#endif
    71 mixloopS16:
    72 		movq	mm1, [esi]		//(%%esi),%%mm1\n" // mm1 = a|b|c|d
    73 		movq	mm2, mm1		//%%mm1,%%mm2\n" // mm2 = a|b|c|d
    74 		movq	mm4, [esi + 8]	//8(%%esi),%%mm4\n" // mm4 = e|f|g|h
    75 		// pre charger le buffer dst dans mm7
    76 		movq	mm7, [edi]		//(%%edi),%%mm7\n" // mm7 = dst[0]"
    77 		// multiplier par le volume
    78 		pmullw	mm1, mm0		//%%mm0,%%mm1\n" // mm1 = l(a*v)|l(b*v)|l(c*v)|l(d*v)
    79 		pmulhw	mm2, mm0		//%%mm0,%%mm2\n" // mm2 = h(a*v)|h(b*v)|h(c*v)|h(d*v)
    80 		movq	mm5, mm4		//%%mm4,%%mm5\n" // mm5 = e|f|g|h
    81 		pmullw	mm4, mm0		//%%mm0,%%mm4\n" // mm4 = l(e*v)|l(f*v)|l(g*v)|l(h*v)
    82 		pmulhw	mm5, mm0		//%%mm0,%%mm5\n" // mm5 = h(e*v)|h(f*v)|h(g*v)|h(h*v)
    83 		movq	mm3, mm1		//%%mm1,%%mm3\n" // mm3 = l(a*v)|l(b*v)|l(c*v)|l(d*v)
    84 		punpckhwd	mm1, mm2	//%%mm2,%%mm1\n" // mm1 = a*v|b*v
    85 		movq		mm6, mm4	//%%mm4,%%mm6\n" // mm6 = l(e*v)|l(f*v)|l(g*v)|l(h*v)
    86 		punpcklwd	mm3, mm2	//%%mm2,%%mm3\n" // mm3 = c*v|d*v
    87 		punpckhwd	mm4, mm5	//%%mm5,%%mm4\n" // mm4 = e*f|f*v
    88 		punpcklwd	mm6, mm5	//%%mm5,%%mm6\n" // mm6 = g*v|h*v
    89 		// pre charger le buffer dst dans mm5
    90 		movq	mm5, [edi + 8]	//8(%%edi),%%mm5\n" // mm5 = dst[1]
    91 		// diviser par 128
    92 		psrad	mm1, 7			//$7,%%mm1\n" // mm1 = a*v/128|b*v/128 , 128 = SDL_MIX_MAXVOLUME
    93 		add		esi, 16			//$16,%%esi\n"
    94 		psrad	mm3, 7			//$7,%%mm3\n" // mm3 = c*v/128|d*v/128
    95 		psrad	mm4, 7			//$7,%%mm4\n" // mm4 = e*v/128|f*v/128
    96 		// mm1 = le sample avec le volume modifie
    97 		packssdw	mm3, mm1	//%%mm1,%%mm3\n" // mm3 = s(a*v|b*v|c*v|d*v)
    98 		psrad	mm6, 7			//$7,%%mm6\n" // mm6= g*v/128|h*v/128
    99 		paddsw	mm3, mm7		//%%mm7,%%mm3\n" // mm3 = adjust_volume(src)+dst
   100 		// mm4 = le sample avec le volume modifie
   101 		packssdw	mm6, mm4	//%%mm4,%%mm6\n" // mm6 = s(e*v|f*v|g*v|h*v)
   102 		movq	[edi], mm3		//%%mm3,(%%edi)\n"
   103 		paddsw	mm6, mm5		//%%mm5,%%mm6\n" // mm6 = adjust_volume(src)+dst
   104 		movq	[edi + 8], mm6	//%%mm6,8(%%edi)\n"
   105 		add		edi, 16			//$16,%%edi\n"
   106 		dec		ebx				//%%ebx\n"
   107 		jnz mixloopS16
   108 
   109 ends16:
   110 		emms
   111 		
   112 		pop		ebx
   113 		pop		esi
   114 		pop		edi
   115 	}
   116     /* *INDENT-ON* */
   117 }
   118 
   119 ////////////////////////////////////////////////
   120 // Mixing for 8 bit signed buffers
   121 ////////////////////////////////////////////////
   122 
   123 void
   124 SDL_MixAudio_MMX_S8_VC(char *dst, char *src, unsigned int nSize, int volume)
   125 {
   126     /* *INDENT-OFF* */
   127 	_asm
   128 	{
   129 
   130 		push	edi
   131 		push	esi
   132 		push	ebx
   133 		
   134 		mov		edi, dst	//movl	%0,%%edi	// edi = dst
   135 		mov		esi, src	//%1,%%esi	// esi = src
   136 		mov		eax, volume	//%3,%%eax	// eax = volume
   137 
   138 		movd	mm0, eax	//%%eax,%%mm0
   139 		movq	mm1, mm0	//%%mm0,%%mm1
   140 		psllq	mm0, 16		//$16,%%mm0
   141 		por		mm0, mm1	//%%mm1,%%mm0
   142 		psllq	mm0, 16		//$16,%%mm0
   143 		por		mm0, mm1	//%%mm1,%%mm0
   144 		psllq	mm0, 16		//$16,%%mm0
   145 		por		mm0, mm1	//%%mm1,%%mm0
   146 		
   147 		mov		ebx, nSize	//%2,%%ebx	// ebx = size
   148 		shr		ebx, 3		//$3,%%ebx	// process 8 bytes per iteration = 8 samples
   149 		cmp		ebx, 0		//$0,%%ebx
   150 		je		endS8
   151 
   152 		#ifndef __WATCOMC__
   153 		align 16
   154 		#endif
   155 mixloopS8:
   156 		pxor	mm2, mm2	//%%mm2,%%mm2		// mm2 = 0
   157 		movq	mm1, [esi]	//(%%esi),%%mm1	// mm1 = a|b|c|d|e|f|g|h
   158 		movq	mm3, mm1	//%%mm1,%%mm3 	// mm3 = a|b|c|d|e|f|g|h
   159 		// on va faire le "sign extension" en faisant un cmp avec 0 qui retourne 1 si <0, 0 si >0
   160 		pcmpgtb		mm2, mm1	//%%mm1,%%mm2	// mm2 = 11111111|00000000|00000000....
   161 		punpckhbw	mm1, mm2	//%%mm2,%%mm1	// mm1 = 0|a|0|b|0|c|0|d
   162 		punpcklbw	mm3, mm2	//%%mm2,%%mm3	// mm3 = 0|e|0|f|0|g|0|h
   163 		movq	mm2, [edi]	//(%%edi),%%mm2	// mm2 = destination
   164 		pmullw	mm1, mm0	//%%mm0,%%mm1	// mm1 = v*a|v*b|v*c|v*d
   165 		add		esi, 8		//$8,%%esi
   166 		pmullw	mm3, mm0	//%%mm0,%%mm3	// mm3 = v*e|v*f|v*g|v*h
   167 		psraw	mm1, 7		//$7,%%mm1		// mm1 = v*a/128|v*b/128|v*c/128|v*d/128 
   168 		psraw	mm3, 7		//$7,%%mm3		// mm3 = v*e/128|v*f/128|v*g/128|v*h/128
   169 		packsswb mm3, mm1	//%%mm1,%%mm3	// mm1 = v*a/128|v*b/128|v*c/128|v*d/128|v*e/128|v*f/128|v*g/128|v*h/128
   170 		paddsb	mm3, mm2	//%%mm2,%%mm3	// add to destination buffer
   171 		movq	[edi], mm3	//%%mm3,(%%edi)	// store back to ram
   172 		add		edi, 8		//$8,%%edi
   173 		dec		ebx			//%%ebx
   174 		jnz		mixloopS8
   175 		
   176 endS8:
   177 		emms
   178 		
   179 		pop		ebx
   180 		pop		esi
   181 		pop		edi
   182 	}
   183     /* *INDENT-ON* */
   184 }
   185 
   186 #endif /* SDL_ASSEMBLY_ROUTINES */
   187 
   188 /* vi: set ts=4 sw=4 expandtab: */