2 SDL - Simple DirectMedia Layer
3 Copyright (C) 1997-2009 Sam Lantinga
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.
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.
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
22 #include "SDL_config.h"
24 #include "SDL_mixer_MMX_VC.h"
26 #if defined(SDL_BUGGY_MMX_MIXERS) /* buggy, so we're disabling them. --ryan. */
27 #if ((defined(_MSC_VER) && defined(_M_IX86)) || defined(__WATCOMC__)) && defined(SDL_ASSEMBLY_ROUTINES)
28 // MMX assembler version of SDL_MixAudio for signed little endian 16 bit samples and signed 8 bit samples
29 // Copyright 2002 Stephane Marchesin (stephane.marchesin@wanadoo.fr)
30 // Converted to Intel ASM notation by Cth
31 // This code is licensed under the LGPL (see COPYING for details)
33 // Assumes buffer size in bytes is a multiple of 16
34 // Assumes SDL_MIX_MAXVOLUME = 128
37 ////////////////////////////////////////////////
38 // Mixing for 16 bit signed buffers
39 ////////////////////////////////////////////////
42 SDL_MixAudio_MMX_S16_VC(char *dst, char *src, unsigned int nSize, int volume)
52 mov edi, dst // edi = dst
53 mov esi, src // esi = src
54 mov eax, volume // eax = volume
55 mov ebx, nSize // ebx = size
56 shr ebx, 4 // process 16 bytes per iteration = 8 samples
60 movd mm0, eax //%%eax,%%mm0
61 movq mm1, mm0 //%%mm0,%%mm1
62 psllq mm0, 16 //$16,%%mm0
63 por mm0, mm1 //%%mm1,%%mm0
64 psllq mm0, 16 //$16,%%mm0
65 por mm0, mm1 //%%mm1,%%mm0
66 psllq mm0, 16 //$16,%%mm0
67 por mm0, mm1 //%%mm1,%%mm0 // mm0 = vol|vol|vol|vol
73 movq mm1, [esi] //(%%esi),%%mm1\n" // mm1 = a|b|c|d
74 movq mm2, mm1 //%%mm1,%%mm2\n" // mm2 = a|b|c|d
75 movq mm4, [esi + 8] //8(%%esi),%%mm4\n" // mm4 = e|f|g|h
76 // pre charger le buffer dst dans mm7
77 movq mm7, [edi] //(%%edi),%%mm7\n" // mm7 = dst[0]"
78 // multiplier par le volume
79 pmullw mm1, mm0 //%%mm0,%%mm1\n" // mm1 = l(a*v)|l(b*v)|l(c*v)|l(d*v)
80 pmulhw mm2, mm0 //%%mm0,%%mm2\n" // mm2 = h(a*v)|h(b*v)|h(c*v)|h(d*v)
81 movq mm5, mm4 //%%mm4,%%mm5\n" // mm5 = e|f|g|h
82 pmullw mm4, mm0 //%%mm0,%%mm4\n" // mm4 = l(e*v)|l(f*v)|l(g*v)|l(h*v)
83 pmulhw mm5, mm0 //%%mm0,%%mm5\n" // mm5 = h(e*v)|h(f*v)|h(g*v)|h(h*v)
84 movq mm3, mm1 //%%mm1,%%mm3\n" // mm3 = l(a*v)|l(b*v)|l(c*v)|l(d*v)
85 punpckhwd mm1, mm2 //%%mm2,%%mm1\n" // mm1 = a*v|b*v
86 movq mm6, mm4 //%%mm4,%%mm6\n" // mm6 = l(e*v)|l(f*v)|l(g*v)|l(h*v)
87 punpcklwd mm3, mm2 //%%mm2,%%mm3\n" // mm3 = c*v|d*v
88 punpckhwd mm4, mm5 //%%mm5,%%mm4\n" // mm4 = e*f|f*v
89 punpcklwd mm6, mm5 //%%mm5,%%mm6\n" // mm6 = g*v|h*v
90 // pre charger le buffer dst dans mm5
91 movq mm5, [edi + 8] //8(%%edi),%%mm5\n" // mm5 = dst[1]
93 psrad mm1, 7 //$7,%%mm1\n" // mm1 = a*v/128|b*v/128 , 128 = SDL_MIX_MAXVOLUME
94 add esi, 16 //$16,%%esi\n"
95 psrad mm3, 7 //$7,%%mm3\n" // mm3 = c*v/128|d*v/128
96 psrad mm4, 7 //$7,%%mm4\n" // mm4 = e*v/128|f*v/128
97 // mm1 = le sample avec le volume modifie
98 packssdw mm3, mm1 //%%mm1,%%mm3\n" // mm3 = s(a*v|b*v|c*v|d*v)
99 psrad mm6, 7 //$7,%%mm6\n" // mm6= g*v/128|h*v/128
100 paddsw mm3, mm7 //%%mm7,%%mm3\n" // mm3 = adjust_volume(src)+dst
101 // mm4 = le sample avec le volume modifie
102 packssdw mm6, mm4 //%%mm4,%%mm6\n" // mm6 = s(e*v|f*v|g*v|h*v)
103 movq [edi], mm3 //%%mm3,(%%edi)\n"
104 paddsw mm6, mm5 //%%mm5,%%mm6\n" // mm6 = adjust_volume(src)+dst
105 movq [edi + 8], mm6 //%%mm6,8(%%edi)\n"
106 add edi, 16 //$16,%%edi\n"
120 ////////////////////////////////////////////////
121 // Mixing for 8 bit signed buffers
122 ////////////////////////////////////////////////
125 SDL_MixAudio_MMX_S8_VC(char *dst, char *src, unsigned int nSize, int volume)
135 mov edi, dst //movl %0,%%edi // edi = dst
136 mov esi, src //%1,%%esi // esi = src
137 mov eax, volume //%3,%%eax // eax = volume
139 movd mm0, eax //%%eax,%%mm0
140 movq mm1, mm0 //%%mm0,%%mm1
141 psllq mm0, 16 //$16,%%mm0
142 por mm0, mm1 //%%mm1,%%mm0
143 psllq mm0, 16 //$16,%%mm0
144 por mm0, mm1 //%%mm1,%%mm0
145 psllq mm0, 16 //$16,%%mm0
146 por mm0, mm1 //%%mm1,%%mm0
148 mov ebx, nSize //%2,%%ebx // ebx = size
149 shr ebx, 3 //$3,%%ebx // process 8 bytes per iteration = 8 samples
150 cmp ebx, 0 //$0,%%ebx
157 pxor mm2, mm2 //%%mm2,%%mm2 // mm2 = 0
158 movq mm1, [esi] //(%%esi),%%mm1 // mm1 = a|b|c|d|e|f|g|h
159 movq mm3, mm1 //%%mm1,%%mm3 // mm3 = a|b|c|d|e|f|g|h
160 // on va faire le "sign extension" en faisant un cmp avec 0 qui retourne 1 si <0, 0 si >0
161 pcmpgtb mm2, mm1 //%%mm1,%%mm2 // mm2 = 11111111|00000000|00000000....
162 punpckhbw mm1, mm2 //%%mm2,%%mm1 // mm1 = 0|a|0|b|0|c|0|d
163 punpcklbw mm3, mm2 //%%mm2,%%mm3 // mm3 = 0|e|0|f|0|g|0|h
164 movq mm2, [edi] //(%%edi),%%mm2 // mm2 = destination
165 pmullw mm1, mm0 //%%mm0,%%mm1 // mm1 = v*a|v*b|v*c|v*d
166 add esi, 8 //$8,%%esi
167 pmullw mm3, mm0 //%%mm0,%%mm3 // mm3 = v*e|v*f|v*g|v*h
168 psraw mm1, 7 //$7,%%mm1 // mm1 = v*a/128|v*b/128|v*c/128|v*d/128
169 psraw mm3, 7 //$7,%%mm3 // mm3 = v*e/128|v*f/128|v*g/128|v*h/128
170 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
171 paddsb mm3, mm2 //%%mm2,%%mm3 // add to destination buffer
172 movq [edi], mm3 //%%mm3,(%%edi) // store back to ram
173 add edi, 8 //$8,%%edi
187 #endif /* SDL_ASSEMBLY_ROUTINES */
188 #endif /* SDL_BUGGY_MMX_MIXERS */
190 /* vi: set ts=4 sw=4 expandtab: */