Mercurial > sdl-ios-xcode

comparison src/audio/SDL_mixer_MMX_VC.c @ 1720:a1ebb17f9c52 SDL-1.3

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Cleaned up a bunch of warnings, started adding Win32 event support
author Sam Lantinga <slouken@libsdl.org>
date Fri, 30 Jun 2006 05:42:49 +0000
parents 4da1ee79c9af
children 1cc762cafff8
comparison
equal deleted inserted replaced
1719:5b9f50c957ed 1720:a1ebb17f9c52
38 //////////////////////////////////////////////// 38 ////////////////////////////////////////////////
39 39
40 void 40 void
41 SDL_MixAudio_MMX_S16_VC(char *dst, char *src, unsigned int nSize, int volume) 41 SDL_MixAudio_MMX_S16_VC(char *dst, char *src, unsigned int nSize, int volume)
42 { 42 {
43 /* *INDENT-OFF* */
43 __asm { 44 __asm {
44 45 push edi
45 push edi push esi push ebx mov edi, dst // edi = dst 46 push esi
47 push ebx
48 mov edi, dst // edi = dst
46 mov esi, src // esi = src 49 mov esi, src // esi = src
47 mov eax, volume // eax = volume 50 mov eax, volume // eax = volume
48 mov ebx, nSize // ebx = size 51 mov ebx, nSize // ebx = size
49 shr ebx, 4 // process 16 bytes per iteration = 8 samples 52 shr ebx, 4 // process 16 bytes per iteration = 8 samples
50 jz endS16 pxor mm0, mm0 movd mm0, eax //%%eax,%%mm0 53 jz endS16
54 pxor mm0, mm0
55 movd mm0, eax //%%eax,%%mm0
51 movq mm1, mm0 //%%mm0,%%mm1 56 movq mm1, mm0 //%%mm0,%%mm1
52 psllq mm0, 16 //$16,%%mm0 57 psllq mm0, 16 //$16,%%mm0
53 por mm0, mm1 //%%mm1,%%mm0 58 por mm0, mm1 //%%mm1,%%mm0
54 psllq mm0, 16 //$16,%%mm0 59 psllq mm0, 16 //$16,%%mm0
55 por mm0, mm1 //%%mm1,%%mm0 60 por mm0, mm1 //%%mm1,%%mm0
56 psllq mm0, 16 //$16,%%mm0 61 psllq mm0, 16 //$16,%%mm0
57 por mm0, mm1 //%%mm1,%%mm0 // mm0 = vol|vol|vol|vol 62 por mm0, mm1 //%%mm1,%%mm0 // mm0 = vol|vol|vol|vol
58 #ifndef __WATCOMC__ 63 #ifndef __WATCOMC__
59 align 16 64 align 16
60 #endif 65 #endif
61 mixloopS16:movq mm1,[esi] //(%%esi),%%mm1\n" // mm1 = a|b|c|d 66 mixloopS16:
62 movq mm2, mm1 //%%mm1,%%mm2\n" // mm2 = a|b|c|d 67 movq mm1,[esi] //(%%esi),%%mm1\n" // mm1 = a|b|c|d
68 movq mm2, mm1 //%%mm1,%%mm2\n" // mm2 = a|b|c|d
63 movq mm4,[esi + 8] //8(%%esi),%%mm4\n" // mm4 = e|f|g|h 69 movq mm4,[esi + 8] //8(%%esi),%%mm4\n" // mm4 = e|f|g|h
64 // pre charger le buffer dst dans mm7 70 // pre charger le buffer dst dans mm7
65 movq mm7,[edi] //(%%edi),%%mm7\n" // mm7 = dst[0]" 71 movq mm7,[edi] //(%%edi),%%mm7\n" // mm7 = dst[0]"
66 // multiplier par le volume 72 // multiplier par le volume
67 pmullw mm1, mm0 //%%mm0,%%mm1\n" // mm1 = l(a*v)|l(b*v)|l(c*v)|l(d*v) 73 pmullw mm1, mm0 //%%mm0,%%mm1\n" // mm1 = l(a*v)|l(b*v)|l(c*v)|l(d*v)
68 pmulhw mm2, mm0 //%%mm0,%%mm2\n" // mm2 = h(a*v)|h(b*v)|h(c*v)|h(d*v) 74 pmulhw mm2, mm0 //%%mm0,%%mm2\n" // mm2 = h(a*v)|h(b*v)|h(c*v)|h(d*v)
69 movq mm5, mm4 //%%mm4,%%mm5\n" // mm5 = e|f|g|h 75 movq mm5, mm4 //%%mm4,%%mm5\n" // mm5 = e|f|g|h
70 pmullw mm4, mm0 //%%mm0,%%mm4\n" // mm4 = l(e*v)|l(f*v)|l(g*v)|l(h*v) 76 pmullw mm4, mm0 //%%mm0,%%mm4\n" // mm4 = l(e*v)|l(f*v)|l(g*v)|l(h*v)
71 pmulhw mm5, mm0 //%%mm0,%%mm5\n" // mm5 = h(e*v)|h(f*v)|h(g*v)|h(h*v) 77 pmulhw mm5, mm0 //%%mm0,%%mm5\n" // mm5 = h(e*v)|h(f*v)|h(g*v)|h(h*v)
72 movq mm3, mm1 //%%mm1,%%mm3\n" // mm3 = l(a*v)|l(b*v)|l(c*v)|l(d*v) 78 movq mm3, mm1 //%%mm1,%%mm3\n" // mm3 = l(a*v)|l(b*v)|l(c*v)|l(d*v)
76 punpckhwd mm4, mm5 //%%mm5,%%mm4\n" // mm4 = e*f|f*v 82 punpckhwd mm4, mm5 //%%mm5,%%mm4\n" // mm4 = e*f|f*v
77 punpcklwd mm6, mm5 //%%mm5,%%mm6\n" // mm6 = g*v|h*v 83 punpcklwd mm6, mm5 //%%mm5,%%mm6\n" // mm6 = g*v|h*v
78 // pre charger le buffer dst dans mm5 84 // pre charger le buffer dst dans mm5
79 movq mm5,[edi + 8] //8(%%edi),%%mm5\n" // mm5 = dst[1] 85 movq mm5,[edi + 8] //8(%%edi),%%mm5\n" // mm5 = dst[1]
80 // diviser par 128 86 // diviser par 128
81 psrad mm1, 7 //$7,%%mm1\n" // mm1 = a*v/128|b*v/128 , 128 = SDL_MIX_MAXVOLUME 87 psrad mm1, 7 //$7,%%mm1\n" // mm1 = a*v/128|b*v/128 , 128 = SDL_MIX_MAXVOLUME
82 add esi, 16 //$16,%%esi\n" 88 add esi, 16 //$16,%%esi\n"
83 psrad mm3, 7 //$7,%%mm3\n" // mm3 = c*v/128|d*v/128 89 psrad mm3, 7 //$7,%%mm3\n" // mm3 = c*v/128|d*v/128
84 psrad mm4, 7 //$7,%%mm4\n" // mm4 = e*v/128|f*v/128 90 psrad mm4, 7 //$7,%%mm4\n" // mm4 = e*v/128|f*v/128
85 // mm1 = le sample avec le volume modifie 91 // mm1 = le sample avec le volume modifie
86 packssdw mm3, mm1 //%%mm1,%%mm3\n" // mm3 = s(a*v|b*v|c*v|d*v) 92 packssdw mm3, mm1 //%%mm1,%%mm3\n" // mm3 = s(a*v|b*v|c*v|d*v)
91 movq[edi], mm3 //%%mm3,(%%edi)\n" 97 movq[edi], mm3 //%%mm3,(%%edi)\n"
92 paddsw mm6, mm5 //%%mm5,%%mm6\n" // mm6 = adjust_volume(src)+dst 98 paddsw mm6, mm5 //%%mm5,%%mm6\n" // mm6 = adjust_volume(src)+dst
93 movq[edi + 8], mm6 //%%mm6,8(%%edi)\n" 99 movq[edi + 8], mm6 //%%mm6,8(%%edi)\n"
94 add edi, 16 //$16,%%edi\n" 100 add edi, 16 //$16,%%edi\n"
95 dec ebx //%%ebx\n" 101 dec ebx //%%ebx\n"
96 jnz mixloopS16 ends16:emms pop ebx pop esi pop edi} 102 jnz mixloopS16
97 103 ends16:
104 emms
105 pop ebx
106 pop esi
107 pop edi
108 }
109 /* *INDENT-ON* */
98 } 110 }
99 111
100 //////////////////////////////////////////////// 112 ////////////////////////////////////////////////
101 // Mixing for 8 bit signed buffers 113 // Mixing for 8 bit signed buffers
102 //////////////////////////////////////////////// 114 ////////////////////////////////////////////////
103 115
104 void 116 void
105 SDL_MixAudio_MMX_S8_VC(char *dst, char *src, unsigned int nSize, int volume) 117 SDL_MixAudio_MMX_S8_VC(char *dst, char *src, unsigned int nSize, int volume)
106 { 118 {
119 /* *INDENT-OFF* */
107 _asm { 120 _asm {
108 121
109 push edi push esi push ebx mov edi, dst //movl %0,%%edi // edi = dst 122 push edi
123 push esi
124 push ebx
125 mov edi, dst //%0,%%edi // edi = dst
110 mov esi, src //%1,%%esi // esi = src 126 mov esi, src //%1,%%esi // esi = src
111 mov eax, volume //%3,%%eax // eax = volume 127 mov eax, volume //%3,%%eax // eax = volume
112 movd mm0, eax //%%eax,%%mm0 128 movd mm0, eax //%%eax,%%mm0
113 movq mm1, mm0 //%%mm0,%%mm1 129 movq mm1, mm0 //%%mm0,%%mm1
114 psllq mm0, 16 //$16,%%mm0 130 psllq mm0, 16 //$16,%%mm0
122 cmp ebx, 0 //$0,%%ebx 138 cmp ebx, 0 //$0,%%ebx
123 je endS8 139 je endS8
124 #ifndef __WATCOMC__ 140 #ifndef __WATCOMC__
125 align 16 141 align 16
126 #endif 142 #endif
127 mixloopS8:pxor mm2, mm2 //%%mm2,%%mm2 // mm2 = 0 143 mixloopS8:
144 pxor mm2, mm2 //%%mm2,%%mm2 // mm2 = 0
128 movq mm1,[esi] //(%%esi),%%mm1 // mm1 = a|b|c|d|e|f|g|h 145 movq mm1,[esi] //(%%esi),%%mm1 // mm1 = a|b|c|d|e|f|g|h
129 movq mm3, mm1 //%%mm1,%%mm3 // mm3 = a|b|c|d|e|f|g|h 146 movq mm3, mm1 //%%mm1,%%mm3 // mm3 = a|b|c|d|e|f|g|h
130 // on va faire le "sign extension" en faisant un cmp avec 0 qui retourne 1 si <0, 0 si >0 147 // on va faire le "sign extension" en faisant un cmp avec 0 qui retourne 1 si <0, 0 si >0
131 pcmpgtb mm2, mm1 //%%mm1,%%mm2 // mm2 = 11111111|00000000|00000000.... 148 pcmpgtb mm2, mm1 //%%mm1,%%mm2 // mm2 = 11111111|00000000|00000000....
132 punpckhbw mm1, mm2 //%%mm2,%%mm1 // mm1 = 0|a|0|b|0|c|0|d 149 punpckhbw mm1, mm2 //%%mm2,%%mm1 // mm1 = 0|a|0|b|0|c|0|d
133 punpcklbw mm3, mm2 //%%mm2,%%mm3 // mm3 = 0|e|0|f|0|g|0|h 150 punpcklbw mm3, mm2 //%%mm2,%%mm3 // mm3 = 0|e|0|f|0|g|0|h
134 movq mm2,[edi] //(%%edi),%%mm2 // mm2 = destination 151 movq mm2,[edi] //(%%edi),%%mm2 // mm2 = destination
135 pmullw mm1, mm0 //%%mm0,%%mm1 // mm1 = v*a|v*b|v*c|v*d 152 pmullw mm1, mm0 //%%mm0,%%mm1 // mm1 = v*a|v*b|v*c|v*d
136 add esi, 8 //$8,%%esi 153 add esi, 8 //$8,%%esi
137 pmullw mm3, mm0 //%%mm0,%%mm3 // mm3 = v*e|v*f|v*g|v*h 154 pmullw mm3, mm0 //%%mm0,%%mm3 // mm3 = v*e|v*f|v*g|v*h
138 psraw mm1, 7 //$7,%%mm1 // mm1 = v*a/128|v*b/128|v*c/128|v*d/128 155 psraw mm1, 7 //$7,%%mm1 // mm1 = v*a/128|v*b/128|v*c/128|v*d/128
139 psraw mm3, 7 //$7,%%mm3 // mm3 = v*e/128|v*f/128|v*g/128|v*h/128 156 psraw mm3, 7 //$7,%%mm3 // mm3 = v*e/128|v*f/128|v*g/128|v*h/128
140 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/128 157 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
141 paddsb mm3, mm2 //%%mm2,%%mm3 // add to destination buffer 158 paddsb mm3, mm2 //%%mm2,%%mm3 // add to destination buffer
142 movq[edi], mm3 //%%mm3,(%%edi) // store back to ram 159 movq[edi], mm3 //%%mm3,(%%edi) // store back to ram
143 add edi, 8 //$8,%%edi 160 add edi, 8 //$8,%%edi
144 dec ebx //%%ebx 161 dec ebx //%%ebx
145 jnz mixloopS8 endS8:emms pop ebx pop esi pop edi} 162 jnz mixloopS8
163 endS8:
164 emms
165 pop ebx
166 pop esi
167 pop edi
168 }
169 /* *INDENT-ON* */
146 } 170 }
147 171
148 #endif /* SDL_ASSEMBLY_ROUTINES */ 172 #endif /* SDL_ASSEMBLY_ROUTINES */
173
149 /* vi: set ts=4 sw=4 expandtab: */ 174 /* vi: set ts=4 sw=4 expandtab: */