Parent Directory
| 
Revision Log
Revision 1790 - (view) (download)
| 1 : | Skal | 1709 | ;/***************************************************************************** |
| 2 : | ; * | ||
| 3 : | ; * XVID MPEG-4 VIDEO CODEC | ||
| 4 : | ; * - GMC core functions - | ||
| 5 : | ; * Copyright(C) 2006 Pascal Massimino <skal@planet-d.net> | ||
| 6 : | ; * | ||
| 7 : | ; * This file is part of XviD, a free MPEG-4 video encoder/decoder | ||
| 8 : | ; * | ||
| 9 : | ; * XviD is free software; you can redistribute it and/or modify it | ||
| 10 : | ; * under the terms of the GNU General Public License as published by | ||
| 11 : | ; * the Free Software Foundation; either version 2 of the License, or | ||
| 12 : | ; * (at your option) any later version. | ||
| 13 : | ; * | ||
| 14 : | ; * This program is distributed in the hope that it will be useful, | ||
| 15 : | ; * but WITHOUT ANY WARRANTY; without even the implied warranty of | ||
| 16 : | ; * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | ||
| 17 : | ; * GNU General Public License for more details. | ||
| 18 : | ; * | ||
| 19 : | ; * You should have received a copy of the GNU General Public License | ||
| 20 : | ; * along with this program; if not, write to the Free Software | ||
| 21 : | ; * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA | ||
| 22 : | ; * | ||
| 23 : | Isibaar | 1790 | ; * $Id: gmc_mmx.asm,v 1.3 2008-08-19 09:06:48 Isibaar Exp $ |
| 24 : | Skal | 1709 | ; * |
| 25 : | ; *************************************************************************/ | ||
| 26 : | |||
| 27 : | ;/************************************************************************** | ||
| 28 : | ; * | ||
| 29 : | ; * History: | ||
| 30 : | ; * | ||
| 31 : | ; * Jun 14 2006: initial version (during Germany/Poland match;) | ||
| 32 : | ; * | ||
| 33 : | ; *************************************************************************/ | ||
| 34 : | |||
| 35 : | bits 32 | ||
| 36 : | |||
| 37 : | %macro cglobal 1 | ||
| 38 : | %ifdef PREFIX | ||
| 39 : | %ifdef MARK_FUNCS | ||
| 40 : | global _%1:function %1.endfunc-%1 | ||
| 41 : | %define %1 _%1:function %1.endfunc-%1 | ||
| 42 : | %else | ||
| 43 : | global _%1 | ||
| 44 : | %define %1 _%1 | ||
| 45 : | %endif | ||
| 46 : | %else | ||
| 47 : | %ifdef MARK_FUNCS | ||
| 48 : | global %1:function %1.endfunc-%1 | ||
| 49 : | %else | ||
| 50 : | global %1 | ||
| 51 : | %endif | ||
| 52 : | %endif | ||
| 53 : | %endmacro | ||
| 54 : | |||
| 55 : | ;////////////////////////////////////////////////////////////////////// | ||
| 56 : | |||
| 57 : | cglobal xvid_GMC_Core_Lin_8_mmx | ||
| 58 : | cglobal xvid_GMC_Core_Lin_8_sse2 | ||
| 59 : | |||
| 60 : | ;////////////////////////////////////////////////////////////////////// | ||
| 61 : | |||
| 62 : | %ifdef FORMAT_COFF | ||
| 63 : | SECTION .rodata | ||
| 64 : | %else | ||
| 65 : | SECTION .rodata align=16 | ||
| 66 : | %endif | ||
| 67 : | |||
| 68 : | align 16 | ||
| 69 : | Cst16: | ||
| 70 : | times 8 dw 16 | ||
| 71 : | |||
| 72 : | SECTION .text | ||
| 73 : | |||
| 74 : | ;////////////////////////////////////////////////////////////////////// | ||
| 75 : | ;// mmx version | ||
| 76 : | |||
| 77 : | %macro GMC_4_SSE 2 ; %1: i %2: out reg (mm5 or mm6) | ||
| 78 : | |||
| 79 : | pcmpeqw mm0, mm0 | ||
| 80 : | movq mm1, [eax+2*(%1) ] ; u0 | u1 | u2 | u3 | ||
| 81 : | psrlw mm0, 12 ; mask 0x000f | ||
| 82 : | movq mm2, [eax+2*(%1)+2*16] ; v0 | v1 | v2 | v3 | ||
| 83 : | |||
| 84 : | pand mm1, mm0 ; u0 | ||
| 85 : | pand mm2, mm0 ; v0 | ||
| 86 : | |||
| 87 : | movq mm0, [Cst16] | ||
| 88 : | movq mm3, mm1 ; u | ... | ||
| 89 : | movq mm4, mm0 | ||
| 90 : | pmullw mm3, mm2 ; u.v | ||
| 91 : | psubw mm0, mm1 ; 16-u | ||
| 92 : | psubw mm4, mm2 ; 16-v | ||
| 93 : | pmullw mm2, mm0 ; (16-u).v | ||
| 94 : | pmullw mm0, mm4 ; (16-u).(16-v) | ||
| 95 : | pmullw mm1, mm4 ; u .(16-v) | ||
| 96 : | |||
| 97 : | movd mm4, [ecx+edx +%1] ; src2 | ||
| 98 : | movd %2, [ecx+edx+1+%1] ; src3 | ||
| 99 : | punpcklbw mm4, mm7 | ||
| 100 : | punpcklbw %2, mm7 | ||
| 101 : | pmullw mm2, mm4 | ||
| 102 : | pmullw mm3, %2 | ||
| 103 : | |||
| 104 : | movd mm4, [ecx +%1] ; src0 | ||
| 105 : | movd %2, [ecx +1+%1] ; src1 | ||
| 106 : | punpcklbw mm4, mm7 | ||
| 107 : | punpcklbw %2, mm7 | ||
| 108 : | pmullw mm4, mm0 | ||
| 109 : | pmullw %2, mm1 | ||
| 110 : | |||
| 111 : | paddw mm2, mm3 | ||
| 112 : | paddw %2, mm4 | ||
| 113 : | |||
| 114 : | paddw %2, mm2 | ||
| 115 : | %endmacro | ||
| 116 : | |||
| 117 : | align 16 | ||
| 118 : | xvid_GMC_Core_Lin_8_mmx: | ||
| 119 : | mov eax, [esp + 8] ; Offsets | ||
| 120 : | mov ecx, [esp +12] ; Src0 | ||
| 121 : | mov edx, [esp +16] ; BpS | ||
| 122 : | |||
| 123 : | pxor mm7, mm7 | ||
| 124 : | |||
| 125 : | GMC_4_SSE 0, mm5 | ||
| 126 : | GMC_4_SSE 4, mm6 | ||
| 127 : | |||
| 128 : | ; pshufw mm4, [esp +20], 01010101b ; Rounder (bits [16..31]) | ||
| 129 : | movd mm4, [esp+20] ; Rounder (bits [16..31]) | ||
| 130 : | mov eax, [esp + 4] ; Dst | ||
| 131 : | punpcklwd mm4, mm4 | ||
| 132 : | punpckhdq mm4, mm4 | ||
| 133 : | |||
| 134 : | paddw mm5, mm4 | ||
| 135 : | paddw mm6, mm4 | ||
| 136 : | psrlw mm5, 8 | ||
| 137 : | psrlw mm6, 8 | ||
| 138 : | packuswb mm5, mm6 | ||
| 139 : | movq [eax], mm5 | ||
| 140 : | |||
| 141 : | ret | ||
| 142 : | .endfunc | ||
| 143 : | |||
| 144 : | ;////////////////////////////////////////////////////////////////////// | ||
| 145 : | ;// SSE2 version | ||
| 146 : | |||
| 147 : | %macro GMC_8_SSE2 0 | ||
| 148 : | |||
| 149 : | pcmpeqw xmm0, xmm0 | ||
| 150 : | movdqa xmm1, [eax ] ; u... | ||
| 151 : | psrlw xmm0, 12 ; mask = 0x000f | ||
| 152 : | movdqa xmm2, [eax+2*16] ; v... | ||
| 153 : | pand xmm1, xmm0 | ||
| 154 : | pand xmm2, xmm0 | ||
| 155 : | |||
| 156 : | movdqa xmm0, [Cst16] | ||
| 157 : | movdqa xmm3, xmm1 ; u | ... | ||
| 158 : | movdqa xmm4, xmm0 | ||
| 159 : | pmullw xmm3, xmm2 ; u.v | ||
| 160 : | psubw xmm0, xmm1 ; 16-u | ||
| 161 : | psubw xmm4, xmm2 ; 16-v | ||
| 162 : | pmullw xmm2, xmm0 ; (16-u).v | ||
| 163 : | pmullw xmm0, xmm4 ; (16-u).(16-v) | ||
| 164 : | pmullw xmm1, xmm4 ; u .(16-v) | ||
| 165 : | |||
| 166 : | movq xmm4, [ecx+edx ] ; src2 | ||
| 167 : | movq xmm5, [ecx+edx+1] ; src3 | ||
| 168 : | punpcklbw xmm4, xmm7 | ||
| 169 : | punpcklbw xmm5, xmm7 | ||
| 170 : | pmullw xmm2, xmm4 | ||
| 171 : | pmullw xmm3, xmm5 | ||
| 172 : | |||
| 173 : | movq xmm4, [ecx ] ; src0 | ||
| 174 : | movq xmm5, [ecx +1] ; src1 | ||
| 175 : | punpcklbw xmm4, xmm7 | ||
| 176 : | punpcklbw xmm5, xmm7 | ||
| 177 : | pmullw xmm4, xmm0 | ||
| 178 : | pmullw xmm5, xmm1 | ||
| 179 : | |||
| 180 : | paddw xmm2, xmm3 | ||
| 181 : | paddw xmm5, xmm4 | ||
| 182 : | |||
| 183 : | paddw xmm5, xmm2 | ||
| 184 : | %endmacro | ||
| 185 : | |||
| 186 : | align 16 | ||
| 187 : | xvid_GMC_Core_Lin_8_sse2: | ||
| 188 : | mov eax, [esp + 8] ; Offsets | ||
| 189 : | mov ecx, [esp +12] ; Src0 | ||
| 190 : | mov edx, [esp +16] ; BpS | ||
| 191 : | |||
| 192 : | pxor xmm7, xmm7 | ||
| 193 : | |||
| 194 : | GMC_8_SSE2 | ||
| 195 : | |||
| 196 : | Skal | 1756 | movd xmm4, [esp +20] |
| 197 : | pshuflw xmm4, xmm4, 01010101b ; Rounder (bits [16..31]) | ||
| 198 : | Skal | 1709 | punpckldq xmm4, xmm4 |
| 199 : | mov eax, [esp + 4] ; Dst | ||
| 200 : | |||
| 201 : | paddw xmm5, xmm4 | ||
| 202 : | psrlw xmm5, 8 | ||
| 203 : | packuswb xmm5, xmm5 | ||
| 204 : | movq [eax], xmm5 | ||
| 205 : | |||
| 206 : | ret | ||
| 207 : | .endfunc | ||
| 208 : | |||
| 209 : | ;////////////////////////////////////////////////////////////////////// | ||
| 210 : | Isibaar | 1790 | |
| 211 : | %ifidn __OUTPUT_FORMAT__,elf | ||
| 212 : | section ".note.GNU-stack" noalloc noexec nowrite progbits | ||
| 213 : | %endif |
| No admin address has been configured | ViewVC Help |
| Powered by ViewVC 1.0.4 |