--- mem_transfer_3dne.asm 2002/12/19 00:41:11 1.1 +++ mem_transfer_3dne.asm 2003/02/15 15:22:19 1.2 @@ -0,0 +1,431 @@ +;/************************************************************************** +; * +; * XVID MPEG-4 VIDEO CODEC +; * mmx 8bit<->16bit transfers +; * +; * This program is an implementation of a part of one or more MPEG-4 +; * Video tools as specified in ISO/IEC 14496-2 standard. Those intending +; * to use this software module in hardware or software products are +; * advised that its use may infringe existing patents or copyrights, and +; * any such use would be at such party's own risk. The original +; * developer of this software module and his/her company, and subsequent +; * editors and their companies, will have no liability for use of this +; * software or modifications or derivatives thereof. +; * +; * This program is free software; you can redistribute it and/or modify +; * it under the terms of the GNU General Public License as published by +; * the Free Software Foundation; either version 2 of the License, or +; * (at your option) any later version. +; * +; * This program 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 General Public License for more details. +; * +; * You should have received a copy of the GNU General Public License +; * along with this program; if not, write to the Free Software +; * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. +; * +; *************************************************************************/ + +; these 3dne functions are compatible with iSSE, but are optimized specifically for +; K7 pipelines +; +;------------------------------------------------------------------------------ +; 09.12.2002 Athlon optimizations contributed by Jaan Kalda +;------------------------------------------------------------------------------ + + +bits 32 +%ifdef FORMAT_COFF +section .data data +%else +section .data data align=16 +%endif + + +align 8 +mm_zero: +dd 0,0 + + +%macro cglobal 1 + %ifdef PREFIX + global _%1 + %define %1 _%1 + %else + global %1 + %endif +%endmacro +%macro nop4 0 +DB 08Dh,074h,026h,0 +%endmacro + +section .text + +cglobal transfer_8to16copy_3dne +cglobal transfer_16to8copy_3dne +cglobal transfer_8to16sub_3dne +cglobal transfer_8to16subro_3dne +cglobal transfer_8to16sub2_3dne +cglobal transfer_16to8add_3dne +cglobal transfer8x8_copy_3dne + +;=========================================================================== +; +; void transfer_8to16copy_3dne(int16_t * const dst, +; const uint8_t * const src, +; uint32_t stride); +; +;=========================================================================== + +align 16 +transfer_8to16copy_3dne: + + mov eax, [esp+ 8] ; Src + mov edx, [esp+12] ; Stride + mov ecx, [esp+ 4] ; Dst + punpcklbw mm0, [byte eax] + punpcklbw mm1, [eax+4] + movq mm2,[eax+edx] + movq mm3,[eax+edx] + pxor mm7,mm7 + lea eax,[eax+2*edx] + punpcklbw mm2,mm7 + punpckhbw mm3,mm7 + psrlw mm0,8 + psrlw mm1,8 + punpcklbw mm4, [eax] + punpcklbw mm5, [eax+edx+4] + movq [byte ecx+0*64], mm0 + movq [ecx+0*64+8], mm1 + punpcklbw mm6, [eax+edx] + punpcklbw mm7, [eax+4] + lea eax,[byte eax+2*edx] + psrlw mm4,8 + psrlw mm5,8 + punpcklbw mm0, [eax] + punpcklbw mm1, [eax+edx+4] + movq [ecx+0*64+16], mm2 + movq [ecx+0*64+24], mm3 + psrlw mm6,8 + psrlw mm7,8 + punpcklbw mm2, [eax+edx] + punpcklbw mm3, [eax+4] + lea eax,[byte eax+2*edx] + movq [byte ecx+0*64+32], mm4 + movq [ecx+0*64+56], mm5 + psrlw mm0,8 + psrlw mm1,8 + punpcklbw mm4, [eax] + punpcklbw mm5, [eax+edx+4] + movq [byte ecx+0*64+48], mm6 + movq [ecx+0*64+40], mm7 + psrlw mm2,8 + psrlw mm3,8 + punpcklbw mm6, [eax+edx] + punpcklbw mm7, [eax+4] + movq [byte ecx+1*64], mm0 + movq [ecx+1*64+24], mm1 + psrlw mm4,8 + psrlw mm5,8 + movq [ecx+1*64+16], mm2 + movq [ecx+1*64+8], mm3 + psrlw mm6,8 + psrlw mm7,8 + movq [byte ecx+1*64+32], mm4 + movq [ecx+1*64+56], mm5 + movq [byte ecx+1*64+48], mm6 + movq [ecx+1*64+40], mm7 +ret + + + +;=========================================================================== +; +; void transfer_16to8copy_3dne(uint8_t * const dst, +; const int16_t * const src, +; uint32_t stride); +; +;=========================================================================== + +align 16 +transfer_16to8copy_3dne: + + mov eax, [esp+ 8] ; Src + mov ecx, [esp+ 4] ; Dst + mov edx, [esp+12] ; Stride + + movq mm0, [byte eax+0*32] + packuswb mm0,[eax+0*32+8] + movq mm1, [eax+0*32+16] + packuswb mm1,[eax+0*32+24] + movq mm5, [eax+2*32+16] + movq mm2, [eax+1*32] + packuswb mm2, [eax+1*32+8] + movq mm3, [eax+1*32+16] + packuswb mm3, [eax+1*32+24] + movq mm6, [eax+3*32] + movq mm4, [eax+2*32] + packuswb mm4, [eax+2*32+8] + packuswb mm5, [eax+2*32+24] + movq mm7, [eax+3*32+16] + packuswb mm7, [eax+3*32+24] + packuswb mm6, [eax+3*32+8] + movq [ecx], mm0 + lea eax,[3*edx] + add eax,ecx + movq [ecx+edx], mm1 + movq [ecx+2*edx], mm2 + movq [byte eax], mm3 + movq [ecx+4*edx], mm4 + lea ecx,[byte ecx+4*edx] + movq [eax+2*edx], mm5 + movq [eax+4*edx], mm7 + movq [ecx+2*edx], mm6 + ret + +;=========================================================================== +; +; void transfer_8to16sub_3dne(int16_t * const dct, +; uint8_t * const cur, +; const uint8_t * const ref, +; const uint32_t stride); +; +;=========================================================================== +;/************************************************************************** +; * +; * History: +; * +; * 27.12.2001 renamed from 'compensate' to 'transfer_8to16sub' +; * 02.12.2001 loop unrolled, code runs 10% faster now (Isibaar) +; * 30.11.2001 16 pixels are processed per iteration (Isibaar) +; * 30.11.2001 .text missing +; * 06.11.2001 inital version; (c)2001 peter ross +; * +; *************************************************************************/ + +; when second argument == 1, reference (ebx) block is to current (eax) +%macro COPY_8_TO_16_SUB 2 + movq mm1, [eax] ; cur + movq mm0, mm1 + movq mm4, [ecx] ; ref + movq mm6, mm4 +%if %2 == 1 + movq [eax], mm4 +%endif + punpckhbw mm1, mm7 + punpckhbw mm6, mm7 + punpcklbw mm4, mm7 +align 8 + movq mm2, [byte eax+edx] + punpcklbw mm0, mm7 + movq mm3, [byte eax+edx] + punpcklbw mm2, mm7 + movq mm5, [byte ecx+edx] ; ref + punpckhbw mm3, mm7 +%if %2 == 1 + movq [byte eax+edx], mm5 +%endif + psubsw mm1, mm6 + + movq mm6, mm5 + psubsw mm0, mm4 +%if (%1 < 3) + lea eax,[eax+2*edx] + lea ecx,[ecx+2*edx] +%else + mov ecx,[esp] + add esp,byte 4 +%endif + movq [edi+%1*32+ 8], mm1 + movq [byte edi+%1*32+ 0], mm0 ; dst + punpcklbw mm5, mm7 + punpckhbw mm6, mm7 + psubsw mm2, mm5 + psubsw mm3, mm6 + movq [edi+%1*32+16], mm2 + movq [edi+%1*32+24], mm3 +%endmacro + +align 16 +transfer_8to16sub_3dne: + mov eax, [esp + 8] ; Cur + mov ecx, [esp +12] ; Ref + push edi + mov edx, [dword esp+4+16] ; Stride + mov edi, [esp+4+ 4] ; Dst + pxor mm7, mm7 + nop +align 4 + COPY_8_TO_16_SUB 0, 1 + COPY_8_TO_16_SUB 1, 1 + COPY_8_TO_16_SUB 2, 1 + COPY_8_TO_16_SUB 3, 1 + mov edi,ecx + ret + +align 16 +transfer_8to16subro_3dne: + mov eax, [esp + 8] ; Cur + mov ecx, [esp +12] ; Ref + push edi + mov edx, [dword esp+4+16] ; Stride + mov edi, [esp+4+ 4] ; Dst + pxor mm7, mm7 + nop +align 4 + COPY_8_TO_16_SUB 0, 0 + COPY_8_TO_16_SUB 1, 0 + COPY_8_TO_16_SUB 2, 0 + COPY_8_TO_16_SUB 3, 0 + mov edi,ecx + ret + + +;=========================================================================== +; +; void transfer_8to16sub2_3dne(int16_t * const dct, +; uint8_t * const cur, +; const uint8_t * ref1, +; const uint8_t * ref2, +; const uint32_t stride) +; +;=========================================================================== + +%macro COPY_8_TO_16_SUB2_SSE 1 + db 0Fh, 6Fh, 44h, 20h, 00 ;movq mm0, [byte eax] ; cur + punpcklbw mm0, mm7 + movq mm2, [byte eax+edx] + punpcklbw mm2, mm7 + db 0Fh, 6Fh, 4ch, 20h, 00 ;movq mm1, [byte eax] + punpckhbw mm1, mm7 + movq mm3, [byte eax+edx] + punpckhbw mm3, mm7 + + movq mm4, [byte ebx] ; ref1 + pavgb mm4, [byte esi] ; ref2 + movq mm5, [ebx+edx] ; ref + pavgb mm5, [esi+edx] ; ref2 + movq mm6, mm4 + punpcklbw mm4, mm7 + punpckhbw mm6, mm7 +%if (%1 < 3) + lea esi,[esi+2*edx] + lea ebx,[byte ebx+2*edx] + lea eax,[eax+2*edx] +%else + mov esi,[esp] + mov ebx,[esp+4] + add esp,byte 8 +%endif + psubsw mm0, mm4 + psubsw mm1, mm6 + movq mm6, mm5 + punpcklbw mm5, mm7 + punpckhbw mm6, mm7 + psubsw mm2, mm5 + psubsw mm3, mm6 + movq [byte ecx+%1*32+ 0], mm0 ; dst + movq [ecx+%1*32+ 8], mm1 + movq [ecx+%1*32+16], mm2 + movq [ecx+%1*32+24], mm3 +%endmacro + +align 16 +transfer_8to16sub2_3dne: + mov edx, [esp +20] ; Stride + mov ecx, [esp + 4] ; Dst + mov eax, [esp + 8] ; Cur + push ebx + lea ebp,[byte ebp] + mov ebx, [esp+4+12] ; Ref1 + push esi + pxor mm7, mm7 + mov esi, [esp+8+16] ; Ref2 + nop4 + COPY_8_TO_16_SUB2_SSE 0 + COPY_8_TO_16_SUB2_SSE 1 + COPY_8_TO_16_SUB2_SSE 2 + COPY_8_TO_16_SUB2_SSE 3 + + ret + + +;=========================================================================== +; +; void transfer_16to8add_3dne(uint8_t * const dst, +; const int16_t * const src, +; uint32_t stride); +; +;=========================================================================== + +%macro COPY_16_TO_8_ADD 1 + db 0Fh, 6Fh, 44h, 21h, 00 ;movq mm0, [byte ecx] + punpcklbw mm0, mm7 + movq mm2, [byte ecx+edx] + punpcklbw mm2, mm7 + db 0Fh, 6Fh, 4ch, 21h, 00 ;movq mm1, [byte ecx] + punpckhbw mm1, mm7 + movq mm3, [byte ecx+edx] + punpckhbw mm3, mm7 + paddsw mm0, [byte eax+%1*32+ 0] + paddsw mm1, [eax+%1*32+ 8] + paddsw mm2, [eax+%1*32+16] + paddsw mm3, [eax+%1*32+24] + packuswb mm0, mm1 + packuswb mm2, mm3 + mov esp,esp + movq [byte ecx], mm0 + movq [ecx+edx], mm2 +%endmacro + + +align 16 +transfer_16to8add_3dne: + mov ecx, [esp+ 4] ; Dst + mov edx, [esp+12] ; Stride + mov eax, [esp+ 8] ; Src + pxor mm7, mm7 + nop + + COPY_16_TO_8_ADD 0 + lea ecx,[byte ecx+2*edx] + COPY_16_TO_8_ADD 1 + lea ecx,[byte ecx+2*edx] + COPY_16_TO_8_ADD 2 + lea ecx,[byte ecx+2*edx] + COPY_16_TO_8_ADD 3 + ret + +;=========================================================================== +; +; void transfer8x8_copy_3dne(uint8_t * const dst, +; const uint8_t * const src, +; const uint32_t stride); +; +; +;=========================================================================== + +%macro COPY_8_TO_8 0 + movq mm0, [byte eax] + movq mm1, [eax+edx] + movq [byte ecx], mm0 + lea eax,[byte eax+2*edx] + movq [ecx+edx], mm1 +%endmacro + +align 16 +transfer8x8_copy_3dne: + mov eax, [esp+ 8] ; Src + mov edx, [esp+12] ; Stride + mov ecx, [esp+ 4] ; Dst + + COPY_8_TO_8 + lea ecx,[byte ecx+2*edx] + COPY_8_TO_8 + lea ecx,[byte ecx+2*edx] + COPY_8_TO_8 + lea ecx,[byte ecx+2*edx] + COPY_8_TO_8 + ret