;/**************************************************************************** ; * ; * XVID MPEG-4 VIDEO CODEC ; * - K7 optimized SAD operators - ; * ; * Copyright(C) 2001 Peter Ross ; * 2001 Michael Militzer ; * 2002 Pascal Massimino ; * 2004 Andre Werthmann ; * ; * 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., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA ; * ; * $Id: sad_xmm.asm,v 1.2 2008-08-19 09:06:48 Isibaar Exp $ ; * ; ***************************************************************************/ BITS 64 %macro cglobal 1 %ifdef PREFIX %ifdef MARK_FUNCS global _%1:function %1.endfunc-%1 %define %1 _%1:function %1.endfunc-%1 %else global _%1 %define %1 _%1 %endif %else %ifdef MARK_FUNCS global %1:function %1.endfunc-%1 %else global %1 %endif %endif %endmacro ;============================================================================= ; Read only data ;============================================================================= %ifdef FORMAT_COFF SECTION .rodata %else SECTION .rodata align=16 %endif ALIGN 16 mmx_one: times 4 dw 1 ;============================================================================= ; Helper macros ;============================================================================= %macro SAD_16x16_SSE 0 movq mm0, [rax] psadbw mm0, [rdx] movq mm1, [rax+8] add rax, rcx psadbw mm1, [rdx+8] paddusw mm5, mm0 add rdx, rcx paddusw mm6, mm1 %endmacro %macro SAD_8x8_SSE 0 movq mm0, [rax] movq mm1, [rax+rcx] psadbw mm0, [rdx] psadbw mm1, [rdx+rcx] add rax, rbx add rdx, rbx paddusw mm5, mm0 paddusw mm6, mm1 %endmacro %macro SADBI_16x16_SSE 0 movq mm0, [rax] movq mm1, [rax+8] movq mm2, [rdx] movq mm3, [rdx+8] pavgb mm2, [rbx] add rdx, rcx pavgb mm3, [rbx+8] add rbx, rcx psadbw mm0, mm2 add rax, rcx psadbw mm1, mm3 paddusw mm5, mm0 paddusw mm6, mm1 %endmacro %macro SADBI_8x8_XMM 0 movq mm0, [rax] movq mm1, [rax+rcx] movq mm2, [rdx] movq mm3, [rdx+rcx] pavgb mm2, [rbx] lea rdx, [rdx+2*rcx] pavgb mm3, [rbx+rcx] lea rbx, [rbx+2*rcx] psadbw mm0, mm2 lea rax, [rax+2*rcx] psadbw mm1, mm3 paddusw mm5, mm0 paddusw mm6, mm1 %endmacro %macro MEAN_16x16_SSE 0 movq mm0, [rax] movq mm1, [rax+8] psadbw mm0, mm7 psadbw mm1, mm7 add rax, rcx paddw mm5, mm0 paddw mm6, mm1 %endmacro %macro ABS_16x16_SSE 0 movq mm0, [rax] movq mm1, [rax+8] psadbw mm0, mm4 psadbw mm1, mm4 lea rax, [rax+rcx] paddw mm5, mm0 paddw mm6, mm1 %endmacro ;============================================================================= ; Code ;============================================================================= SECTION .text align=16 cglobal sad16_x86_64 cglobal sad8_x86_64 cglobal sad16bi_x86_64 cglobal sad8bi_x86_64 cglobal dev16_x86_64 cglobal sad16v_x86_64 ;----------------------------------------------------------------------------- ; ; uint32_t sad16_x86_64(const uint8_t * const cur, ; const uint8_t * const ref, ; const uint32_t stride, ; const uint32_t best_sad); ; ;----------------------------------------------------------------------------- ALIGN 16 sad16_x86_64: mov rcx, rdx ; stride mov rdx, rsi ; src2 (64bit pointer) mov rax, rdi ; src1 (64bit pointer) pxor mm5, mm5 ; accum1 pxor mm6, mm6 ; accum2 SAD_16x16_SSE SAD_16x16_SSE SAD_16x16_SSE SAD_16x16_SSE SAD_16x16_SSE SAD_16x16_SSE SAD_16x16_SSE SAD_16x16_SSE SAD_16x16_SSE SAD_16x16_SSE SAD_16x16_SSE SAD_16x16_SSE SAD_16x16_SSE SAD_16x16_SSE SAD_16x16_SSE SAD_16x16_SSE paddusw mm6,mm5 movd eax, mm6 ret .endfunc ;----------------------------------------------------------------------------- ; ; uint32_t sad8_x86_64(const uint8_t * const cur, ; const uint8_t * const ref, ; const uint32_t stride); ; ;----------------------------------------------------------------------------- ALIGN 16 sad8_x86_64: mov rcx, rdx ; stride mov rdx, rsi ; src2 mov rax, rdi ; src1 push rbx lea rbx, [rcx+rcx] pxor mm5, mm5 ; accum1 pxor mm6, mm6 ; accum2 SAD_8x8_SSE SAD_8x8_SSE SAD_8x8_SSE movq mm0, [rax] movq mm1, [rax+rcx] psadbw mm0, [rdx] psadbw mm1, [rdx+rcx] pop rbx paddusw mm5,mm0 paddusw mm6,mm1 paddusw mm6,mm5 movd eax, mm6 ret .endfunc ;----------------------------------------------------------------------------- ; ; uint32_t sad16bi_x86_64(const uint8_t * const cur, ; const uint8_t * const ref1, ; const uint8_t * const ref2, ; const uint32_t stride); ; ;----------------------------------------------------------------------------- ALIGN 16 sad16bi_x86_64: push rbx ; rcx is stride (4. arg) mov rbx, rdx ; ref2 mov rdx, rsi ; ref1 mov rax, rdi ; src pxor mm5, mm5 ; accum1 pxor mm6, mm6 ; accum2 SADBI_16x16_SSE SADBI_16x16_SSE SADBI_16x16_SSE SADBI_16x16_SSE SADBI_16x16_SSE SADBI_16x16_SSE SADBI_16x16_SSE SADBI_16x16_SSE SADBI_16x16_SSE SADBI_16x16_SSE SADBI_16x16_SSE SADBI_16x16_SSE SADBI_16x16_SSE SADBI_16x16_SSE SADBI_16x16_SSE SADBI_16x16_SSE paddusw mm6,mm5 movd eax, mm6 pop rbx ret .endfunc ;----------------------------------------------------------------------------- ; ; uint32_t sad8bi_x86_64(const uint8_t * const cur, ; const uint8_t * const ref1, ; const uint8_t * const ref2, ; const uint32_t stride); ; ;----------------------------------------------------------------------------- ALIGN 16 sad8bi_x86_64: push rbx ; rcx is stride mov rbx, rdx ; ref2 mov rdx, rsi ; ref1 mov rax, rdi ; src pxor mm5, mm5 ; accum1 pxor mm6, mm6 ; accum2 .Loop SADBI_8x8_XMM SADBI_8x8_XMM SADBI_8x8_XMM SADBI_8x8_XMM paddusw mm6,mm5 movd eax, mm6 pop rbx ret .endfunc ;----------------------------------------------------------------------------- ; ; uint32_t dev16_x86_64(const uint8_t * const cur, ; const uint32_t stride); ; ;----------------------------------------------------------------------------- ALIGN 16 dev16_x86_64: mov rcx, rsi ; stride mov rax, rdi ; src pxor mm7, mm7 ; zero pxor mm5, mm5 ; mean accums pxor mm6, mm6 MEAN_16x16_SSE MEAN_16x16_SSE MEAN_16x16_SSE MEAN_16x16_SSE MEAN_16x16_SSE MEAN_16x16_SSE MEAN_16x16_SSE MEAN_16x16_SSE MEAN_16x16_SSE MEAN_16x16_SSE MEAN_16x16_SSE MEAN_16x16_SSE MEAN_16x16_SSE MEAN_16x16_SSE MEAN_16x16_SSE MEAN_16x16_SSE paddusw mm6, mm5 movq mm4, mm6 psllq mm4, 32 paddd mm4, mm6 psrld mm4, 8 ; /= (16*16) packssdw mm4, mm4 packuswb mm4, mm4 ; mm4 contains the mean mov rax, rdi ; src pxor mm5, mm5 ; sums pxor mm6, mm6 ABS_16x16_SSE ABS_16x16_SSE ABS_16x16_SSE ABS_16x16_SSE ABS_16x16_SSE ABS_16x16_SSE ABS_16x16_SSE ABS_16x16_SSE ABS_16x16_SSE ABS_16x16_SSE ABS_16x16_SSE ABS_16x16_SSE ABS_16x16_SSE ABS_16x16_SSE ABS_16x16_SSE ABS_16x16_SSE paddusw mm6, mm5 movq mm7, mm6 psllq mm7, 32 paddd mm6, mm7 movd eax, mm6 ret .endfunc ;----------------------------------------------------------------------------- ;int sad16v_x86_64(const uint8_t * const cur, ; const uint8_t * const ref, ; const uint32_t stride, ; int* sad8); ;----------------------------------------------------------------------------- ALIGN 16 sad16v_x86_64: push rbx mov rbx, rcx ; sad ptr (64bit) mov rcx, rdx ; stride mov rdx, rsi ; src2 mov rax, rdi ; src1 pxor mm5, mm5 ; accum1 pxor mm6, mm6 ; accum2 pxor mm7, mm7 ; total SAD_16x16_SSE SAD_16x16_SSE SAD_16x16_SSE SAD_16x16_SSE SAD_16x16_SSE SAD_16x16_SSE SAD_16x16_SSE SAD_16x16_SSE paddusw mm7, mm5 paddusw mm7, mm6 movd [rbx], mm5 movd [rbx+4], mm6 pxor mm5, mm5 ; accum1 pxor mm6, mm6 ; accum2 SAD_16x16_SSE SAD_16x16_SSE SAD_16x16_SSE SAD_16x16_SSE SAD_16x16_SSE SAD_16x16_SSE SAD_16x16_SSE SAD_16x16_SSE paddusw mm7, mm5 paddusw mm7, mm6 movd [rbx+8], mm5 movd [rbx+12], mm6 movd eax, mm7 pop rbx ret .endfunc %ifidn __OUTPUT_FORMAT__,elf section ".note.GNU-stack" noalloc noexec nowrite progbits %endif