ref: 2b176495d3f8f8c904d8b2bc603427b1f069a181
dir: /codec/processing/src/x86/vaa.asm/
;*! ;* \copy ;* Copyright (c) 2010-2013, Cisco Systems ;* All rights reserved. ;* ;* Redistribution and use in source and binary forms, with or without ;* modification, are permitted provided that the following conditions ;* are met: ;* ;* * Redistributions of source code must retain the above copyright ;* notice, this list of conditions and the following disclaimer. ;* ;* * Redistributions in binary form must reproduce the above copyright ;* notice, this list of conditions and the following disclaimer in ;* the documentation and/or other materials provided with the ;* distribution. ;* ;* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS ;* "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT ;* LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS ;* FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE ;* COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, ;* INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, ;* BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; ;* LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER ;* CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT ;* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ;* ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE ;* POSSIBILITY OF SUCH DAMAGE. ;* ;* ;* vaa.asm ;* ;* Abstract ;* sse2 for pVaa routines ;* ;* History ;* 04/14/2010 Created ;* 06/07/2010 Added AnalysisVaaInfoIntra_sse2(ssse3) ;* 06/10/2010 Tune rc_sad_frame_sse2 and got about 40% improvement ;* 08/11/2010 Added abs_difference_mbrow_sse2 & sum_sqrsum_mbrow_sse2 ;* ;*************************************************************************/ %include "asm_inc.asm" ;*********************************************************************** ; Macros and other preprocessor constants ;*********************************************************************** %macro SUM_SQR_SSE2 3 ; dst, pSrc, zero movdqa %1, %2 punpcklbw %1, %3 punpckhbw %2, %3 pmaddwd %1, %1 pmaddwd %2, %2 paddd %1, %2 pshufd %2, %1, 04Eh ; 01001110 B paddd %1, %2 pshufd %2, %1, 0B1h ; 10110001 B paddd %1, %2 %endmacro ; END OF SUM_SQR_SSE2 %macro WELS_SAD_16x2_SSE2 3 ;esi :%1 edi:%2 ebx:%3 movdqa xmm1, [%1] movdqa xmm2, [%2] movdqa xmm3, [%1+%3] movdqa xmm4, [%2+%3] psadbw xmm1, xmm2 psadbw xmm3, xmm4 paddd xmm6, xmm1 paddd xmm6, xmm3 lea %1, [%1+%3*2] lea %2, [%2+%3*2] %endmacro ; by comparing it outperforms than phaddw(SSSE3) sets %macro SUM_WORD_8x2_SSE2 2 ; dst(pSrc), tmp ; @sum_8x2 begin pshufd %2, %1, 04Eh ; 01001110 B paddw %1, %2 pshuflw %2, %1, 04Eh ; 01001110 B paddw %1, %2 pshuflw %2, %1, 0B1h ; 10110001 B paddw %1, %2 ; end of @sum_8x2 %endmacro ; END of SUM_WORD_8x2_SSE2 %macro WELS_SAD_SUM_SQSUM_16x1_SSE2 3 ;esi:%1,edi:%2,ebx:%3 movdqa xmm1, [%1] movdqa xmm2, [%2] movdqa xmm3, xmm1 psadbw xmm3, xmm2 paddd xmm6, xmm3 movdqa xmm3, xmm1 psadbw xmm3, xmm0 paddd xmm5, xmm3 movdqa xmm2, xmm1 punpcklbw xmm1, xmm0 punpckhbw xmm2, xmm0 pmaddwd xmm1, xmm1 pmaddwd xmm2, xmm2 paddd xmm4, xmm1 paddd xmm4, xmm2 add %1, %3 add %2, %3 %endmacro %macro WELS_SAD_SUM_SQSUM_SQDIFF_16x1_SSE2 3 ;esi:%1 edi:%2 ebx:%3 movdqa xmm1, [%1] movdqa xmm2, [%2] movdqa xmm3, xmm1 psadbw xmm3, xmm2 paddd xmm7, xmm3 ; sad movdqa xmm3, xmm1 pmaxub xmm3, xmm2 pminub xmm2, xmm1 psubb xmm3, xmm2 ; diff movdqa xmm2, xmm1 psadbw xmm2, xmm0 paddd xmm6, xmm2 ; sum movdqa xmm2, xmm1 punpcklbw xmm1, xmm0 punpckhbw xmm2, xmm0 pmaddwd xmm1, xmm1 pmaddwd xmm2, xmm2 paddd xmm5, xmm1 paddd xmm5, xmm2 ; sqsum movdqa xmm1, xmm3 punpcklbw xmm1, xmm0 punpckhbw xmm3, xmm0 pmaddwd xmm1, xmm1 pmaddwd xmm3, xmm3 paddd xmm4, xmm1 paddd xmm4, xmm3 ; sqdiff add %1, %3 add %2, %3 %endmacro %macro WELS_SAD_SD_MAD_16x1_SSE2 7 ;esi:%5 edi:%6 ebx:%7 %define sad_reg %1 %define sum_cur_reg %2 %define sum_ref_reg %3 %define mad_reg %4 movdqa xmm1, [%5] movdqa xmm2, [%6] movdqa xmm3, xmm1 psadbw xmm3, xmm0 paddd sum_cur_reg, xmm3 ; sum_cur movdqa xmm3, xmm2 psadbw xmm3, xmm0 paddd sum_ref_reg, xmm3 ; sum_ref movdqa xmm3, xmm1 pmaxub xmm3, xmm2 pminub xmm2, xmm1 psubb xmm3, xmm2 ; abs diff pmaxub mad_reg, xmm3 ; max abs diff psadbw xmm3, xmm0 paddd sad_reg, xmm3 ; sad add %5, %7 add %6, %7 %endmacro %macro WELS_MAX_REG_SSE2 1 ; xmm1, xmm2, xmm3 can be used %define max_reg %1 movdqa xmm1, max_reg psrldq xmm1, 4 pmaxub max_reg, xmm1 movdqa xmm1, max_reg psrldq xmm1, 2 pmaxub max_reg, xmm1 movdqa xmm1, max_reg psrldq xmm1, 1 pmaxub max_reg, xmm1 %endmacro %macro WELS_SAD_BGD_SQDIFF_16x1_SSE2 7 ;esi:%5 edi:%6 ebx:%7 %define sad_reg %1 %define sum_reg %2 %define mad_reg %3 %define sqdiff_reg %4 movdqa xmm1, [%5] movdqa xmm2, xmm1 movdqa xmm3, xmm1 punpcklbw xmm2, xmm0 punpckhbw xmm3, xmm0 pmaddwd xmm2, xmm2 pmaddwd xmm3, xmm3 paddd xmm2, xmm3 movdqa xmm3, xmm2 psllq xmm2, 32 psrlq xmm3, 32 psllq xmm3, 32 paddd xmm2, xmm3 paddd sad_reg, xmm2 ; sqsum movdqa xmm2, [%6] movdqa xmm3, xmm1 psadbw xmm3, xmm0 paddd sum_reg, xmm3 ; sum_cur movdqa xmm3, xmm2 psadbw xmm3, xmm0 pslldq xmm3, 4 paddd sum_reg, xmm3 ; sum_ref movdqa xmm3, xmm1 pmaxub xmm3, xmm2 pminub xmm2, xmm1 psubb xmm3, xmm2 ; abs diff pmaxub mad_reg, xmm3 ; max abs diff movdqa xmm1, xmm3 psadbw xmm3, xmm0 paddd sad_reg, xmm3 ; sad movdqa xmm3, xmm1 punpcklbw xmm1, xmm0 punpckhbw xmm3, xmm0 pmaddwd xmm1, xmm1 pmaddwd xmm3, xmm3 paddd sqdiff_reg, xmm1 paddd sqdiff_reg, xmm3 ; sqdiff add %5, %7 add %6, %7 %endmacro ;*********************************************************************** ; Code ;*********************************************************************** SECTION .text %ifdef X86_32 ;*********************************************************************** ; void SampleVariance16x16_sse2( uint8_t * y_ref, int32_t y_ref_stride, uint8_t * y_src, int32_t y_src_stride,SMotionTextureUnit* pMotionTexture ); ;*********************************************************************** WELS_EXTERN SampleVariance16x16_sse2 push esi push edi push ebx sub esp, 16 %define SUM [esp] %define SUM_CUR [esp+4] %define SQR [esp+8] %define SQR_CUR [esp+12] %define PUSH_SIZE 28 ; 12 + 16 mov edi, [esp+PUSH_SIZE+4] ; y_ref mov edx, [esp+PUSH_SIZE+8] ; y_ref_stride mov esi, [esp+PUSH_SIZE+12] ; y_src mov eax, [esp+PUSH_SIZE+16] ; y_src_stride mov ecx, 010h ; height = 16 pxor xmm7, xmm7 movdqu SUM, xmm7 .hloops: movdqa xmm0, [edi] ; y_ref movdqa xmm1, [esi] ; y_src movdqa xmm2, xmm0 ; store first for future process movdqa xmm3, xmm1 ; sum += diff; movdqa xmm4, xmm0 psadbw xmm4, xmm1 ; 2 parts, [0,..,15], [64,..,79] ; to be continued for sum pshufd xmm5, xmm4, 0C6h ; 11000110 B paddw xmm4, xmm5 movd ebx, xmm4 add SUM, ebx ; sqr += diff * diff; pmaxub xmm0, xmm1 pminub xmm1, xmm2 psubb xmm0, xmm1 ; diff SUM_SQR_SSE2 xmm1, xmm0, xmm7 ; dst, pSrc, zero movd ebx, xmm1 add SQR, ebx ; sum_cur += y_src[x]; movdqa xmm0, xmm3 ; cur_orig movdqa xmm1, xmm0 punpcklbw xmm0, xmm7 punpckhbw xmm1, xmm7 paddw xmm0, xmm1 ; 8x2 SUM_WORD_8x2_SSE2 xmm0, xmm1 movd ebx, xmm0 and ebx, 0ffffh add SUM_CUR, ebx ; sqr_cur += y_src[x] * y_src[x]; SUM_SQR_SSE2 xmm0, xmm3, xmm7 ; dst, pSrc, zero movd ebx, xmm0 add SQR_CUR, ebx lea edi, [edi+edx] lea esi, [esi+eax] dec ecx jnz near .hloops mov ebx, 0 mov bx, word SUM sar ebx, 8 imul ebx, ebx mov ecx, SQR sar ecx, 8 sub ecx, ebx mov edi, [esp+PUSH_SIZE+20] ; pMotionTexture mov [edi], cx ; to store uiMotionIndex mov ebx, 0 mov bx, word SUM_CUR sar ebx, 8 imul ebx, ebx mov ecx, SQR_CUR sar ecx, 8 sub ecx, ebx mov [edi+2], cx ; to store uiTextureIndex %undef SUM %undef SUM_CUR %undef SQR %undef SQR_CUR %undef PUSH_SIZE add esp, 16 pop ebx pop edi pop esi ret ;************************************************************************************************************* ;void VAACalcSad_sse2( const uint8_t *cur_data, const uint8_t *ref_data, int32_t iPicWidth, int32_t iPicHeight ; int32_t iPicStride, int32_t *psadframe, int32_t *psad8x8) ;************************************************************************************************************* WELS_EXTERN VAACalcSad_sse2 %define cur_data esp + pushsize + 4 %define ref_data esp + pushsize + 8 %define iPicWidth esp + pushsize + 12 %define iPicHeight esp + pushsize + 16 %define iPicStride esp + pushsize + 20 %define psadframe esp + pushsize + 24 %define psad8x8 esp + pushsize + 28 %define pushsize 12 push esi push edi push ebx mov esi, [cur_data] mov edi, [ref_data] mov ebx, [iPicStride] mov edx, [psad8x8] mov eax, ebx shr dword [iPicWidth], 4 ; iPicWidth/16 shr dword [iPicHeight], 4 ; iPicHeight/16 shl eax, 4 ; iPicStride*16 pxor xmm0, xmm0 pxor xmm7, xmm7 ; iFrameSad height_loop: mov ecx, dword [iPicWidth] push esi push edi width_loop: pxor xmm6, xmm6 ; WELS_SAD_16x2_SSE2 esi,edi,ebx WELS_SAD_16x2_SSE2 esi,edi,ebx WELS_SAD_16x2_SSE2 esi,edi,ebx WELS_SAD_16x2_SSE2 esi,edi,ebx paddd xmm7, xmm6 movd [edx], xmm6 psrldq xmm6, 8 movd [edx+4], xmm6 pxor xmm6, xmm6 WELS_SAD_16x2_SSE2 esi,edi,ebx WELS_SAD_16x2_SSE2 esi,edi,ebx WELS_SAD_16x2_SSE2 esi,edi,ebx WELS_SAD_16x2_SSE2 esi,edi,ebx paddd xmm7, xmm6 movd [edx+8], xmm6 psrldq xmm6, 8 movd [edx+12], xmm6 add edx, 16 sub esi, eax sub edi, eax add esi, 16 add edi, 16 dec ecx jnz width_loop pop edi pop esi add esi, eax add edi, eax dec dword [iPicHeight] jnz height_loop mov edx, [psadframe] movdqa xmm5, xmm7 psrldq xmm7, 8 paddd xmm7, xmm5 movd [edx], xmm7 %undef cur_data %undef ref_data %undef iPicWidth %undef iPicHeight %undef iPicStride %undef psadframe %undef psad8x8 %undef pushsize pop ebx pop edi pop esi ret %else ;64-bit ;*********************************************************************** ; void SampleVariance16x16_sse2( uint8_t * y_ref, int32_t y_ref_stride, uint8_t * y_src, int32_t y_src_stride,SMotionTextureUnit* pMotionTexture ); ;*********************************************************************** WELS_EXTERN SampleVariance16x16_sse2 %define SUM r10;[esp] %define SUM_CUR r11;[esp+4] %define SQR r13;[esp+8] %define SQR_CUR r15;[esp+12] push r12 push r13 push r14 push r15 %assign push_num 4 LOAD_5_PARA PUSH_XMM 8 SIGN_EXTENSION r1,r1d SIGN_EXTENSION r3,r3d mov r12,010h pxor xmm7, xmm7 movq SUM, xmm7 movq SUM_CUR,xmm7 movq SQR,xmm7 movq SQR_CUR,xmm7 .hloops: mov r14,0 movdqa xmm0, [r0] ; y_ref movdqa xmm1, [r2] ; y_src movdqa xmm2, xmm0 ; store first for future process movdqa xmm3, xmm1 ; sum += diff; movdqa xmm4, xmm0 psadbw xmm4, xmm1 ; 2 parts, [0,..,15], [64,..,79] ; to be continued for sum pshufd xmm5, xmm4, 0C6h ; 11000110 B paddw xmm4, xmm5 movd r14d, xmm4 add SUM, r14 ; sqr += diff * diff; pmaxub xmm0, xmm1 pminub xmm1, xmm2 psubb xmm0, xmm1 ; diff SUM_SQR_SSE2 xmm1, xmm0, xmm7 ; dst, pSrc, zero movd r14d, xmm1 add SQR, r14 ; sum_cur += y_src[x]; movdqa xmm0, xmm3 ; cur_orig movdqa xmm1, xmm0 punpcklbw xmm0, xmm7 punpckhbw xmm1, xmm7 paddw xmm0, xmm1 ; 8x2 SUM_WORD_8x2_SSE2 xmm0, xmm1 movd r14d, xmm0 and r14, 0ffffh add SUM_CUR, r14 ; sqr_cur += y_src[x] * y_src[x]; SUM_SQR_SSE2 xmm0, xmm3, xmm7 ; dst, pSrc, zero movd r14d, xmm0 add SQR_CUR, r14 lea r0, [r0+r1] lea r2, [r2+r3] dec r12 jnz near .hloops mov r0, SUM sar r0, 8 imul r0, r0 mov r1, SQR sar r1, 8 sub r1, r0 mov [r4], r1w ; to store uiMotionIndex mov r0, SUM_CUR sar r0, 8 imul r0, r0 mov r1, SQR_CUR sar r1, 8 sub r1, r0 mov [r4+2], r1w ; to store uiTextureIndex POP_XMM LOAD_5_PARA_POP pop r15 pop r14 pop r13 pop r12 %assign push_num 0 ret ;************************************************************************************************************* ;void VAACalcSad_sse2( const uint8_t *cur_data, const uint8_t *ref_data, int32_t iPicWidth, int32_t iPicHeight ; int32_t iPicStride, int32_t *psadframe, int32_t *psad8x8) ;************************************************************************************************************* WELS_EXTERN VAACalcSad_sse2 %define cur_data r0 %define ref_data r1 %define iPicWidth r2 %define iPicHeight r3 %define iPicStride r4 %define psadframe r5 %define psad8x8 r6 push r12 push r13 %assign push_num 2 LOAD_7_PARA PUSH_XMM 8 SIGN_EXTENSION r2,r2d SIGN_EXTENSION r3,r3d SIGN_EXTENSION r4,r4d mov r12,r4 shr r2, 4 ; iPicWidth/16 shr r3, 4 ; iPicHeight/16 shl r12, 4 ; iPicStride*16 pxor xmm0, xmm0 pxor xmm7, xmm7 ; iFrameSad height_loop: mov r13, r2 push r0 push r1 width_loop: pxor xmm6, xmm6 WELS_SAD_16x2_SSE2 r0,r1,r4 WELS_SAD_16x2_SSE2 r0,r1,r4 WELS_SAD_16x2_SSE2 r0,r1,r4 WELS_SAD_16x2_SSE2 r0,r1,r4 paddd xmm7, xmm6 movd [r6], xmm6 psrldq xmm6, 8 movd [r6+4], xmm6 pxor xmm6, xmm6 WELS_SAD_16x2_SSE2 r0,r1,r4 WELS_SAD_16x2_SSE2 r0,r1,r4 WELS_SAD_16x2_SSE2 r0,r1,r4 WELS_SAD_16x2_SSE2 r0,r1,r4 paddd xmm7, xmm6 movd [r6+8], xmm6 psrldq xmm6, 8 movd [r6+12], xmm6 add r6, 16 sub r0, r12 sub r1, r12 add r0, 16 add r1, 16 dec r13 jnz width_loop pop r1 pop r0 add r0, r12 add r1, r12 dec r3 jnz height_loop ;mov r13, [psadframe] movdqa xmm5, xmm7 psrldq xmm7, 8 paddd xmm7, xmm5 movd [psadframe], xmm7 %undef cur_data %undef ref_data %undef iPicWidth %undef iPicHeight %undef iPicStride %undef psadframe %undef psad8x8 %undef pushsize POP_XMM LOAD_7_PARA_POP pop r13 pop r12 %assign push_num 0 ret %endif %ifdef X86_32 ;************************************************************************************************************* ;void VAACalcSadVar_sse2( const uint8_t *cur_data, const uint8_t *ref_data, int32_t iPicWidth, int32_t iPicHeight ; int32_t iPicStride, int32_t *psadframe, int32_t *psad8x8, int32_t *psum16x16, int32_t *psqsum16x16) ;************************************************************************************************************* WELS_EXTERN VAACalcSadVar_sse2 %define localsize 8 %define cur_data esp + pushsize + localsize + 4 %define ref_data esp + pushsize + localsize + 8 %define iPicWidth esp + pushsize + localsize + 12 %define iPicHeight esp + pushsize + localsize + 16 %define iPicStride esp + pushsize + localsize + 20 %define psadframe esp + pushsize + localsize + 24 %define psad8x8 esp + pushsize + localsize + 28 %define psum16x16 esp + pushsize + localsize + 32 %define psqsum16x16 esp + pushsize + localsize + 36 %define tmp_esi esp + 0 %define tmp_edi esp + 4 %define pushsize 16 push ebp push esi push edi push ebx sub esp, localsize mov esi, [cur_data] mov edi, [ref_data] mov ebx, [iPicStride] mov edx, [psad8x8] mov eax, ebx shr dword [iPicWidth], 4 ; iPicWidth/16 shr dword [iPicHeight], 4 ; iPicHeight/16 shl eax, 4 ; iPicStride*16 pxor xmm0, xmm0 pxor xmm7, xmm7 ; iFrameSad var_height_loop: mov ecx, dword [iPicWidth] mov [tmp_esi], esi mov [tmp_edi], edi var_width_loop: pxor xmm6, xmm6 ; hiQuad_loQuad pSad8x8 pxor xmm5, xmm5 ; pSum16x16 pxor xmm4, xmm4 ; sqsum_16x16 WELS_SAD_SUM_SQSUM_16x1_SSE2 esi,edi,ebx WELS_SAD_SUM_SQSUM_16x1_SSE2 esi,edi,ebx WELS_SAD_SUM_SQSUM_16x1_SSE2 esi,edi,ebx WELS_SAD_SUM_SQSUM_16x1_SSE2 esi,edi,ebx WELS_SAD_SUM_SQSUM_16x1_SSE2 esi,edi,ebx WELS_SAD_SUM_SQSUM_16x1_SSE2 esi,edi,ebx WELS_SAD_SUM_SQSUM_16x1_SSE2 esi,edi,ebx WELS_SAD_SUM_SQSUM_16x1_SSE2 esi,edi,ebx paddd xmm7, xmm6 movd [edx], xmm6 psrldq xmm6, 8 movd [edx+4], xmm6 pxor xmm6, xmm6 WELS_SAD_SUM_SQSUM_16x1_SSE2 esi,edi,ebx WELS_SAD_SUM_SQSUM_16x1_SSE2 esi,edi,ebx WELS_SAD_SUM_SQSUM_16x1_SSE2 esi,edi,ebx WELS_SAD_SUM_SQSUM_16x1_SSE2 esi,edi,ebx WELS_SAD_SUM_SQSUM_16x1_SSE2 esi,edi,ebx WELS_SAD_SUM_SQSUM_16x1_SSE2 esi,edi,ebx WELS_SAD_SUM_SQSUM_16x1_SSE2 esi,edi,ebx WELS_SAD_SUM_SQSUM_16x1_SSE2 esi,edi,ebx paddd xmm7, xmm6 movd [edx+8], xmm6 psrldq xmm6, 8 movd [edx+12], xmm6 mov ebp, [psum16x16] movdqa xmm1, xmm5 psrldq xmm1, 8 paddd xmm5, xmm1 movd [ebp], xmm5 add dword [psum16x16], 4 movdqa xmm5, xmm4 psrldq xmm5, 8 paddd xmm4, xmm5 movdqa xmm3, xmm4 psrldq xmm3, 4 paddd xmm4, xmm3 mov ebp, [psqsum16x16] movd [ebp], xmm4 add dword [psqsum16x16], 4 add edx, 16 sub esi, eax sub edi, eax add esi, 16 add edi, 16 dec ecx jnz var_width_loop mov esi, [tmp_esi] mov edi, [tmp_edi] add esi, eax add edi, eax dec dword [iPicHeight] jnz var_height_loop mov edx, [psadframe] movdqa xmm5, xmm7 psrldq xmm7, 8 paddd xmm7, xmm5 movd [edx], xmm7 add esp, localsize pop ebx pop edi pop esi pop ebp %undef cur_data %undef ref_data %undef iPicWidth %undef iPicHeight %undef iPicStride %undef psadframe %undef psad8x8 %undef psum16x16 %undef psqsum16x16 %undef tmp_esi %undef tmp_edi %undef pushsize %undef localsize ret %else ;64-bit ;************************************************************************************************************* ;void VAACalcSadVar_sse2( const uint8_t *cur_data, const uint8_t *ref_data, int32_t iPicWidth, int32_t iPicHeight ; int32_t iPicStride, int32_t *psadframe, int32_t *psad8x8, int32_t *psum16x16, int32_t *psqsum16x16) ;************************************************************************************************************* WELS_EXTERN VAACalcSadVar_sse2 %define cur_data arg1 ;r0 %define ref_data arg2 ;r1 %define iPicWidth arg3 ;r2 %define iPicHeight arg4 ;r3 %define iPicStride arg5 %define psadframe arg6 %define psad8x8 arg7 %define psum16x16 arg8 %define psqsum16x16 arg9 push r12 push r13 push r14 push r15 %assign push_num 4 PUSH_XMM 8 %ifdef WIN64 mov r4, arg5 ;iPicStride mov r5, arg6 ;psad8x8 %endif mov r14,arg7 SIGN_EXTENSION r2,r2d SIGN_EXTENSION r3,r3d SIGN_EXTENSION r4,r4d mov r13,r4 shr r2,4 shr r3,4 shl r13,4 ; iPicStride*16 pxor xmm0, xmm0 pxor xmm7, xmm7 ; iFrameSad var_height_loop: push r2 %assign push_num push_num+1 mov r11, r0 mov r12, r1 var_width_loop: pxor xmm6, xmm6 ; hiQuad_loQuad pSad8x8 pxor xmm5, xmm5 ; pSum16x16 pxor xmm4, xmm4 ; sqsum_16x16 WELS_SAD_SUM_SQSUM_16x1_SSE2 r0,r1,r4 WELS_SAD_SUM_SQSUM_16x1_SSE2 r0,r1,r4 WELS_SAD_SUM_SQSUM_16x1_SSE2 r0,r1,r4 WELS_SAD_SUM_SQSUM_16x1_SSE2 r0,r1,r4 WELS_SAD_SUM_SQSUM_16x1_SSE2 r0,r1,r4 WELS_SAD_SUM_SQSUM_16x1_SSE2 r0,r1,r4 WELS_SAD_SUM_SQSUM_16x1_SSE2 r0,r1,r4 WELS_SAD_SUM_SQSUM_16x1_SSE2 r0,r1,r4 paddd xmm7, xmm6 movd [r14], xmm6 psrldq xmm6, 8 movd [r14+4], xmm6 pxor xmm6, xmm6 WELS_SAD_SUM_SQSUM_16x1_SSE2 r0,r1,r4 WELS_SAD_SUM_SQSUM_16x1_SSE2 r0,r1,r4 WELS_SAD_SUM_SQSUM_16x1_SSE2 r0,r1,r4 WELS_SAD_SUM_SQSUM_16x1_SSE2 r0,r1,r4 WELS_SAD_SUM_SQSUM_16x1_SSE2 r0,r1,r4 WELS_SAD_SUM_SQSUM_16x1_SSE2 r0,r1,r4 WELS_SAD_SUM_SQSUM_16x1_SSE2 r0,r1,r4 WELS_SAD_SUM_SQSUM_16x1_SSE2 r0,r1,r4 paddd xmm7, xmm6 movd [r14+8], xmm6 psrldq xmm6, 8 movd [r14+12], xmm6 mov r15, psum16x16 movdqa xmm1, xmm5 psrldq xmm1, 8 paddd xmm5, xmm1 movd [r15], xmm5 add dword psum16x16, 4 movdqa xmm5, xmm4 psrldq xmm5, 8 paddd xmm4, xmm5 movdqa xmm3, xmm4 psrldq xmm3, 4 paddd xmm4, xmm3 mov r15, psqsum16x16 movd [r15], xmm4 add dword psqsum16x16, 4 add r14,16 sub r0, r13 sub r1, r13 add r0, 16 add r1, 16 dec r2 jnz var_width_loop pop r2 %assign push_num push_num-1 mov r0, r11 mov r1, r12 add r0, r13 add r1, r13 dec r3 jnz var_height_loop mov r15, psadframe movdqa xmm5, xmm7 psrldq xmm7, 8 paddd xmm7, xmm5 movd [r15], xmm7 POP_XMM pop r15 pop r14 pop r13 pop r12 %assign push_num 0 %undef cur_data %undef ref_data %undef iPicWidth %undef iPicHeight %undef iPicStride %undef psadframe %undef psad8x8 %undef psum16x16 %undef psqsum16x16 %undef tmp_esi %undef tmp_edi %undef pushsize %undef localsize ret %endif %ifdef X86_32 ;************************************************************************************************************* ;void VAACalcSadSsd_sse2(const uint8_t *cur_data, const uint8_t *ref_data, int32_t iPicWidth, int32_t iPicHeight, ; int32_t iPicStride,int32_t *psadframe, int32_t *psad8x8, int32_t *psum16x16, int32_t *psqsum16x16, int32_t *psqdiff16x16) ;************************************************************************************************************* WELS_EXTERN VAACalcSadSsd_sse2 %define localsize 12 %define cur_data esp + pushsize + localsize + 4 %define ref_data esp + pushsize + localsize + 8 %define iPicWidth esp + pushsize + localsize + 12 %define iPicHeight esp + pushsize + localsize + 16 %define iPicStride esp + pushsize + localsize + 20 %define psadframe esp + pushsize + localsize + 24 %define psad8x8 esp + pushsize + localsize + 28 %define psum16x16 esp + pushsize + localsize + 32 %define psqsum16x16 esp + pushsize + localsize + 36 %define psqdiff16x16 esp + pushsize + localsize + 40 %define tmp_esi esp + 0 %define tmp_edi esp + 4 %define tmp_sadframe esp + 8 %define pushsize 16 push ebp push esi push edi push ebx sub esp, localsize mov ecx, [iPicWidth] mov ecx, [iPicHeight] mov esi, [cur_data] mov edi, [ref_data] mov ebx, [iPicStride] mov edx, [psad8x8] mov eax, ebx shr dword [iPicWidth], 4 ; iPicWidth/16 shr dword [iPicHeight], 4 ; iPicHeight/16 shl eax, 4 ; iPicStride*16 mov ecx, [iPicWidth] mov ecx, [iPicHeight] pxor xmm0, xmm0 movd [tmp_sadframe], xmm0 sqdiff_height_loop: mov ecx, dword [iPicWidth] mov [tmp_esi], esi mov [tmp_edi], edi sqdiff_width_loop: pxor xmm7, xmm7 ; hiQuad_loQuad pSad8x8 pxor xmm6, xmm6 ; pSum16x16 pxor xmm5, xmm5 ; sqsum_16x16 four dword pxor xmm4, xmm4 ; sqdiff_16x16 four Dword WELS_SAD_SUM_SQSUM_SQDIFF_16x1_SSE2 esi,edi,ebx WELS_SAD_SUM_SQSUM_SQDIFF_16x1_SSE2 esi,edi,ebx WELS_SAD_SUM_SQSUM_SQDIFF_16x1_SSE2 esi,edi,ebx WELS_SAD_SUM_SQSUM_SQDIFF_16x1_SSE2 esi,edi,ebx WELS_SAD_SUM_SQSUM_SQDIFF_16x1_SSE2 esi,edi,ebx WELS_SAD_SUM_SQSUM_SQDIFF_16x1_SSE2 esi,edi,ebx WELS_SAD_SUM_SQSUM_SQDIFF_16x1_SSE2 esi,edi,ebx WELS_SAD_SUM_SQSUM_SQDIFF_16x1_SSE2 esi,edi,ebx movdqa xmm1, xmm7 movd [edx], xmm7 psrldq xmm7, 8 paddd xmm1, xmm7 movd [edx+4], xmm7 movd ebp, xmm1 add [tmp_sadframe], ebp pxor xmm7, xmm7 WELS_SAD_SUM_SQSUM_SQDIFF_16x1_SSE2 esi,edi,ebx WELS_SAD_SUM_SQSUM_SQDIFF_16x1_SSE2 esi,edi,ebx WELS_SAD_SUM_SQSUM_SQDIFF_16x1_SSE2 esi,edi,ebx WELS_SAD_SUM_SQSUM_SQDIFF_16x1_SSE2 esi,edi,ebx WELS_SAD_SUM_SQSUM_SQDIFF_16x1_SSE2 esi,edi,ebx WELS_SAD_SUM_SQSUM_SQDIFF_16x1_SSE2 esi,edi,ebx WELS_SAD_SUM_SQSUM_SQDIFF_16x1_SSE2 esi,edi,ebx WELS_SAD_SUM_SQSUM_SQDIFF_16x1_SSE2 esi,edi,ebx movdqa xmm1, xmm7 movd [edx+8], xmm7 psrldq xmm7, 8 paddd xmm1, xmm7 movd [edx+12], xmm7 movd ebp, xmm1 add [tmp_sadframe], ebp mov ebp, [psum16x16] movdqa xmm1, xmm6 psrldq xmm1, 8 paddd xmm6, xmm1 movd [ebp], xmm6 add dword [psum16x16], 4 mov ebp, [psqsum16x16] pshufd xmm6, xmm5, 14 ;00001110 paddd xmm6, xmm5 pshufd xmm5, xmm6, 1 ;00000001 paddd xmm5, xmm6 movd [ebp], xmm5 add dword [psqsum16x16], 4 mov ebp, [psqdiff16x16] pshufd xmm5, xmm4, 14 ; 00001110 paddd xmm5, xmm4 pshufd xmm4, xmm5, 1 ; 00000001 paddd xmm4, xmm5 movd [ebp], xmm4 add dword [psqdiff16x16], 4 add edx, 16 sub esi, eax sub edi, eax add esi, 16 add edi, 16 dec ecx jnz sqdiff_width_loop mov esi, [tmp_esi] mov edi, [tmp_edi] add esi, eax add edi, eax dec dword [iPicHeight] jnz sqdiff_height_loop mov ebx, [tmp_sadframe] mov eax, [psadframe] mov [eax], ebx add esp, localsize pop ebx pop edi pop esi pop ebp %undef cur_data %undef ref_data %undef iPicWidth %undef iPicHeight %undef iPicStride %undef psadframe %undef psad8x8 %undef psum16x16 %undef psqsum16x16 %undef psqdiff16x16 %undef tmp_esi %undef tmp_edi %undef tmp_sadframe %undef pushsize %undef localsize ret %else ;************************************************************************************************************* ;void VAACalcSadSsd_sse2(const uint8_t *cur_data, const uint8_t *ref_data, int32_t iPicWidth, int32_t iPicHeight, ; int32_t iPicStride,int32_t *psadframe, int32_t *psad8x8, int32_t *psum16x16, int32_t *psqsum16x16, int32_t *psqdiff16x16) ;************************************************************************************************************* WELS_EXTERN VAACalcSadSsd_sse2 %define localsize 12 %define cur_data arg1;r0 %define ref_data arg2;r1 %define iPicWidth arg3;r2 %define iPicHeight arg4;r3 %define iPicStride arg5; %define psadframe arg6; %define psad8x8 arg7; %define psum16x16 arg8; %define psqsum16x16 arg9; %define psqdiff16x16 arg10 push r12 push r13 push r14 push r15 %assign push_num 4 PUSH_XMM 10 %ifdef WIN64 mov r4,arg5 %endif mov r14,arg7 SIGN_EXTENSION r2,r2d SIGN_EXTENSION r3,r3d SIGN_EXTENSION r4,r4d mov r13,r4 shr r2,4 ; iPicWidth/16 shr r3,4 ; iPicHeight/16 shl r13,4 ; iPicStride*16 pxor xmm0, xmm0 pxor xmm8, xmm8 ;framesad pxor xmm9, xmm9 sqdiff_height_loop: ;mov ecx, dword [iPicWidth] ;mov r14,r2 push r2 %assign push_num push_num +1 mov r10, r0 mov r11, r1 sqdiff_width_loop: pxor xmm7, xmm7 ; hiQuad_loQuad pSad8x8 pxor xmm6, xmm6 ; pSum16x16 pxor xmm5, xmm5 ; sqsum_16x16 four dword pxor xmm4, xmm4 ; sqdiff_16x16 four Dword WELS_SAD_SUM_SQSUM_SQDIFF_16x1_SSE2 r0,r1,r4 WELS_SAD_SUM_SQSUM_SQDIFF_16x1_SSE2 r0,r1,r4 WELS_SAD_SUM_SQSUM_SQDIFF_16x1_SSE2 r0,r1,r4 WELS_SAD_SUM_SQSUM_SQDIFF_16x1_SSE2 r0,r1,r4 WELS_SAD_SUM_SQSUM_SQDIFF_16x1_SSE2 r0,r1,r4 WELS_SAD_SUM_SQSUM_SQDIFF_16x1_SSE2 r0,r1,r4 WELS_SAD_SUM_SQSUM_SQDIFF_16x1_SSE2 r0,r1,r4 WELS_SAD_SUM_SQSUM_SQDIFF_16x1_SSE2 r0,r1,r4 movdqa xmm1, xmm7 movd [r14], xmm7 psrldq xmm7, 8 paddd xmm1, xmm7 movd [r14+4], xmm7 movd r15d, xmm1 movd xmm9, r15d paddd xmm8,xmm9 pxor xmm7, xmm7 WELS_SAD_SUM_SQSUM_SQDIFF_16x1_SSE2 r0,r1,r4 WELS_SAD_SUM_SQSUM_SQDIFF_16x1_SSE2 r0,r1,r4 WELS_SAD_SUM_SQSUM_SQDIFF_16x1_SSE2 r0,r1,r4 WELS_SAD_SUM_SQSUM_SQDIFF_16x1_SSE2 r0,r1,r4 WELS_SAD_SUM_SQSUM_SQDIFF_16x1_SSE2 r0,r1,r4 WELS_SAD_SUM_SQSUM_SQDIFF_16x1_SSE2 r0,r1,r4 WELS_SAD_SUM_SQSUM_SQDIFF_16x1_SSE2 r0,r1,r4 WELS_SAD_SUM_SQSUM_SQDIFF_16x1_SSE2 r0,r1,r4 movdqa xmm1, xmm7 movd [r14+8], xmm7 psrldq xmm7, 8 paddd xmm1, xmm7 movd [r14+12], xmm7 movd r15d, xmm1 movd xmm9, r15d paddd xmm8,xmm9 mov r15, psum16x16 movdqa xmm1, xmm6 psrldq xmm1, 8 paddd xmm6, xmm1 movd [r15], xmm6 add dword psum16x16, 4 mov r15, psqsum16x16 pshufd xmm6, xmm5, 14 ;00001110 paddd xmm6, xmm5 pshufd xmm5, xmm6, 1 ;00000001 paddd xmm5, xmm6 movd [r15], xmm5 add dword psqsum16x16, 4 mov r15, psqdiff16x16 pshufd xmm5, xmm4, 14 ; 00001110 paddd xmm5, xmm4 pshufd xmm4, xmm5, 1 ; 00000001 paddd xmm4, xmm5 movd [r15], xmm4 add dword psqdiff16x16, 4 add r14,16 sub r0, r13 sub r1, r13 add r0, 16 add r1, 16 dec r2 jnz sqdiff_width_loop pop r2 %assign push_num push_num -1 mov r0, r10 mov r1, r11 add r0, r13 add r1, r13 dec r3 jnz sqdiff_height_loop mov r13, psadframe movd [r13], xmm8 POP_XMM pop r15 pop r14 pop r13 pop r12 %assign push_num 0 %undef cur_data %undef ref_data %undef iPicWidth %undef iPicHeight %undef iPicStride %undef psadframe %undef psad8x8 %undef psum16x16 %undef psqsum16x16 %undef psqdiff16x16 %undef tmp_esi %undef tmp_edi %undef tmp_sadframe %undef pushsize %undef localsize ret %endif %ifdef X86_32 ;************************************************************************************************************* ;void VAACalcSadBgd_sse2(const uint8_t *cur_data, const uint8_t *ref_data, int32_t iPicWidth, int32_t iPicHeight, ; int32_t iPicStride, int32_t *psadframe, int32_t *psad8x8, int32_t *p_sd8x8, uint8_t *p_mad8x8) ;************************************************************************************************************* WELS_EXTERN VAACalcSadBgd_sse2 %define localsize 12 %define cur_data esp + pushsize + localsize + 4 %define ref_data esp + pushsize + localsize + 8 %define iPicWidth esp + pushsize + localsize + 12 %define iPicHeight esp + pushsize + localsize + 16 %define iPicStride esp + pushsize + localsize + 20 %define psadframe esp + pushsize + localsize + 24 %define psad8x8 esp + pushsize + localsize + 28 %define p_sd8x8 esp + pushsize + localsize + 32 %define p_mad8x8 esp + pushsize + localsize + 36 %define tmp_esi esp + 0 %define tmp_edi esp + 4 %define tmp_ecx esp + 8 %define pushsize 16 push ebp push esi push edi push ebx sub esp, localsize mov esi, [cur_data] mov edi, [ref_data] mov ebx, [iPicStride] mov eax, ebx shr dword [iPicWidth], 4 ; iPicWidth/16 shr dword [iPicHeight], 4 ; iPicHeight/16 shl eax, 4 ; iPicStride*16 xor ebp, ebp pxor xmm0, xmm0 bgd_height_loop: mov ecx, dword [iPicWidth] mov [tmp_esi], esi mov [tmp_edi], edi bgd_width_loop: pxor xmm7, xmm7 ; pSad8x8 pxor xmm6, xmm6 ; sum_cur_8x8 pxor xmm5, xmm5 ; sum_ref_8x8 pxor xmm4, xmm4 ; pMad8x8 WELS_SAD_SD_MAD_16x1_SSE2 xmm7, xmm6, xmm5, xmm4 ,esi ,edi, ebx WELS_SAD_SD_MAD_16x1_SSE2 xmm7, xmm6, xmm5, xmm4 ,esi ,edi, ebx WELS_SAD_SD_MAD_16x1_SSE2 xmm7, xmm6, xmm5, xmm4 ,esi ,edi, ebx WELS_SAD_SD_MAD_16x1_SSE2 xmm7, xmm6, xmm5, xmm4 ,esi ,edi, ebx WELS_SAD_SD_MAD_16x1_SSE2 xmm7, xmm6, xmm5, xmm4 ,esi ,edi, ebx WELS_SAD_SD_MAD_16x1_SSE2 xmm7, xmm6, xmm5, xmm4 ,esi ,edi, ebx WELS_SAD_SD_MAD_16x1_SSE2 xmm7, xmm6, xmm5, xmm4 ,esi ,edi, ebx WELS_SAD_SD_MAD_16x1_SSE2 xmm7, xmm6, xmm5, xmm4 ,esi ,edi, ebx mov edx, [p_mad8x8] WELS_MAX_REG_SSE2 xmm4 ;movdqa xmm1, xmm4 ;punpcklbw xmm1, xmm0 ;punpcklwd xmm1, xmm0 ;movd [edx], xmm1 ;punpckhbw xmm4, xmm0 ;punpcklwd xmm4, xmm0 ;movd [edx+4], xmm4 ;add edx, 8 ;mov [p_mad8x8], edx mov [tmp_ecx], ecx movhlps xmm1, xmm4 movd ecx, xmm4 mov [edx], cl movd ecx, xmm1 mov [edx+1],cl add edx, 2 mov [p_mad8x8], edx pslldq xmm7, 4 pslldq xmm6, 4 pslldq xmm5, 4 pxor xmm4, xmm4 ; pMad8x8 WELS_SAD_SD_MAD_16x1_SSE2 xmm7, xmm6, xmm5, xmm4 ,esi ,edi, ebx WELS_SAD_SD_MAD_16x1_SSE2 xmm7, xmm6, xmm5, xmm4 ,esi ,edi, ebx WELS_SAD_SD_MAD_16x1_SSE2 xmm7, xmm6, xmm5, xmm4 ,esi ,edi, ebx WELS_SAD_SD_MAD_16x1_SSE2 xmm7, xmm6, xmm5, xmm4 ,esi ,edi, ebx WELS_SAD_SD_MAD_16x1_SSE2 xmm7, xmm6, xmm5, xmm4 ,esi ,edi, ebx WELS_SAD_SD_MAD_16x1_SSE2 xmm7, xmm6, xmm5, xmm4 ,esi ,edi, ebx WELS_SAD_SD_MAD_16x1_SSE2 xmm7, xmm6, xmm5, xmm4 ,esi ,edi, ebx WELS_SAD_SD_MAD_16x1_SSE2 xmm7, xmm6, xmm5, xmm4 ,esi ,edi, ebx mov edx, [p_mad8x8] WELS_MAX_REG_SSE2 xmm4 ;movdqa xmm1, xmm4 ;punpcklbw xmm1, xmm0 ;punpcklwd xmm1, xmm0 ;movd [edx], xmm1 ;punpckhbw xmm4, xmm0 ;punpcklwd xmm4, xmm0 ;movd [edx+4], xmm4 ;add edx, 8 ;mov [p_mad8x8], edx movhlps xmm1, xmm4 movd ecx, xmm4 mov [edx], cl movd ecx, xmm1 mov [edx+1],cl add edx, 2 mov [p_mad8x8], edx ; data in xmm7, xmm6, xmm5: D1 D3 D0 D2 mov edx, [psad8x8] pshufd xmm1, xmm7, 10001101b ; D3 D2 D1 D0 movdqa [edx], xmm1 add edx, 16 mov [psad8x8], edx ; sad8x8 paddd xmm1, xmm7 ; D1+3 D3+2 D0+1 D2+0 pshufd xmm2, xmm1, 00000011b paddd xmm1, xmm2 movd edx, xmm1 add ebp, edx ; sad frame mov edx, [p_sd8x8] psubd xmm6, xmm5 pshufd xmm1, xmm6, 10001101b movdqa [edx], xmm1 add edx, 16 mov [p_sd8x8], edx add edx, 16 sub esi, eax sub edi, eax add esi, 16 add edi, 16 mov ecx, [tmp_ecx] dec ecx jnz bgd_width_loop mov esi, [tmp_esi] mov edi, [tmp_edi] add esi, eax add edi, eax dec dword [iPicHeight] jnz bgd_height_loop mov edx, [psadframe] mov [edx], ebp add esp, localsize pop ebx pop edi pop esi pop ebp %undef cur_data %undef ref_data %undef iPicWidth %undef iPicHeight %undef iPicStride %undef psadframe %undef psad8x8 %undef p_sd8x8 %undef p_mad8x8 %undef tmp_esi %undef tmp_edi %undef pushsize %undef localsize ret ;************************************************************************************************************* ;void VAACalcSadSsdBgd_sse2(const uint8_t *cur_data, const uint8_t *ref_data, int32_t iPicWidth, int32_t iPicHeight, ; int32_t iPicStride, int32_t *psadframe, int32_t *psad8x8, int32_t *psum16x16, int32_t *psqsum16x16, ; int32_t *psqdiff16x16, int32_t *p_sd8x8, uint8_t *p_mad8x8) ;************************************************************************************************************* WELS_EXTERN VAACalcSadSsdBgd_sse2 %define localsize 16 %define cur_data esp + pushsize + localsize + 4 %define ref_data esp + pushsize + localsize + 8 %define iPicWidth esp + pushsize + localsize + 12 %define iPicHeight esp + pushsize + localsize + 16 %define iPicStride esp + pushsize + localsize + 20 %define psadframe esp + pushsize + localsize + 24 %define psad8x8 esp + pushsize + localsize + 28 %define psum16x16 esp + pushsize + localsize + 32 %define psqsum16x16 esp + pushsize + localsize + 36 %define psqdiff16x16 esp + pushsize + localsize + 40 %define p_sd8x8 esp + pushsize + localsize + 44 %define p_mad8x8 esp + pushsize + localsize + 48 %define tmp_esi esp + 0 %define tmp_edi esp + 4 %define tmp_sadframe esp + 8 %define tmp_ecx esp + 12 %define pushsize 16 push ebp push esi push edi push ebx sub esp, localsize mov esi, [cur_data] mov edi, [ref_data] mov ebx, [iPicStride] mov eax, ebx shr dword [iPicWidth], 4 ; iPicWidth/16 shr dword [iPicHeight], 4 ; iPicHeight/16 shl eax, 4 ; iPicStride*16 pxor xmm0, xmm0 movd [tmp_sadframe], xmm0 sqdiff_bgd_height_loop: mov ecx, dword [iPicWidth] mov [tmp_esi], esi mov [tmp_edi], edi sqdiff_bgd_width_loop: pxor xmm7, xmm7 ; pSad8x8 interleaves sqsum16x16: sqsum1 sad1 sqsum0 sad0 pxor xmm6, xmm6 ; sum_8x8 interleaves cur and pRef in Dword, Sref1 Scur1 Sref0 Scur0 pxor xmm5, xmm5 ; pMad8x8 pxor xmm4, xmm4 ; sqdiff_16x16 four Dword WELS_SAD_BGD_SQDIFF_16x1_SSE2 xmm7, xmm6, xmm5, xmm4, esi , edi , ebx WELS_SAD_BGD_SQDIFF_16x1_SSE2 xmm7, xmm6, xmm5, xmm4, esi , edi , ebx WELS_SAD_BGD_SQDIFF_16x1_SSE2 xmm7, xmm6, xmm5, xmm4, esi , edi , ebx WELS_SAD_BGD_SQDIFF_16x1_SSE2 xmm7, xmm6, xmm5, xmm4, esi , edi , ebx WELS_SAD_BGD_SQDIFF_16x1_SSE2 xmm7, xmm6, xmm5, xmm4, esi , edi , ebx WELS_SAD_BGD_SQDIFF_16x1_SSE2 xmm7, xmm6, xmm5, xmm4, esi , edi , ebx WELS_SAD_BGD_SQDIFF_16x1_SSE2 xmm7, xmm6, xmm5, xmm4, esi , edi , ebx WELS_SAD_BGD_SQDIFF_16x1_SSE2 xmm7, xmm6, xmm5, xmm4, esi , edi , ebx mov edx, [psad8x8] movdqa xmm2, xmm7 pshufd xmm1, xmm2, 00001110b movd [edx], xmm2 movd [edx+4], xmm1 add edx, 8 mov [psad8x8], edx ; sad8x8 paddd xmm1, xmm2 movd edx, xmm1 add [tmp_sadframe], edx ; iFrameSad mov edx, [psum16x16] movdqa xmm1, xmm6 pshufd xmm2, xmm1, 00001110b paddd xmm1, xmm2 movd [edx], xmm1 ; sum mov edx, [p_sd8x8] pshufd xmm1, xmm6, 11110101b ; Sref1 Sref1 Sref0 Sref0 psubd xmm6, xmm1 ; 00 diff1 00 diff0 pshufd xmm1, xmm6, 00001000b ; xx xx diff1 diff0 movq [edx], xmm1 add edx, 8 mov [p_sd8x8], edx mov edx, [p_mad8x8] WELS_MAX_REG_SSE2 xmm5 ;movdqa xmm1, xmm5 ;punpcklbw xmm1, xmm0 ;punpcklwd xmm1, xmm0 ;movd [edx], xmm1 ;punpckhbw xmm5, xmm0 ;punpcklwd xmm5, xmm0 ;movd [edx+4], xmm5 ;add edx, 8 ;mov [p_mad8x8], edx mov [tmp_ecx], ecx movhlps xmm1, xmm5 movd ecx, xmm5 mov [edx], cl movd ecx, xmm1 mov [edx+1],cl add edx, 2 mov [p_mad8x8], edx psrlq xmm7, 32 psllq xmm7, 32 ; clear sad pxor xmm6, xmm6 ; sum_8x8 interleaves cur and pRef in Dword, Sref1 Scur1 Sref0 Scur0 pxor xmm5, xmm5 ; pMad8x8 WELS_SAD_BGD_SQDIFF_16x1_SSE2 xmm7, xmm6, xmm5, xmm4, esi , edi , ebx WELS_SAD_BGD_SQDIFF_16x1_SSE2 xmm7, xmm6, xmm5, xmm4, esi , edi , ebx WELS_SAD_BGD_SQDIFF_16x1_SSE2 xmm7, xmm6, xmm5, xmm4, esi , edi , ebx WELS_SAD_BGD_SQDIFF_16x1_SSE2 xmm7, xmm6, xmm5, xmm4, esi , edi , ebx WELS_SAD_BGD_SQDIFF_16x1_SSE2 xmm7, xmm6, xmm5, xmm4, esi , edi , ebx WELS_SAD_BGD_SQDIFF_16x1_SSE2 xmm7, xmm6, xmm5, xmm4, esi , edi , ebx WELS_SAD_BGD_SQDIFF_16x1_SSE2 xmm7, xmm6, xmm5, xmm4, esi , edi , ebx WELS_SAD_BGD_SQDIFF_16x1_SSE2 xmm7, xmm6, xmm5, xmm4, esi , edi , ebx mov edx, [psad8x8] movdqa xmm2, xmm7 pshufd xmm1, xmm2, 00001110b movd [edx], xmm2 movd [edx+4], xmm1 add edx, 8 mov [psad8x8], edx ; sad8x8 paddd xmm1, xmm2 movd edx, xmm1 add [tmp_sadframe], edx ; iFrameSad mov edx, [psum16x16] movdqa xmm1, xmm6 pshufd xmm2, xmm1, 00001110b paddd xmm1, xmm2 movd ebp, xmm1 ; sum add [edx], ebp add edx, 4 mov [psum16x16], edx mov edx, [psqsum16x16] psrlq xmm7, 32 pshufd xmm2, xmm7, 00001110b paddd xmm2, xmm7 movd [edx], xmm2 ; sqsum add edx, 4 mov [psqsum16x16], edx mov edx, [p_sd8x8] pshufd xmm1, xmm6, 11110101b ; Sref1 Sref1 Sref0 Sref0 psubd xmm6, xmm1 ; 00 diff1 00 diff0 pshufd xmm1, xmm6, 00001000b ; xx xx diff1 diff0 movq [edx], xmm1 add edx, 8 mov [p_sd8x8], edx mov edx, [p_mad8x8] WELS_MAX_REG_SSE2 xmm5 ;movdqa xmm1, xmm5 ;punpcklbw xmm1, xmm0 ;punpcklwd xmm1, xmm0 ;movd [edx], xmm1 ;punpckhbw xmm5, xmm0 ;punpcklwd xmm5, xmm0 ;movd [edx+4], xmm5 ;add edx, 8 ;mov [p_mad8x8], edx movhlps xmm1, xmm5 movd ecx, xmm5 mov [edx], cl movd ecx, xmm1 mov [edx+1],cl add edx, 2 mov [p_mad8x8], edx mov edx, [psqdiff16x16] pshufd xmm1, xmm4, 00001110b paddd xmm4, xmm1 pshufd xmm1, xmm4, 00000001b paddd xmm4, xmm1 movd [edx], xmm4 add edx, 4 mov [psqdiff16x16], edx add edx, 16 sub esi, eax sub edi, eax add esi, 16 add edi, 16 mov ecx, [tmp_ecx] dec ecx jnz sqdiff_bgd_width_loop mov esi, [tmp_esi] mov edi, [tmp_edi] add esi, eax add edi, eax dec dword [iPicHeight] jnz sqdiff_bgd_height_loop mov edx, [psadframe] mov ebp, [tmp_sadframe] mov [edx], ebp add esp, localsize pop ebx pop edi pop esi pop ebp %undef cur_data %undef ref_data %undef iPicWidth %undef iPicHeight %undef iPicStride %undef psadframe %undef psad8x8 %undef psum16x16 %undef psqsum16x16 %undef psqdiff16x16 %undef p_sd8x8 %undef p_mad8x8 %undef tmp_esi %undef tmp_edi %undef pushsize %undef localsize ret %else ;************************************************************************************************************* ;void VAACalcSadBgd_sse2(const uint8_t *cur_data, const uint8_t *ref_data, int32_t iPicWidth, int32_t iPicHeight, ; int32_t iPicStride, int32_t *psadframe, int32_t *psad8x8, int32_t *p_sd8x8, uint8_t *p_mad8x8) ;************************************************************************************************************* WELS_EXTERN VAACalcSadBgd_sse2 %define cur_data arg1; %define ref_data arg2; %define iPicWidth arg3; %define iPicHeight arg4; %define iPicStride arg5; %define psadframe arg6; %define psad8x8 arg7; %define p_sd8x8 arg8; %define p_mad8x8 arg9; push r12 push r13 push r14 push r15 %assign push_num 4 PUSH_XMM 10 %ifdef WIN64 mov r4,arg5 ; mov r5,arg6 %endif mov r14,arg7 SIGN_EXTENSION r2,r2d SIGN_EXTENSION r3,r3d SIGN_EXTENSION r4,r4d mov r13,r4 mov r15,r0 shr r2,4 shr r3,4 shl r13,4 pxor xmm0, xmm0 pxor xmm8, xmm8 pxor xmm9, xmm9 bgd_height_loop: ;mov ecx, dword [iPicWidth] push r2 %assign push_num push_num+1 mov r10, r15 mov r11, r1 bgd_width_loop: pxor xmm7, xmm7 ; pSad8x8 pxor xmm6, xmm6 ; sum_cur_8x8 pxor xmm5, xmm5 ; sum_ref_8x8 pxor xmm4, xmm4 ; pMad8x8 WELS_SAD_SD_MAD_16x1_SSE2 xmm7, xmm6, xmm5, xmm4 ,r15 ,r1, r4 WELS_SAD_SD_MAD_16x1_SSE2 xmm7, xmm6, xmm5, xmm4 ,r15 ,r1, r4 WELS_SAD_SD_MAD_16x1_SSE2 xmm7, xmm6, xmm5, xmm4 ,r15 ,r1, r4 WELS_SAD_SD_MAD_16x1_SSE2 xmm7, xmm6, xmm5, xmm4 ,r15 ,r1, r4 WELS_SAD_SD_MAD_16x1_SSE2 xmm7, xmm6, xmm5, xmm4 ,r15 ,r1, r4 WELS_SAD_SD_MAD_16x1_SSE2 xmm7, xmm6, xmm5, xmm4 ,r15 ,r1, r4 WELS_SAD_SD_MAD_16x1_SSE2 xmm7, xmm6, xmm5, xmm4 ,r15 ,r1, r4 WELS_SAD_SD_MAD_16x1_SSE2 xmm7, xmm6, xmm5, xmm4 ,r15 ,r1, r4 mov r14, p_mad8x8 WELS_MAX_REG_SSE2 xmm4 ;mov [tmp_ecx], ecx movhlps xmm1, xmm4 movd r0d, xmm4 mov [r14], r0b movd r0d, xmm1 mov [r14+1],r0b add r14, 2 ;mov p_mad8x8, r14 pslldq xmm7, 4 pslldq xmm6, 4 pslldq xmm5, 4 pxor xmm4, xmm4 ; pMad8x8 WELS_SAD_SD_MAD_16x1_SSE2 xmm7, xmm6, xmm5, xmm4 ,r15 ,r1, r4 WELS_SAD_SD_MAD_16x1_SSE2 xmm7, xmm6, xmm5, xmm4 ,r15 ,r1, r4 WELS_SAD_SD_MAD_16x1_SSE2 xmm7, xmm6, xmm5, xmm4 ,r15 ,r1, r4 WELS_SAD_SD_MAD_16x1_SSE2 xmm7, xmm6, xmm5, xmm4 ,r15 ,r1, r4 WELS_SAD_SD_MAD_16x1_SSE2 xmm7, xmm6, xmm5, xmm4 ,r15 ,r1, r4 WELS_SAD_SD_MAD_16x1_SSE2 xmm7, xmm6, xmm5, xmm4 ,r15 ,r1, r4 WELS_SAD_SD_MAD_16x1_SSE2 xmm7, xmm6, xmm5, xmm4 ,r15 ,r1, r4 WELS_SAD_SD_MAD_16x1_SSE2 xmm7, xmm6, xmm5, xmm4 ,r15 ,r1, r4 ;mov r14, [p_mad8x8] WELS_MAX_REG_SSE2 xmm4 movhlps xmm1, xmm4 movd r0d, xmm4 mov [r14], r0b movd r0d, xmm1 mov [r14+1],r0b add r14, 2 mov p_mad8x8, r14 ; data in xmm7, xmm6, xmm5: D1 D3 D0 D2 mov r14, psad8x8 pshufd xmm1, xmm7, 10001101b ; D3 D2 D1 D0 movdqa [r14], xmm1 add r14, 16 mov psad8x8, r14 ; sad8x8 paddd xmm1, xmm7 ; D1+3 D3+2 D0+1 D2+0 pshufd xmm2, xmm1, 00000011b paddd xmm1, xmm2 movd r14d, xmm1 movd xmm9, r14d paddd xmm8, xmm9 ; sad frame mov r14, p_sd8x8 psubd xmm6, xmm5 pshufd xmm1, xmm6, 10001101b movdqa [r14], xmm1 add r14, 16 mov p_sd8x8, r14 ;add edx, 16 sub r15, r13 sub r1, r13 add r15, 16 add r1, 16 dec r2 jnz bgd_width_loop pop r2 %assign push_num push_num-1 mov r15, r10 mov r1, r11 add r15, r13 add r1, r13 dec r3 jnz bgd_height_loop mov r13, psadframe movd [r13], xmm8 POP_XMM pop r15 pop r14 pop r13 pop r12 %assign push_num 0 %undef cur_data %undef ref_data %undef iPicWidth %undef iPicHeight %undef iPicStride %undef psadframe %undef psad8x8 %undef p_sd8x8 %undef p_mad8x8 %undef tmp_esi %undef tmp_edi %undef pushsize %undef localsize ret ;************************************************************************************************************* ;void VAACalcSadSsdBgd_sse2(const uint8_t *cur_data, const uint8_t *ref_data, int32_t iPicWidth, int32_t iPicHeight, ; int32_t iPicStride, int32_t *psadframe, int32_t *psad8x8, int32_t *psum16x16, int32_t *psqsum16x16, ; int32_t *psqdiff16x16, int32_t *p_sd8x8, uint8_t *p_mad8x8) ;************************************************************************************************************* WELS_EXTERN VAACalcSadSsdBgd_sse2 %define cur_data arg1; %define ref_data arg2; %define iPicWidth arg3; %define iPicHeight arg4; %define iPicStride arg5; %define psadframe arg6; %define psad8x8 arg7; %define psum16x16 arg8; %define psqsum16x16 arg9; %define psqdiff16x16 arg10; %define p_sd8x8 arg11 %define p_mad8x8 arg12 push r12 push r13 push r14 push r15 %assign push_num 4 PUSH_XMM 10 %ifdef WIN64 mov r4,arg5 ;mov r5,arg6 %endif SIGN_EXTENSION r2,r2d SIGN_EXTENSION r3,r3d SIGN_EXTENSION r4,r4d mov r13,r4 shr r2, 4 ; iPicWidth/16 shr r3, 4 ; iPicHeight/16 shl r13, 4 ; iPicStride*16 pxor xmm0, xmm0 pxor xmm8, xmm8 pxor xmm9, xmm9 sqdiff_bgd_height_loop: mov r10, r0 mov r11, r1 push r2 %assign push_num push_num+1 sqdiff_bgd_width_loop: pxor xmm7, xmm7 ; pSad8x8 interleaves sqsum16x16: sqsum1 sad1 sqsum0 sad0 pxor xmm6, xmm6 ; sum_8x8 interleaves cur and pRef in Dword, Sref1 Scur1 Sref0 Scur0 pxor xmm5, xmm5 ; pMad8x8 pxor xmm4, xmm4 ; sqdiff_16x16 four Dword WELS_SAD_BGD_SQDIFF_16x1_SSE2 xmm7, xmm6, xmm5, xmm4, r0 , r1 , r4 WELS_SAD_BGD_SQDIFF_16x1_SSE2 xmm7, xmm6, xmm5, xmm4, r0 , r1 , r4 WELS_SAD_BGD_SQDIFF_16x1_SSE2 xmm7, xmm6, xmm5, xmm4, r0 , r1 , r4 WELS_SAD_BGD_SQDIFF_16x1_SSE2 xmm7, xmm6, xmm5, xmm4, r0 , r1 , r4 WELS_SAD_BGD_SQDIFF_16x1_SSE2 xmm7, xmm6, xmm5, xmm4, r0 , r1 , r4 WELS_SAD_BGD_SQDIFF_16x1_SSE2 xmm7, xmm6, xmm5, xmm4, r0 , r1 , r4 WELS_SAD_BGD_SQDIFF_16x1_SSE2 xmm7, xmm6, xmm5, xmm4, r0 , r1 , r4 WELS_SAD_BGD_SQDIFF_16x1_SSE2 xmm7, xmm6, xmm5, xmm4, r0 , r1 , r4 mov r14, psad8x8 movdqa xmm2, xmm7 pshufd xmm1, xmm2, 00001110b movd [r14], xmm2 movd [r14+4], xmm1 add r14, 8 mov psad8x8, r14 ; sad8x8 paddd xmm1, xmm2 movd r14d, xmm1 movd xmm9,r14d paddd xmm8, xmm9 ; iFrameSad mov r14, psum16x16 movdqa xmm1, xmm6 pshufd xmm2, xmm1, 00001110b paddd xmm1, xmm2 movd [r14], xmm1 ; sum mov r14, p_sd8x8 pshufd xmm1, xmm6, 11110101b ; Sref1 Sref1 Sref0 Sref0 psubd xmm6, xmm1 ; 00 diff1 00 diff0 pshufd xmm1, xmm6, 00001000b ; xx xx diff1 diff0 movq [r14], xmm1 add r14, 8 mov p_sd8x8, r14 mov r14, p_mad8x8 WELS_MAX_REG_SSE2 xmm5 movhlps xmm1, xmm5 push r0 movd r0d, xmm5 mov [r14], r0b movd r0d, xmm1 mov [r14+1],r0b pop r0 add r14, 2 mov p_mad8x8, r14 psrlq xmm7, 32 psllq xmm7, 32 ; clear sad pxor xmm6, xmm6 ; sum_8x8 interleaves cur and pRef in Dword, Sref1 Scur1 Sref0 Scur0 pxor xmm5, xmm5 ; pMad8x8 WELS_SAD_BGD_SQDIFF_16x1_SSE2 xmm7, xmm6, xmm5, xmm4, r0 , r1 , r4 WELS_SAD_BGD_SQDIFF_16x1_SSE2 xmm7, xmm6, xmm5, xmm4, r0 , r1 , r4 WELS_SAD_BGD_SQDIFF_16x1_SSE2 xmm7, xmm6, xmm5, xmm4, r0 , r1 , r4 WELS_SAD_BGD_SQDIFF_16x1_SSE2 xmm7, xmm6, xmm5, xmm4, r0 , r1 , r4 WELS_SAD_BGD_SQDIFF_16x1_SSE2 xmm7, xmm6, xmm5, xmm4, r0 , r1 , r4 WELS_SAD_BGD_SQDIFF_16x1_SSE2 xmm7, xmm6, xmm5, xmm4, r0 , r1 , r4 WELS_SAD_BGD_SQDIFF_16x1_SSE2 xmm7, xmm6, xmm5, xmm4, r0 , r1 , r4 WELS_SAD_BGD_SQDIFF_16x1_SSE2 xmm7, xmm6, xmm5, xmm4, r0 , r1 , r4 mov r14, psad8x8 movdqa xmm2, xmm7 pshufd xmm1, xmm2, 00001110b movd [r14], xmm2 movd [r14+4], xmm1 add r14, 8 mov psad8x8, r14 ; sad8x8 paddd xmm1, xmm2 movd r14d, xmm1 movd xmm9, r14d paddd xmm8, xmm9 ; iFrameSad mov r14, psum16x16 movdqa xmm1, xmm6 pshufd xmm2, xmm1, 00001110b paddd xmm1, xmm2 movd r15d, xmm1 ; sum add [r14], r15d add r14, 4 mov psum16x16, r14 mov r14, psqsum16x16 psrlq xmm7, 32 pshufd xmm2, xmm7, 00001110b paddd xmm2, xmm7 movd [r14], xmm2 ; sqsum add r14, 4 mov psqsum16x16, r14 mov r14, p_sd8x8 pshufd xmm1, xmm6, 11110101b ; Sref1 Sref1 Sref0 Sref0 psubd xmm6, xmm1 ; 00 diff1 00 diff0 pshufd xmm1, xmm6, 00001000b ; xx xx diff1 diff0 movq [r14], xmm1 add r14, 8 mov p_sd8x8, r14 mov r14, p_mad8x8 WELS_MAX_REG_SSE2 xmm5 movhlps xmm1, xmm5 push r0 movd r0d, xmm5 mov [r14], r0b movd r0d, xmm1 mov [r14+1],r0b pop r0 add r14, 2 mov p_mad8x8, r14 mov r14, psqdiff16x16 pshufd xmm1, xmm4, 00001110b paddd xmm4, xmm1 pshufd xmm1, xmm4, 00000001b paddd xmm4, xmm1 movd [r14], xmm4 add r14, 4 mov psqdiff16x16, r14 add r14, 16 sub r0, r13 sub r1, r13 add r0, 16 add r1, 16 dec r2 jnz sqdiff_bgd_width_loop pop r2 %assign push_num push_num-1 mov r0, r10 mov r1, r11 add r0, r13 add r1, r13 dec r3 jnz sqdiff_bgd_height_loop mov r14, psadframe movd [r14], xmm8 POP_XMM pop r15 pop r14 pop r13 pop r12 %assign push_num 0 %undef cur_data %undef ref_data %undef iPicWidth %undef iPicHeight %undef iPicStride %undef psadframe %undef psad8x8 %undef psum16x16 %undef psqsum16x16 %undef psqdiff16x16 %undef p_sd8x8 %undef p_mad8x8 %undef tmp_esi %undef tmp_edi %undef pushsize %undef localsize ret %endif %ifdef X86_32 %define ptrword dword %else %define ptrword qword %endif %define xmm_width 16 %define ymm_width 32 %macro PUSHM 1-* %rep %0 push %1 %rotate 1 %endrep %assign push_num push_num + %0 %endmacro %macro POPM 1-* %rep %0 %rotate -1 pop %1 %endrep %assign push_num push_num - %0 %endmacro %ifdef X86_32 %define stack_alloc_min 4 %else %define stack_alloc_min 8 %endif ; Allocate aligned stack space. ; address_out=%1 size=%2 alignment=%3 %macro STACK_ALLOC 3 %if (%3) & ((%3) - 1) %error non-power-of-2 alignment requested. %endif %if (%3) > 0 %assign stack_alloc_align ((%3) + stack_alloc_min - 1) / stack_alloc_min %else %assign stack_alloc_align 1 %endif %assign stack_alloc_num ((%2) + stack_alloc_min - 1) / stack_alloc_min + stack_alloc_align - 1 %assign push_num push_num + stack_alloc_num sub r7, stack_alloc_min * stack_alloc_num %if stack_alloc_align == 1 mov %1, r7 %else lea %1, [r7 + stack_alloc_min * (stack_alloc_align - 1)] and %1, -(stack_alloc_min * stack_alloc_align) %endif %endmacro ; Deallocate stack space allocated with STACK_ALLOC. %macro STACK_DEALLOC 0 add r7, stack_alloc_min * stack_alloc_num %assign push_num push_num - stack_alloc_num %endmacro %ifdef HAVE_AVX2 ; Max unsigned byte per quadword ; out=%1 in=%2 tmp=%3 %macro AVX2_Maxubq 3 vpsrlq %3, %2, 32 vpmaxub %1, %2, %3 vpsrlq %3, %1, 16 vpmaxub %1, %1, %3 vpsrlq %3, %1, 8 vpmaxub %1, %1, %3 %endmacro ; Max unsigned byte per quadword. 2 register input. ; Results interleaved as least significant byte of even/odd doublewords. ; out=%1 in_a=%2 in_b=%3 tmp=%4 %macro AVX2_Maxubq2 4 vpblendd %4, %2, %3, 10101010b vpshufd %4, %4, 10110001b vpblendd %1, %2, %3, 01010101b vpmaxub %1, %4, %1 vpsrld %4, %1, 16 vpmaxub %1, %1, %4 vpsrld %4, %1, 8 vpmaxub %1, %1, %4 %endmacro ; res=%1 src=%2 zero=%3 tmp=%4 add_to_res=%5 %macro AVX2_Sqsumbdw 5 vpunpcklbw %4, %2, %3 %if %5 vpmaddwd %4, %4, %4 vpaddd %1, %1, %4 %else vpmaddwd %1, %4, %4 %endif vpunpckhbw %4, %2, %3 vpmaddwd %4, %4, %4 vpaddd %1, %1, %4 %endmacro ; res=%1 src=%2 zero=%3 tmp=%4 add_to_res=%5 %macro AVX2_Sumbdw 5 %if %5 vpsadbw %4, %2, %3 vpaddd %1, %1, %4 %else vpsadbw %1, %2, %3 %endif %endmacro ; res=%1 a=%2 b=%3 a=%4 tmp=%5 %macro AVX2_AbsDiffub 5 vpsubusb %5, %2, %3 vpsubusb %1, %3, %4 vpor %1, %5, %1 %endmacro ; sad=%1 cur_data=%2 ref_data=%3 tmp=%4 accumulate_results=%5 %macro AVX2_Sadbdw 5 %if %5 vpsadbw %4, %2, %3 vpaddd %1, %1, %4 %else vpsadbw %1, %2, %3 %endif %endmacro ; sad=%1 sum_cur=%2 sqsum_cur=%3 cur_data=%4 ref_data=%5 zero=%6 tmp=%7 accumulate_results=%8 %macro AVX2_SadSumSqsumbdw 8 AVX2_Sadbdw %1, %4, %5, %7, %8 AVX2_Sumbdw %2, %4, %6, %7, %8 AVX2_Sqsumbdw %3, %4, %6, %7, %8 %endmacro ; sad=%1 pCur=%2 pRef=%3 tmp=%4 accumulate_results=%5 %macro AVX2_Sad 5 vmovdqu %4, [%2] AVX2_Sadbdw %1, %4, [%3], %4, %5 %endmacro ; sad=%1 sum_cur=%2 sqsum_cur=%3 pCur=%4 pRef=%5 zero=%6 tmp=%7,%8 accumulate_results=%9 %macro AVX2_SadSumSqsum 9 vmovdqu %7, [%4] AVX2_SadSumSqsumbdw %1, %2, %3, %7, [%5], %6, %8, %9 %endmacro ; sad=%1 sum_cur=%2 sqsum_cur=%3 sqdiff=%4 pCur=%5 pRef=%6 zero=%7 tmp=%8,%9,%10 accumulate_results=%11 %macro AVX2_SadSumSqsumSqdiff 11 vmovdqu %8, [%5] vmovdqu %9, [%6] AVX2_SadSumSqsumbdw %1, %2, %3, %8, %9, %7, %10, %11 AVX2_AbsDiffub %9, %8, %9, %8, %10 AVX2_Sqsumbdw %4, %9, %7, %10, %11 %endmacro ; sad=%1 sum_cur=%2 sum_ref=%3 mad=%4 pCur=%5 pRef=%6 zero=%7 tmp=%8,%9,%10 accumulate_results=%11 %macro AVX2_SadSdMad 11 vmovdqu %8, [%5] vmovdqu %9, [%6] AVX2_Sumbdw %2, %8, %7, %10, %11 AVX2_Sumbdw %3, %9, %7, %10, %11 AVX2_Sadbdw %1, %8, %9, %10, %11 %if %11 AVX2_AbsDiffub %9, %8, %9, %8, %10 vpmaxub %4, %4, %9 %else AVX2_AbsDiffub %4, %8, %9, %8, %10 %endif %endmacro ; sad=%1 sum_cur=%2 sum_ref=%3 mad=%4 sqdiff=%5 sqsum_cur=%6 pCur=%7 pRef=%8 zero=%9 tmp=%10,%11,%12 accumulate_results=%13 %macro AVX2_SadBgdSqdiff 13 %ifidn %12, 0 vmovdqu %10, [%7] AVX2_Sumbdw %2, %10, %9, %11, %13 AVX2_Sqsumbdw %6, %10, %9, %11, %13 vmovdqu %11, [%8] AVX2_Sadbdw %1, %10, %11, %10, %13 AVX2_Sumbdw %3, %11, %9, %10, %13 vmovdqu %10, [%7] %if %13 AVX2_AbsDiffub %11, %10, %11, [%7], %10 vpmaxub %4, %4, %11 AVX2_Sqsumbdw %5, %11, %9, %10, %13 %else AVX2_AbsDiffub %4, %10, %11, [%7], %10 AVX2_Sqsumbdw %5, %4, %9, %10, %13 %endif %else vmovdqu %10, [%7] vmovdqu %11, [%8] AVX2_Sadbdw %1, %10, %11, %12, %13 AVX2_Sumbdw %2, %10, %9, %12, %13 AVX2_Sumbdw %3, %11, %9, %12, %13 AVX2_Sqsumbdw %6, %10, %9, %12, %13 %if %13 AVX2_AbsDiffub %11, %10, %11, %10, %12 vpmaxub %4, %4, %11 AVX2_Sqsumbdw %5, %11, %9, %10, %13 %else AVX2_AbsDiffub %4, %10, %11, %10, %12 AVX2_Sqsumbdw %5, %4, %9, %10, %13 %endif %endif %endmacro ; p_dst=%1 mmreg_prefix=%2 data=%3 tmp=%4 second_blocks=%5 %macro AVX2_Store8x8Accdw 5 vpshufd %2%4, %2%3, 1000b %ifidni %2, x vmovlps [%1 + 8 * %5], x%4 %elif %5 == 0 vmovdqu [%1], %2%4 %else vmovlps [%1 + 8], x%4 vextracti128 x%4, %2%4, 1 vmovlps [%1 + 24], x%4 %endif %endmacro ; p_dst=%1 mmreg_prefix=%2 data=%3 tmp=%4 second_blocks=%5 %macro AVX2_Store8x8Accb 5 vpunpckhqdq %2%4, %2%3, %2%3 vpunpcklbw %2%4, %2%3, %2%4 %if %5 == 0 vmovd [%1 + 0], x%4 %ifidni %2, y vextracti128 x%4, %2%4, 1 vmovd [%1 + 4], x%4 %endif %else vpextrw [%1 + 2], x%4, 0 %ifidni %2, y vextracti128 x%4, %2%4, 1 vpextrw [%1 + 6], x%4, 0 %endif %endif %endmacro ; p_dst=%1 data=%2 tmp=%3,%4 second_blocks=%5 %macro AVX2_Store2x8x8Accb 5 vpunpckhqdq y%3, y%2, y%2 vpunpcklbw y%3, y%2, y%3 vextracti128 x%4, y%3, 1 vpsllq x%4, x%4, 32 vpblendd x%4, x%3, x%4, 1010b %if %5 vpslld x%4, x%4, 16 vpblendw x%4, x%4, [%1], 01010101b %endif vmovdqu [%1], x%4 %endmacro ; p_dst=%1 mmreg_prefix=%2 data=%3 tmp=%4 add_to_dst=%5 %macro AVX2_Store16x16Accdw 5 %ifidni %2, x %if %5 vmovd x%4, [%1 + 0] vpaddd x%3, x%4, x%3 %endif vmovd [%1 + 0], x%3 %elif %5 == 0 vmovd [%1 + 0], x%3 vextracti128 x%3, %2%3, 1 vmovd [%1 + 4], x%3 %else vextracti128 x%4, %2%3, 1 vpunpckldq x%4, x%3, x%4 vmovq x%3, [%1 + 0] vpaddd x%3, x%3, x%4 vmovlps [%1 + 0], x%3 %endif %endmacro ; p_dst1=%1 p_dst2=%2 i_dst_offset=%3 gpr_tmp=%4 mmreg_prefix=%5 data=%6 mm_tmp=%7 add_to_dst=%8 %macro AVX2_Store2x16x16Accdw 8 %ifidni %5, x mov %4, %1 %if %8 == 0 vmovd [%4 + %3], x%6 mov %4, %2 vpextrd [%4 + %3], x%6, 2 %else vmovd x%7, [%4 + %3] vpaddd x%7, x%7, x%6 vmovd [%4 + %3], x%7 mov %4, %2 vpbroadcastd x%7, [%4 + %3] vpaddd x%7, x%7, x%6 vpextrd [%4 + %3], x%7, 2 %endif %else vextracti128 x%7, %5%6, 1 vpblendd x%6, x%6, x%7, 1010b mov %4, %1 %if %8 == 0 vmovlps [%4 + %3], x%6 mov %4, %2 vmovhps [%4 + %3], x%6 %else vmovq x%7, [%4 + %3] vpaddd x%7, x%7, x%6 vmovlps [%4 + %3], x%7 mov %4, %2 vpbroadcastq x%7, [%4 + %3] vpaddd x%7, x%7, x%6 vmovhps [%4 + %3], x%7 %endif %endif %endmacro ; x/y-mm_prefix=%1 mm_clobber=%2,%3,%4,%5,%6 b_second_blocks=%7 %macro AVX2_CalcSad_8Lines 7 %define mm_tmp0 %2 %define mm_sad %3 %define mm_sad2 %4 %define mm_sad3 %5 %define mm_sad4 %6 %define b_second_blocks %7 %ifdef i_stride5 %define i_stride5_ i_stride5 %else lea r_tmp, [5 * i_stride] %define i_stride5_ r_tmp %endif ; Use multiple accumulators to shorten dependency chains and enable more parallelism. AVX2_Sad %1 %+ mm_sad, p_cur, p_ref, %1 %+ mm_tmp0, 0 AVX2_Sad %1 %+ mm_sad2, p_cur + 1 * i_stride, p_ref + 1 * i_stride, %1 %+ mm_tmp0, 0 AVX2_Sad %1 %+ mm_sad3, p_cur + 2 * i_stride, p_ref + 2 * i_stride, %1 %+ mm_tmp0, 0 AVX2_Sad %1 %+ mm_sad4, p_cur + 1 * i_stride3, p_ref + 1 * i_stride3, %1 %+ mm_tmp0, 0 AVX2_Sad %1 %+ mm_sad, p_cur + 4 * i_stride, p_ref + 4 * i_stride, %1 %+ mm_tmp0, 1 AVX2_Sad %1 %+ mm_sad2, p_cur + 1 * i_stride5_, p_ref + 1 * i_stride5_, %1 %+ mm_tmp0, 1 %ifdef i_stride7 %define i_stride7_ i_stride7 %else lea r_tmp, [i_stride + 2 * i_stride3] %define i_stride7_ r_tmp %endif AVX2_Sad %1 %+ mm_sad3, p_cur + 2 * i_stride3, p_ref + 2 * i_stride3, %1 %+ mm_tmp0, 1 AVX2_Sad %1 %+ mm_sad4, p_cur + 1 * i_stride7_, p_ref + 1 * i_stride7_, %1 %+ mm_tmp0, 1 %undef i_stride5_ %undef i_stride7_ ; Increment addresses for the next iteration. Doing this early is beneficial on Haswell. add p_cur, %1 %+ mm_width add p_ref, %1 %+ mm_width ; Collapse accumulators. vpaddd %1 %+ mm_sad, %1 %+ mm_sad, %1 %+ mm_sad2 vpaddd %1 %+ mm_sad3, %1 %+ mm_sad3, %1 %+ mm_sad4 vpaddd %1 %+ mm_sad, %1 %+ mm_sad, %1 %+ mm_sad3 AVX2_Store8x8Accdw p_sad8x8 + xcnt_unit * i_xcnt, %1, mm_sad, mm_tmp0, b_second_blocks vpaddd y %+ mm_sadframe, y %+ mm_sadframe, y %+ mm_sad %undef mm_tmp0 %undef mm_sad %undef mm_sad2 %undef mm_sad3 %undef mm_sad4 %undef b_second_blocks %endmacro ;************************************************************************************************************* ;void VAACalcSad_avx2( const uint8_t *cur_data, const uint8_t *ref_data, int32_t iPicWidth, int32_t iPicHeight ; int32_t iPicStride, int32_t *psadframe, int32_t *psad8x8) ;************************************************************************************************************* WELS_EXTERN VAACalcSad_avx2 %define p_sadframe ptrword arg6 %define p_sad8x8 ptrword arg7 %ifdef X86_32 %define saveregs r5, r6 %else %define saveregs rbx, rbp, r12 %endif %assign push_num 0 LOAD_5_PARA PUSH_XMM 7 SIGN_EXTENSION r2, r2d SIGN_EXTENSION r3, r3d SIGN_EXTENSION r4, r4d PUSHM saveregs %define mm_zero mm0 %define mm_sadframe mm6 vpxor x %+ mm_zero, x %+ mm_zero, x %+ mm_zero vmovdqa y %+ mm_sadframe, y %+ mm_zero and r2, -16 ; iPicWidth &= -16 jle .done ; bail if iPicWidth < 16 sar r3, 4 ; iPicHeight / 16 jle .done ; bail if iPicHeight < 16 shr r2, 2 ; iPicWidth / 4 %define p_cur r0 %define p_ref r1 %define i_xcnt r2 %define i_ycnt ptrword arg4 %define i_stride r4 %define xcnt_unit 4 %ifdef X86_32 mov i_ycnt, r3 mov r5, p_sad8x8 %define i_stride3 r3 %undef p_sad8x8 %define p_sad8x8 r5 %define r_tmp r6 lea i_stride3, [3 * i_stride] %else mov rbp, p_sad8x8 %define i_stride3 rbx %define i_stride5 r12 %define i_stride7 r6 %undef p_sad8x8 %define p_sad8x8 rbp lea i_stride3, [3 * i_stride] lea i_stride5, [5 * i_stride] lea i_stride7, [i_stride + 2 * i_stride3] %endif ; offset pointer so as to compensate for the i_xcnt offset below. sub p_sad8x8, 4 * 16 / xcnt_unit push i_xcnt %assign push_num push_num + 1 %define i_xcnt_load ptrword [r7] .height_loop: ; use end-of-line pointers so as to enable use of a negative counter as index. lea p_sad8x8, [p_sad8x8 + xcnt_unit * i_xcnt] ; use a negative loop counter so as to enable counting toward zero and indexing with the same counter. neg i_xcnt add i_xcnt, 16 / xcnt_unit jz .width_loop_upper8_remaining16 .width_loop_upper8: AVX2_CalcSad_8Lines y, mm1, mm2, mm3, mm4, mm5, 0 add i_xcnt, 32 / xcnt_unit jl .width_loop_upper8 jg .width_loop_upper8_end .width_loop_upper8_remaining16: AVX2_CalcSad_8Lines x, mm1, mm2, mm3, mm4, mm5, 0 .width_loop_upper8_end: lea p_cur, [p_cur + 8 * i_stride] lea p_ref, [p_ref + 8 * i_stride] xor i_xcnt, i_xcnt sub i_xcnt, i_xcnt_load lea p_cur, [p_cur + xcnt_unit * i_xcnt] lea p_ref, [p_ref + xcnt_unit * i_xcnt] add i_xcnt, 16 / xcnt_unit jz .width_loop_lower8_remaining16 .width_loop_lower8: AVX2_CalcSad_8Lines y, mm1, mm2, mm3, mm4, mm5, 1 add i_xcnt, 32 / xcnt_unit jl .width_loop_lower8 jg .width_loop_lower8_end .width_loop_lower8_remaining16: AVX2_CalcSad_8Lines x, mm1, mm2, mm3, mm4, mm5, 1 .width_loop_lower8_end: lea p_cur, [p_cur + 8 * i_stride] lea p_ref, [p_ref + 8 * i_stride] xor i_xcnt, i_xcnt sub i_xcnt, i_xcnt_load lea p_cur, [p_cur + xcnt_unit * i_xcnt] lea p_ref, [p_ref + xcnt_unit * i_xcnt] neg i_xcnt sub i_ycnt, 1 jnz .height_loop pop i_xcnt %assign push_num push_num - 1 %undef i_xcnt_load .done: mov r6, p_sadframe vextracti128 xmm2, y %+ mm_sadframe, 1 vpaddd xmm2, x %+ mm_sadframe, xmm2 vpunpckhqdq xmm1, xmm2, xmm2 vpaddd xmm2, xmm2, xmm1 vmovd [r6], xmm2 vzeroupper POPM saveregs POP_XMM LOAD_5_PARA_POP %undef p_cur %undef p_ref %undef i_xcnt %undef i_ycnt %undef i_stride %undef r_tmp %undef xcnt_unit %undef i_stride3 %undef i_stride5 %undef i_stride7 %undef mm_sadframe %undef mm_zero %undef saveregs %undef p_sadframe %undef p_sad8x8 ret ; x/y-mm_prefix=%1 mm_clobber=%2,%3,%4,%5,%6 b_second_blocks=%7 %macro AVX2_CalcSadVar_8Lines 7 %define mm_tmp0 %2 %define mm_tmp1 %3 %define mm_sad %4 %define mm_sum %5 %define mm_sqsum %6 %define b_second_blocks %7 ; Unroll for better performance on Haswell. ; Avoid unrolling for the 16 px case so as to reduce the code footprint. %ifidni %1, y lea r_tmp, [5 * i_stride] AVX2_SadSumSqsum %1 %+ mm_sad, %1 %+ mm_sum, %1 %+ mm_sqsum, p_cur, p_ref, %1 %+ mm_zero, %1 %+ mm_tmp0, %1 %+ mm_tmp1, 0 AVX2_SadSumSqsum %1 %+ mm_sad, %1 %+ mm_sum, %1 %+ mm_sqsum, p_cur + 1 * i_stride, p_ref + 1 * i_stride, %1 %+ mm_zero, %1 %+ mm_tmp0, %1 %+ mm_tmp1, 1 AVX2_SadSumSqsum %1 %+ mm_sad, %1 %+ mm_sum, %1 %+ mm_sqsum, p_cur + 2 * i_stride, p_ref + 2 * i_stride, %1 %+ mm_zero, %1 %+ mm_tmp0, %1 %+ mm_tmp1, 1 AVX2_SadSumSqsum %1 %+ mm_sad, %1 %+ mm_sum, %1 %+ mm_sqsum, p_cur + 1 * i_stride3, p_ref + 1 * i_stride3, %1 %+ mm_zero, %1 %+ mm_tmp0, %1 %+ mm_tmp1, 1 AVX2_SadSumSqsum %1 %+ mm_sad, %1 %+ mm_sum, %1 %+ mm_sqsum, p_cur + 4 * i_stride, p_ref + 4 * i_stride, %1 %+ mm_zero, %1 %+ mm_tmp0, %1 %+ mm_tmp1, 1 AVX2_SadSumSqsum %1 %+ mm_sad, %1 %+ mm_sum, %1 %+ mm_sqsum, p_cur + r_tmp, p_ref + r_tmp, %1 %+ mm_zero, %1 %+ mm_tmp0, %1 %+ mm_tmp1, 1 lea r_tmp, [i_stride + 2 * i_stride3] AVX2_SadSumSqsum %1 %+ mm_sad, %1 %+ mm_sum, %1 %+ mm_sqsum, p_cur + 2 * i_stride3, p_ref + 2 * i_stride3, %1 %+ mm_zero, %1 %+ mm_tmp0, %1 %+ mm_tmp1, 1 AVX2_SadSumSqsum %1 %+ mm_sad, %1 %+ mm_sum, %1 %+ mm_sqsum, p_cur + r_tmp, p_ref + r_tmp, %1 %+ mm_zero, %1 %+ mm_tmp0, %1 %+ mm_tmp1, 1 ; Increment addresses for the next iteration. Doing this early is beneficial on Haswell. add p_cur, %1 %+ mm_width add p_ref, %1 %+ mm_width %else vpxor x %+ mm_sad, x %+ mm_sad, x %+ mm_sad vpxor x %+ mm_sum, x %+ mm_sum, x %+ mm_sum vpxor x %+ mm_sqsum, x %+ mm_sqsum, x %+ mm_sqsum lea r_tmp, [8 * i_stride] add p_cur, r_tmp add p_ref, r_tmp neg r_tmp %%loop: AVX2_SadSumSqsum %1 %+ mm_sad, %1 %+ mm_sum, %1 %+ mm_sqsum, p_cur + r_tmp, p_ref + r_tmp, %1 %+ mm_zero, %1 %+ mm_tmp0, %1 %+ mm_tmp1, 1 add r_tmp, i_stride jl %%loop ; Increment addresses for the next iteration. Doing this early is beneficial on Haswell. lea r_tmp, [8 * i_stride - %1 %+ mm_width] sub p_cur, r_tmp sub p_ref, r_tmp %endif AVX2_Store8x8Accdw p_sad8x8 + 4 * i_xcnt, %1, mm_sad, mm_tmp1, b_second_blocks vpaddd y %+ mm_sadframe, y %+ mm_sadframe, y %+ mm_sad vpunpcklqdq %1 %+ mm_tmp0, %1 %+ mm_sum, %1 %+ mm_sqsum vpunpckhqdq %1 %+ mm_tmp1, %1 %+ mm_sum, %1 %+ mm_sqsum vpaddd %1 %+ mm_tmp0, %1 %+ mm_tmp0, %1 %+ mm_tmp1 vpshufd %1 %+ mm_tmp1, %1 %+ mm_tmp0, 10110001b vpaddd %1 %+ mm_tmp0, %1 %+ mm_tmp0, %1 %+ mm_tmp1 AVX2_Store2x16x16Accdw p_sum16x16, p_sqsum16x16, i_xcnt, r_tmp, %1, mm_tmp0, mm_tmp1, b_second_blocks %undef mm_tmp0 %undef mm_tmp1 %undef mm_sad %undef mm_sum %undef mm_sqsum %undef b_second_blocks %endmacro ;************************************************************************************************************* ;void VAACalcSadVar_avx2( const uint8_t *cur_data, const uint8_t *ref_data, int32_t iPicWidth, int32_t iPicHeight ; int32_t iPicStride, int32_t *psadframe, int32_t *psad8x8, int32_t *psum16x16, int32_t *psqsum16x16) ;************************************************************************************************************* WELS_EXTERN VAACalcSadVar_avx2 %define p_sadframe ptrword arg6 %define p_sad8x8 ptrword arg7 %define p_sum16x16 ptrword arg8 %define p_sqsum16x16 ptrword arg9 %ifdef X86_32 %define saveregs r5, r6 %else %define saveregs rbx, rbp, r12, r13 %endif %assign push_num 0 LOAD_5_PARA PUSH_XMM 7 SIGN_EXTENSION r2, r2d SIGN_EXTENSION r3, r3d SIGN_EXTENSION r4, r4d PUSHM saveregs %define mm_zero mm0 %define mm_sadframe mm6 vpxor x %+ mm_zero, x %+ mm_zero, x %+ mm_zero vmovdqa y %+ mm_sadframe, y %+ mm_zero and r2, -16 ; iPicWidth &= -16 jle .done ; bail if iPicWidth < 16 sar r3, 4 ; iPicHeight / 16 jle .done ; bail if iPicHeight < 16 shr r2, 2 ; iPicWidth / 4 %define p_cur r0 %define p_ref r1 %define i_xcnt r2 %define i_ycnt ptrword arg4 %define i_stride r4 %define r_tmp r6 %define xcnt_unit 4 %ifdef X86_32 mov i_ycnt, r3 mov r3, p_sad8x8 %undef p_sad8x8 %define p_sad8x8 r3 %define i_stride3 r5 %else mov rbp, p_sad8x8 mov r12, p_sum16x16 mov r13, p_sqsum16x16 %undef p_sad8x8 %undef p_sum16x16 %undef p_sqsum16x16 %define p_sad8x8 rbp %define p_sum16x16 r12 %define p_sqsum16x16 r13 %define i_stride3 rbx %endif lea i_stride3, [3 * i_stride] ; offset pointers so as to compensate for the i_xcnt offset below. sub p_sad8x8, 4 * 16 / xcnt_unit sub p_sum16x16, 1 * 16 / xcnt_unit sub p_sqsum16x16, 1 * 16 / xcnt_unit ; use a negative loop counter so as to enable counting toward zero and indexing with the same counter. neg i_xcnt .height_loop: push i_xcnt %assign push_num push_num + 1 %define i_xcnt_load ptrword [r7] ; use end-of-line pointers so as to enable use of a negative counter as index. lea r_tmp, [xcnt_unit * i_xcnt] sub p_sad8x8, r_tmp sub p_sum16x16, i_xcnt sub p_sqsum16x16, i_xcnt add i_xcnt, 16 / xcnt_unit jz .width_loop_upper8_remaining16 .width_loop_upper8: AVX2_CalcSadVar_8Lines y, mm1, mm2, mm3, mm4, mm5, 0 add i_xcnt, 32 / xcnt_unit jl .width_loop_upper8 jg .width_loop_upper8_end .width_loop_upper8_remaining16: AVX2_CalcSadVar_8Lines x, mm1, mm2, mm3, mm4, mm5, 0 .width_loop_upper8_end: lea p_cur, [p_cur + 8 * i_stride] lea p_ref, [p_ref + 8 * i_stride] mov i_xcnt, i_xcnt_load lea p_cur, [p_cur + xcnt_unit * i_xcnt] lea p_ref, [p_ref + xcnt_unit * i_xcnt] add i_xcnt, 16 / xcnt_unit jz .width_loop_lower8_remaining16 .width_loop_lower8: AVX2_CalcSadVar_8Lines y, mm1, mm2, mm3, mm4, mm5, 1 add i_xcnt, 32 / xcnt_unit jl .width_loop_lower8 jg .width_loop_lower8_end .width_loop_lower8_remaining16: AVX2_CalcSadVar_8Lines x, mm1, mm2, mm3, mm4, mm5, 1 .width_loop_lower8_end: lea p_cur, [p_cur + 8 * i_stride] lea p_ref, [p_ref + 8 * i_stride] %undef i_xcnt_load pop i_xcnt %assign push_num push_num - 1 lea p_cur, [p_cur + xcnt_unit * i_xcnt] lea p_ref, [p_ref + xcnt_unit * i_xcnt] sub i_ycnt, 1 jnz .height_loop .done: mov r_tmp, p_sadframe vextracti128 xmm2, y %+ mm_sadframe, 1 vpaddd xmm2, x %+ mm_sadframe, xmm2 vpunpckhqdq xmm1, xmm2, xmm2 vpaddd xmm2, xmm2, xmm1 vmovd [r_tmp], xmm2 vzeroupper POPM saveregs POP_XMM LOAD_5_PARA_POP %undef p_cur %undef p_ref %undef i_xcnt %undef i_ycnt %undef i_stride %undef i_stride3 %undef r_tmp %undef xcnt_unit %undef mm_sadframe %undef mm_zero %undef saveregs %undef p_sadframe %undef p_sad8x8 %undef p_sum16x16 %undef p_sqsum16x16 ret ; x/y-mm_prefix=%1 mm_clobber=%2,%3,%4,%5,%6,%7,%8 b_second_blocks=%9 %macro AVX2_CalcSadSsd_8Lines 9 %define mm_tmp0 %2 %define mm_tmp1 %3 %define mm_tmp2 %4 %define mm_sad %5 %define mm_sum %6 %define mm_sqsum %7 %define mm_sqdiff %8 %define b_second_blocks %9 ; Unroll for better performance on Haswell. ; Avoid unrolling for the 16 px case so as to reduce the code footprint. %ifidni %1, y %ifdef i_stride5 lea r_tmp, [i_stride + 2 * i_stride3] %define i_stride5_ i_stride5 %else lea r_tmp, [5 * i_stride] %define i_stride5_ r_tmp %endif AVX2_SadSumSqsumSqdiff %1 %+ mm_sad, %1 %+ mm_sum, %1 %+ mm_sqsum, %1 %+ mm_sqdiff, p_cur, p_ref, %1 %+ mm_zero, %1 %+ mm_tmp0, %1 %+ mm_tmp1, %1 %+ mm_tmp2, 0 AVX2_SadSumSqsumSqdiff %1 %+ mm_sad, %1 %+ mm_sum, %1 %+ mm_sqsum, %1 %+ mm_sqdiff, p_cur + 1 * i_stride, p_ref + 1 * i_stride, %1 %+ mm_zero, %1 %+ mm_tmp0, %1 %+ mm_tmp1, %1 %+ mm_tmp2, 1 AVX2_SadSumSqsumSqdiff %1 %+ mm_sad, %1 %+ mm_sum, %1 %+ mm_sqsum, %1 %+ mm_sqdiff, p_cur + 2 * i_stride, p_ref + 2 * i_stride, %1 %+ mm_zero, %1 %+ mm_tmp0, %1 %+ mm_tmp1, %1 %+ mm_tmp2, 1 AVX2_SadSumSqsumSqdiff %1 %+ mm_sad, %1 %+ mm_sum, %1 %+ mm_sqsum, %1 %+ mm_sqdiff, p_cur + 1 * i_stride3, p_ref + 1 * i_stride3, %1 %+ mm_zero, %1 %+ mm_tmp0, %1 %+ mm_tmp1, %1 %+ mm_tmp2, 1 AVX2_SadSumSqsumSqdiff %1 %+ mm_sad, %1 %+ mm_sum, %1 %+ mm_sqsum, %1 %+ mm_sqdiff, p_cur + 4 * i_stride, p_ref + 4 * i_stride, %1 %+ mm_zero, %1 %+ mm_tmp0, %1 %+ mm_tmp1, %1 %+ mm_tmp2, 1 AVX2_SadSumSqsumSqdiff %1 %+ mm_sad, %1 %+ mm_sum, %1 %+ mm_sqsum, %1 %+ mm_sqdiff, p_cur + 1 * i_stride5_, p_ref + 1 * i_stride5_, %1 %+ mm_zero, %1 %+ mm_tmp0, %1 %+ mm_tmp1, %1 %+ mm_tmp2, 1 %ifndef i_stride5 lea r_tmp, [i_stride + 2 * i_stride3] %endif %undef i_stride5_ AVX2_SadSumSqsumSqdiff %1 %+ mm_sad, %1 %+ mm_sum, %1 %+ mm_sqsum, %1 %+ mm_sqdiff, p_cur + 2 * i_stride3, p_ref + 2 * i_stride3, %1 %+ mm_zero, %1 %+ mm_tmp0, %1 %+ mm_tmp1, %1 %+ mm_tmp2, 1 AVX2_SadSumSqsumSqdiff %1 %+ mm_sad, %1 %+ mm_sum, %1 %+ mm_sqsum, %1 %+ mm_sqdiff, p_cur + r_tmp, p_ref + r_tmp, %1 %+ mm_zero, %1 %+ mm_tmp0, %1 %+ mm_tmp1, %1 %+ mm_tmp2, 1 ; Increment addresses for the next iteration. Doing this early is beneficial on Haswell. add p_cur, %1 %+ mm_width add p_ref, %1 %+ mm_width %else vpxor x %+ mm_sad, x %+ mm_sad, x %+ mm_sad vpxor x %+ mm_sum, x %+ mm_sum, x %+ mm_sum vpxor x %+ mm_sqsum, x %+ mm_sqsum, x %+ mm_sqsum vpxor x %+ mm_sqdiff, x %+ mm_sqdiff, x %+ mm_sqdiff lea r_tmp, [8 * i_stride] add p_cur, r_tmp add p_ref, r_tmp neg r_tmp %%loop: AVX2_SadSumSqsumSqdiff %1 %+ mm_sad, %1 %+ mm_sum, %1 %+ mm_sqsum, %1 %+ mm_sqdiff, p_cur + r_tmp, p_ref + r_tmp, %1 %+ mm_zero, %1 %+ mm_tmp0, %1 %+ mm_tmp1, %1 %+ mm_tmp2, 1 add r_tmp, i_stride jl %%loop ; Increment addresses for the next iteration. Doing this early is beneficial on Haswell. lea r_tmp, [8 * i_stride - %1 %+ mm_width] sub p_cur, r_tmp sub p_ref, r_tmp %endif mov r_tmp, p_sad8x8 AVX2_Store8x8Accdw r_tmp + 4 * i_xcnt, %1, mm_sad, mm_tmp1, b_second_blocks %ifdef X86_32 vpaddd y %+ mm_tmp1, y %+ mm_sad, sadframe_acc vmovdqa sadframe_acc, y %+ mm_tmp1 %else vpaddd sadframe_acc, sadframe_acc, y %+ mm_sad %endif mov r_tmp, i_xcnt add r_tmp, p_sum16x16 vpunpckhqdq %1 %+ mm_tmp1, %1 %+ mm_sum, %1 %+ mm_sum vpaddd %1 %+ mm_tmp0, %1 %+ mm_sum, %1 %+ mm_tmp1 AVX2_Store16x16Accdw r_tmp, %1, mm_tmp0, mm_tmp1, b_second_blocks vpunpcklqdq %1 %+ mm_tmp0, %1 %+ mm_sqsum, %1 %+ mm_sqdiff vpunpckhqdq %1 %+ mm_tmp1, %1 %+ mm_sqsum, %1 %+ mm_sqdiff vpaddd %1 %+ mm_tmp0, %1 %+ mm_tmp0, %1 %+ mm_tmp1 vpshufd %1 %+ mm_tmp1, %1 %+ mm_tmp0, 10110001b vpaddd %1 %+ mm_tmp0, %1 %+ mm_tmp0, %1 %+ mm_tmp1 AVX2_Store2x16x16Accdw p_sqsum16x16, p_sqdiff16x16, i_xcnt, r_tmp, %1, mm_tmp0, mm_tmp1, b_second_blocks %undef mm_tmp0 %undef mm_tmp1 %undef mm_tmp2 %undef mm_sad %undef mm_sum %undef mm_sqsum %undef mm_sqdiff %undef b_second_blocks %endmacro ;************************************************************************************************************* ;void VAACalcSadSsd_avx2(const uint8_t *cur_data, const uint8_t *ref_data, int32_t iPicWidth, int32_t iPicHeight, ; int32_t iPicStride,int32_t *psadframe, int32_t *psad8x8, int32_t *psum16x16, int32_t *psqsum16x16, int32_t *psqdiff16x16) ;************************************************************************************************************* WELS_EXTERN VAACalcSadSsd_avx2 %define p_sadframe ptrword arg6 %define p_sad8x8 ptrword arg7 %define p_sum16x16 ptrword arg8 %define p_sqsum16x16 ptrword arg9 %define p_sqdiff16x16 ptrword arg10 %ifdef X86_32 %define saveregs r5, r6 %else %define saveregs rbx, rbp, r12, r13, r14, r15 %endif %assign push_num 0 LOAD_5_PARA PUSH_XMM 9 SIGN_EXTENSION r2, r2d SIGN_EXTENSION r3, r3d SIGN_EXTENSION r4, r4d PUSHM saveregs %define mm_zero mm0 vpxor x %+ mm_zero, x %+ mm_zero, x %+ mm_zero %ifdef X86_32 STACK_ALLOC r5, ymm_width, ymm_width %define sadframe_acc_addr r5 %define sadframe_acc [sadframe_acc_addr] %else %define sadframe_acc ymm8 %define xsadframe_acc xmm8 %endif vmovdqa sadframe_acc, y %+ mm_zero and r2, -16 ; iPicWidth &= -16 jle .done ; bail if iPicWidth < 16 sar r3, 4 ; iPicHeight / 16 jle .done ; bail if iPicHeight < 16 shr r2, 2 ; iPicWidth / 4 %define p_cur r0 %define p_ref r1 %define i_xcnt r2 %define i_ycnt ptrword arg4 %define i_stride r4 %define r_tmp r6 %define xcnt_unit 4 %ifdef X86_32 mov i_ycnt, r3 %define i_stride3 r3 %else mov r12, p_sad8x8 mov r13, p_sum16x16 mov r14, p_sqsum16x16 mov r15, p_sqdiff16x16 %undef p_sad8x8 %undef p_sum16x16 %undef p_sqsum16x16 %undef p_sqdiff16x16 %define p_sad8x8 r12 %define p_sum16x16 r13 %define p_sqsum16x16 r14 %define p_sqdiff16x16 r15 %define i_stride3 rbx %define i_stride5 rbp lea i_stride5, [5 * i_stride] %endif lea i_stride3, [3 * i_stride] ; offset pointers so as to compensate for i_xcnt offset below. sub p_sad8x8, 4 * 16 / xcnt_unit sub p_sum16x16, 1 * 16 / xcnt_unit sub p_sqsum16x16, 1 * 16 / xcnt_unit sub p_sqdiff16x16, 1 * 16 / xcnt_unit ; use a negative loop counter so as to enable counting toward zero and indexing with the same counter. neg i_xcnt .height_loop: push i_xcnt %assign push_num push_num + 1 %define i_xcnt_load ptrword [r7] ; use end-of-line pointers so as to enable use of a negative counter as index. lea r_tmp, [xcnt_unit * i_xcnt] sub p_sad8x8, r_tmp sub p_sum16x16, i_xcnt sub p_sqsum16x16, i_xcnt sub p_sqdiff16x16, i_xcnt add i_xcnt, 16 / xcnt_unit jz .width_loop_upper8_remaining16 .width_loop_upper8: AVX2_CalcSadSsd_8Lines y, mm1, mm2, mm3, mm4, mm5, mm6, mm7, 0 add i_xcnt, 32 / xcnt_unit jl .width_loop_upper8 jg .width_loop_upper8_end .width_loop_upper8_remaining16: AVX2_CalcSadSsd_8Lines x, mm1, mm2, mm3, mm4, mm5, mm6, mm7, 0 .width_loop_upper8_end: lea p_cur, [p_cur + 8 * i_stride] lea p_ref, [p_ref + 8 * i_stride] mov i_xcnt, i_xcnt_load lea p_cur, [p_cur + xcnt_unit * i_xcnt] lea p_ref, [p_ref + xcnt_unit * i_xcnt] add i_xcnt, 16 / xcnt_unit jz .width_loop_lower8_remaining16 .width_loop_lower8: AVX2_CalcSadSsd_8Lines y, mm1, mm2, mm3, mm4, mm5, mm6, mm7, 1 add i_xcnt, 32 / xcnt_unit jl .width_loop_lower8 jg .width_loop_lower8_end .width_loop_lower8_remaining16: AVX2_CalcSadSsd_8Lines x, mm1, mm2, mm3, mm4, mm5, mm6, mm7, 1 .width_loop_lower8_end: lea p_cur, [p_cur + 8 * i_stride] lea p_ref, [p_ref + 8 * i_stride] %undef i_xcnt_load pop i_xcnt %assign push_num push_num - 1 lea p_cur, [p_cur + xcnt_unit * i_xcnt] lea p_ref, [p_ref + xcnt_unit * i_xcnt] sub i_ycnt, 1 jnz .height_loop .done: mov r_tmp, p_sadframe %ifdef X86_32 vmovdqa xmm2, sadframe_acc vpaddd xmm2, xmm2, [sadframe_acc_addr + xmm_width] %else vextracti128 xmm2, sadframe_acc, 1 vpaddd xmm2, xsadframe_acc, xmm2 %endif vpunpckhqdq xmm1, xmm2, xmm2 vpaddd xmm2, xmm2, xmm1 vmovd [r_tmp], xmm2 vzeroupper %ifdef X86_32 STACK_DEALLOC %endif POPM saveregs POP_XMM LOAD_5_PARA_POP %undef p_cur %undef p_ref %undef i_xcnt %undef i_ycnt %undef i_stride %undef i_stride3 %undef i_stride5 %undef r_tmp %undef xcnt_unit %undef sadframe_acc %undef sadframe_acc_addr %undef xsadframe_acc %undef mm_zero %undef saveregs %undef p_sadframe %undef p_sad8x8 %undef p_sum16x16 %undef p_sqsum16x16 %undef p_sqdiff16x16 ret ; x/y-mm_prefix=%1 mm_clobber=%2,%3,%4,%5,%6,%7,%8 b_second_blocks=%9 %macro AVX2_CalcSadBgd_8Lines 9 %define mm_tmp0 %2 %define mm_tmp1 %3 %define mm_tmp2 %8 %define mm_mad %4 %define mm_sumcur %5 %define mm_sumref %6 %define mm_sad %7 %define b_second_blocks %9 ; Unroll for better performance on Haswell. ; Avoid unrolling for the 16 px case so as to reduce the code footprint. %ifidni %1, y lea r_tmp, [5 * i_stride] AVX2_SadSdMad %1 %+ mm_sad, %1 %+ mm_sumcur, %1 %+ mm_sumref, %1 %+ mm_mad, p_cur, p_ref, %1 %+ mm_zero, %1 %+ mm_tmp0, %1 %+ mm_tmp1, %1 %+ mm_tmp2, 0 AVX2_SadSdMad %1 %+ mm_sad, %1 %+ mm_sumcur, %1 %+ mm_sumref, %1 %+ mm_mad, p_cur + 1 * i_stride, p_ref + 1 * i_stride, %1 %+ mm_zero, %1 %+ mm_tmp0, %1 %+ mm_tmp1, %1 %+ mm_tmp2, 1 AVX2_SadSdMad %1 %+ mm_sad, %1 %+ mm_sumcur, %1 %+ mm_sumref, %1 %+ mm_mad, p_cur + 2 * i_stride, p_ref + 2 * i_stride, %1 %+ mm_zero, %1 %+ mm_tmp0, %1 %+ mm_tmp1, %1 %+ mm_tmp2, 1 AVX2_SadSdMad %1 %+ mm_sad, %1 %+ mm_sumcur, %1 %+ mm_sumref, %1 %+ mm_mad, p_cur + 1 * i_stride3, p_ref + 1 * i_stride3, %1 %+ mm_zero, %1 %+ mm_tmp0, %1 %+ mm_tmp1, %1 %+ mm_tmp2, 1 AVX2_SadSdMad %1 %+ mm_sad, %1 %+ mm_sumcur, %1 %+ mm_sumref, %1 %+ mm_mad, p_cur + 4 * i_stride, p_ref + 4 * i_stride, %1 %+ mm_zero, %1 %+ mm_tmp0, %1 %+ mm_tmp1, %1 %+ mm_tmp2, 1 AVX2_SadSdMad %1 %+ mm_sad, %1 %+ mm_sumcur, %1 %+ mm_sumref, %1 %+ mm_mad, p_cur + r_tmp, p_ref + r_tmp, %1 %+ mm_zero, %1 %+ mm_tmp0, %1 %+ mm_tmp1, %1 %+ mm_tmp2, 1 lea r_tmp, [i_stride + 2 * i_stride3] AVX2_SadSdMad %1 %+ mm_sad, %1 %+ mm_sumcur, %1 %+ mm_sumref, %1 %+ mm_mad, p_cur + 2 * i_stride3, p_ref + 2 * i_stride3, %1 %+ mm_zero, %1 %+ mm_tmp0, %1 %+ mm_tmp1, %1 %+ mm_tmp2, 1 AVX2_SadSdMad %1 %+ mm_sad, %1 %+ mm_sumcur, %1 %+ mm_sumref, %1 %+ mm_mad, p_cur + r_tmp, p_ref + r_tmp, %1 %+ mm_zero, %1 %+ mm_tmp0, %1 %+ mm_tmp1, %1 %+ mm_tmp2, 1 ; Increment addresses for the next iteration. Doing this early is beneficial on Haswell. add p_cur, %1 %+ mm_width add p_ref, %1 %+ mm_width %else vpxor x %+ mm_sad, x %+ mm_sad, x %+ mm_sad vpxor x %+ mm_sumcur, x %+ mm_sumcur, x %+ mm_sumcur vpxor x %+ mm_sumref, x %+ mm_sumref, x %+ mm_sumref vpxor x %+ mm_mad, x %+ mm_mad, x %+ mm_mad lea r_tmp, [8 * i_stride] add p_cur, r_tmp add p_ref, r_tmp neg r_tmp %%loop: AVX2_SadSdMad %1 %+ mm_sad, %1 %+ mm_sumcur, %1 %+ mm_sumref, %1 %+ mm_mad, p_cur + r_tmp, p_ref + r_tmp, %1 %+ mm_zero, %1 %+ mm_tmp0, %1 %+ mm_tmp1, %1 %+ mm_tmp2, 1 add r_tmp, i_stride jl %%loop ; Increment addresses for the next iteration. Doing this early is beneficial on Haswell. lea r_tmp, [8 * i_stride - %1 %+ mm_width] sub p_cur, r_tmp sub p_ref, r_tmp %endif mov r_tmp, p_sad8x8 AVX2_Store8x8Accdw r_tmp + 4 * i_xcnt, %1, mm_sad, mm_tmp1, b_second_blocks %ifdef X86_32 vpaddd y %+ mm_tmp1, y %+ mm_sad, sadframe_acc vmovdqa sadframe_acc, y %+ mm_tmp1 %else vpaddd sadframe_acc, sadframe_acc, y %+ mm_sad %endif mov r_tmp, p_sd8x8 vpsubd %1 %+ mm_tmp0, %1 %+ mm_sumcur, %1 %+ mm_sumref AVX2_Store8x8Accdw r_tmp + 4 * i_xcnt, %1, mm_tmp0, mm_tmp1, b_second_blocks ; Coalesce store and horizontal reduction of MAD accumulator for even and ; odd iterations so as to enable more parallelism. %ifidni %1, y test i_xcnt, 32 / xcnt_unit jz %%preserve_mad mov r_tmp, p_mad8x8 AVX2_Maxubq2 y %+ mm_mad, y %+ mm_mad, prev_mad, y %+ mm_tmp0 AVX2_Store2x8x8Accb r_tmp + i_xcnt - 8, mm_mad, mm_tmp0, mm_tmp1, b_second_blocks %%preserve_mad: vmovdqa prev_mad, y %+ mm_mad %else mov r_tmp, p_mad8x8 AVX2_Maxubq %1 %+ mm_mad, %1 %+ mm_mad, %1 %+ mm_tmp0 AVX2_Store8x8Accb r_tmp + i_xcnt, %1, mm_mad, mm_tmp0, b_second_blocks %endif %undef mm_tmp0 %undef mm_tmp1 %undef mm_tmp2 %undef mm_mad %undef mm_sumcur %undef mm_sumref %undef mm_sad %undef b_second_blocks %endmacro ; Store remaining MAD accumulator for width & 32 cases. ; width/xcnt_unit=%1 mm_tmp=%2,%3 b_second_blocks=%4 %macro AVX2_StoreRemainingSingleMad 4 test %1, 32 / xcnt_unit jz %%skip mov r_tmp, p_mad8x8 vmovdqa y%2, prev_mad AVX2_Maxubq y%2, y%2, y%3 AVX2_Store8x8Accb r_tmp + i_xcnt - 8, y, %2, %3, %4 %%skip: %endmacro ;************************************************************************************************************* ;void VAACalcSadBgd_avx2(const uint8_t *cur_data, const uint8_t *ref_data, int32_t iPicWidth, int32_t iPicHeight, ; int32_t iPicStride, int32_t *psadframe, int32_t *psad8x8, int32_t *p_sd8x8, uint8_t *p_mad8x8) ;************************************************************************************************************* WELS_EXTERN VAACalcSadBgd_avx2 %define p_sadframe arg6 %define p_sad8x8 arg7 %define p_sd8x8 arg8 %define p_mad8x8 arg9 %ifdef X86_32 %define saveregs r5, r6 %else %define saveregs rbx, rbp, r12, r13 %endif %assign push_num 0 LOAD_5_PARA PUSH_XMM 10 SIGN_EXTENSION r2, r2d SIGN_EXTENSION r3, r3d SIGN_EXTENSION r4, r4d PUSHM saveregs %define mm_zero mm0 vpxor x %+ mm_zero, x %+ mm_zero, x %+ mm_zero %ifdef X86_32 STACK_ALLOC r5, 2 * ymm_width, ymm_width %define sadframe_acc_addr r5 %define sadframe_acc [sadframe_acc_addr] %define prev_mad [r5 + ymm_width] %else %define sadframe_acc ymm8 %define xsadframe_acc xmm8 %define prev_mad ymm9 %endif vmovdqa sadframe_acc, y %+ mm_zero and r2, -16 ; iPicWidth &= -16 jle .done ; bail if iPicWidth < 16 sar r3, 4 ; iPicHeight / 16 jle .done ; bail if iPicHeight < 16 shr r2, 2 ; iPicWidth / 4 %define p_cur r0 %define p_ref r1 %define i_xcnt r2 %define i_ycnt ptrword arg4 %define i_stride r4 %define r_tmp r6 %define xcnt_unit 4 %ifdef X86_32 mov i_ycnt, r3 %define i_stride3 r3 %else mov rbp, p_sad8x8 mov r12, p_sd8x8 mov r13, p_mad8x8 %undef p_sad8x8 %undef p_sd8x8 %undef p_mad8x8 %define p_sad8x8 rbp %define p_sd8x8 r12 %define p_mad8x8 r13 %define i_stride3 rbx %endif lea i_stride3, [3 * i_stride] ; offset pointers to compensate for the i_xcnt offset below. mov r_tmp, i_xcnt and r_tmp, 64 / xcnt_unit - 1 sub p_mad8x8, r_tmp shl r_tmp, 2 sub p_sad8x8, r_tmp sub p_sd8x8, r_tmp .height_loop: push i_xcnt %assign push_num push_num + 1 %define i_xcnt_load ptrword [r7] ; use end-of-line pointers so as to enable use of a negative counter as index. lea r_tmp, [xcnt_unit * i_xcnt] add p_sad8x8, r_tmp add p_sd8x8, r_tmp add p_mad8x8, i_xcnt and i_xcnt, -(64 / xcnt_unit) jz .width_loop_upper8_64x_end ; use a negative loop counter to enable counting toward zero and indexing with the same counter. neg i_xcnt .width_loop_upper8: AVX2_CalcSadBgd_8Lines y, mm1, mm2, mm3, mm4, mm5, mm6, mm7, 0 add i_xcnt, 32 / xcnt_unit jl .width_loop_upper8 jg .width_loop_upper8_32x_end .width_loop_upper8_64x_end: test i_xcnt_load, 32 / xcnt_unit jnz .width_loop_upper8 .width_loop_upper8_32x_end: AVX2_StoreRemainingSingleMad i_xcnt_load, mm1, mm2, 0 test i_xcnt_load, 16 / xcnt_unit jz .width_loop_upper8_end ; remaining 16. AVX2_CalcSadBgd_8Lines x, mm1, mm2, mm3, mm4, mm5, mm6, mm7, 0 .width_loop_upper8_end: lea p_cur, [p_cur + 8 * i_stride] lea p_ref, [p_ref + 8 * i_stride] mov i_xcnt, i_xcnt_load lea r_tmp, [xcnt_unit * i_xcnt] sub p_cur, r_tmp sub p_ref, r_tmp and i_xcnt, -(64 / xcnt_unit) jz .width_loop_lower8_64x_end neg i_xcnt .width_loop_lower8: AVX2_CalcSadBgd_8Lines y, mm1, mm2, mm3, mm4, mm5, mm6, mm7, 1 add i_xcnt, 32 / xcnt_unit jl .width_loop_lower8 jg .width_loop_lower8_32x_end .width_loop_lower8_64x_end: test i_xcnt_load, 32 / xcnt_unit jnz .width_loop_lower8 .width_loop_lower8_32x_end: AVX2_StoreRemainingSingleMad i_xcnt_load, mm1, mm2, 1 test i_xcnt_load, 16 / xcnt_unit jz .width_loop_lower8_end ; remaining 16. AVX2_CalcSadBgd_8Lines x, mm1, mm2, mm3, mm4, mm5, mm6, mm7, 1 .width_loop_lower8_end: lea p_cur, [p_cur + 8 * i_stride] lea p_ref, [p_ref + 8 * i_stride] pop i_xcnt %undef i_xcnt_load %assign push_num push_num - 1 lea r_tmp, [xcnt_unit * i_xcnt] sub p_cur, r_tmp sub p_ref, r_tmp sub i_ycnt, 1 jnz .height_loop .done: mov r_tmp, p_sadframe %ifdef X86_32 vmovdqa xmm2, sadframe_acc vpaddd xmm2, xmm2, [sadframe_acc_addr + xmm_width] %else vextracti128 xmm2, sadframe_acc, 1 vpaddd xmm2, xsadframe_acc, xmm2 %endif vpunpckhqdq xmm1, xmm2, xmm2 vpaddd xmm2, xmm2, xmm1 vmovd [r_tmp], xmm2 vzeroupper %ifdef X86_32 STACK_DEALLOC %endif POPM saveregs POP_XMM LOAD_5_PARA_POP %undef p_cur %undef p_ref %undef i_xcnt %undef i_ycnt %undef i_stride %undef i_stride3 %undef r_tmp %undef xcnt_unit %undef sadframe_acc %undef sadframe_acc_addr %undef xsadframe_acc %undef prev_mad %undef mm_zero %undef saveregs %undef p_sadframe %undef p_sad8x8 %undef p_sd8x8 %undef p_mad8x8 ret ; x/y-mm_prefix=%1 mm_clobber=%2,%3,%4,%5,%6,%7,%8,%9,%10 b_second_blocks=%11 %macro AVX2_CalcSadSsdBgd_8Lines 11 %define mm_tmp0 %2 %define mm_tmp1 %3 %define mm_sad %4 %define mm_sum %5 %define mm_sumref %6 %define mm_mad %7 %define mm_sqsum %8 %define mm_sqdiff %9 %ifidn %10, 0 %define tmp2 0 %else %define tmp2 %1 %+ %10 %endif %define b_second_blocks %11 ; Unroll for better performance on Haswell. ; Avoid unrolling for the 16 px case so as to reduce the code footprint. %ifidni %1, y lea r_tmp, [5 * i_stride] AVX2_SadBgdSqdiff %1 %+ mm_sad, %1 %+ mm_sum, %1 %+ mm_sumref, %1 %+ mm_mad, %1 %+ mm_sqdiff, %1 %+ mm_sqsum, p_cur, p_ref, %1 %+ mm_zero, %1 %+ mm_tmp0, %1 %+ mm_tmp1, tmp2, 0 AVX2_SadBgdSqdiff %1 %+ mm_sad, %1 %+ mm_sum, %1 %+ mm_sumref, %1 %+ mm_mad, %1 %+ mm_sqdiff, %1 %+ mm_sqsum, p_cur + 1 * i_stride, p_ref + 1 * i_stride, %1 %+ mm_zero, %1 %+ mm_tmp0, %1 %+ mm_tmp1, tmp2, 1 AVX2_SadBgdSqdiff %1 %+ mm_sad, %1 %+ mm_sum, %1 %+ mm_sumref, %1 %+ mm_mad, %1 %+ mm_sqdiff, %1 %+ mm_sqsum, p_cur + 2 * i_stride, p_ref + 2 * i_stride, %1 %+ mm_zero, %1 %+ mm_tmp0, %1 %+ mm_tmp1, tmp2, 1 AVX2_SadBgdSqdiff %1 %+ mm_sad, %1 %+ mm_sum, %1 %+ mm_sumref, %1 %+ mm_mad, %1 %+ mm_sqdiff, %1 %+ mm_sqsum, p_cur + 1 * i_stride3, p_ref + 1 * i_stride3, %1 %+ mm_zero, %1 %+ mm_tmp0, %1 %+ mm_tmp1, tmp2, 1 AVX2_SadBgdSqdiff %1 %+ mm_sad, %1 %+ mm_sum, %1 %+ mm_sumref, %1 %+ mm_mad, %1 %+ mm_sqdiff, %1 %+ mm_sqsum, p_cur + 4 * i_stride, p_ref + 4 * i_stride, %1 %+ mm_zero, %1 %+ mm_tmp0, %1 %+ mm_tmp1, tmp2, 1 AVX2_SadBgdSqdiff %1 %+ mm_sad, %1 %+ mm_sum, %1 %+ mm_sumref, %1 %+ mm_mad, %1 %+ mm_sqdiff, %1 %+ mm_sqsum, p_cur + r_tmp, p_ref + r_tmp, %1 %+ mm_zero, %1 %+ mm_tmp0, %1 %+ mm_tmp1, tmp2, 1 lea r_tmp, [i_stride + 2 * i_stride3] AVX2_SadBgdSqdiff %1 %+ mm_sad, %1 %+ mm_sum, %1 %+ mm_sumref, %1 %+ mm_mad, %1 %+ mm_sqdiff, %1 %+ mm_sqsum, p_cur + 2 * i_stride3, p_ref + 2 * i_stride3, %1 %+ mm_zero, %1 %+ mm_tmp0, %1 %+ mm_tmp1, tmp2, 1 AVX2_SadBgdSqdiff %1 %+ mm_sad, %1 %+ mm_sum, %1 %+ mm_sumref, %1 %+ mm_mad, %1 %+ mm_sqdiff, %1 %+ mm_sqsum, p_cur + r_tmp, p_ref + r_tmp, %1 %+ mm_zero, %1 %+ mm_tmp0, %1 %+ mm_tmp1, tmp2, 1 ; Increment addresses for the next iteration. Doing this early is beneficial on Haswell. add p_cur, %1 %+ mm_width add p_ref, %1 %+ mm_width %else vpxor x %+ mm_sad, x %+ mm_sad, x %+ mm_sad vpxor x %+ mm_sum, x %+ mm_sum, x %+ mm_sum vpxor x %+ mm_sumref, x %+ mm_sumref, x %+ mm_sumref vpxor x %+ mm_mad, x %+ mm_mad, x %+ mm_mad vpxor x %+ mm_sqsum, x %+ mm_sqsum, x %+ mm_sqsum vpxor x %+ mm_sqdiff, x %+ mm_sqdiff, x %+ mm_sqdiff lea r_tmp, [8 * i_stride] add p_cur, r_tmp add p_ref, r_tmp neg r_tmp %%loop: AVX2_SadBgdSqdiff %1 %+ mm_sad, %1 %+ mm_sum, %1 %+ mm_sumref, %1 %+ mm_mad, %1 %+ mm_sqdiff, %1 %+ mm_sqsum, p_cur + r_tmp, p_ref + r_tmp, %1 %+ mm_zero, %1 %+ mm_tmp0, %1 %+ mm_tmp1, tmp2, 1 add r_tmp, i_stride jl %%loop ; Increment addresses for the next iteration. Doing this early is beneficial on Haswell. lea r_tmp, [8 * i_stride - %1 %+ mm_width] sub p_cur, r_tmp sub p_ref, r_tmp %endif mov r_tmp, p_sad8x8 AVX2_Store8x8Accdw r_tmp + 4 * i_xcnt, %1, mm_sad, mm_tmp1, b_second_blocks %ifdef X86_32 vpaddd y %+ mm_tmp1, y %+ mm_sad, sadframe_acc vmovdqa sadframe_acc, y %+ mm_tmp1 %else vpaddd sadframe_acc, sadframe_acc, y %+ mm_sad %endif mov r_tmp, i_xcnt add r_tmp, p_sum16x16 vpunpckhqdq %1 %+ mm_tmp1, %1 %+ mm_sum, %1 %+ mm_sum vpaddd %1 %+ mm_tmp0, %1 %+ mm_sum, %1 %+ mm_tmp1 AVX2_Store16x16Accdw r_tmp, %1, mm_tmp0, mm_tmp1, b_second_blocks mov r_tmp, p_sd8x8 vpsubd %1 %+ mm_sum, %1 %+ mm_sum, %1 %+ mm_sumref AVX2_Store8x8Accdw r_tmp + 4 * i_xcnt, %1, mm_sum, mm_tmp0, b_second_blocks ; Coalesce store and horizontal reduction of MAD accumulator for even and ; odd iterations so as to enable more parallelism. %ifidni %1, y test i_xcnt, 32 / xcnt_unit jz %%preserve_mad mov r_tmp, p_mad8x8 AVX2_Maxubq2 y %+ mm_mad, y %+ mm_mad, prev_mad, y %+ mm_tmp0 AVX2_Store2x8x8Accb r_tmp + i_xcnt - 8, mm_mad, mm_tmp0, mm_tmp1, b_second_blocks %%preserve_mad: vmovdqa prev_mad, y %+ mm_mad %else mov r_tmp, p_mad8x8 AVX2_Maxubq %1 %+ mm_mad, %1 %+ mm_mad, %1 %+ mm_tmp0 AVX2_Store8x8Accb r_tmp + i_xcnt, %1, mm_mad, mm_tmp0, b_second_blocks %endif vpunpcklqdq %1 %+ mm_tmp0, %1 %+ mm_sqsum, %1 %+ mm_sqdiff vpunpckhqdq %1 %+ mm_tmp1, %1 %+ mm_sqsum, %1 %+ mm_sqdiff vpaddd %1 %+ mm_tmp0, %1 %+ mm_tmp0, %1 %+ mm_tmp1 vpshufd %1 %+ mm_tmp1, %1 %+ mm_tmp0, 10110001b vpaddd %1 %+ mm_tmp0, %1 %+ mm_tmp0, %1 %+ mm_tmp1 AVX2_Store2x16x16Accdw p_sqsum16x16, p_sqdiff16x16, i_xcnt, r_tmp, %1, mm_tmp0, mm_tmp1, b_second_blocks %undef mm_tmp0 %undef mm_tmp1 %undef mm_sqsum %undef mm_sqdiff %undef mm_mad %undef mm_sum %undef mm_sumref %undef mm_sad %undef tmp2 %undef b_second_blocks %endmacro ;************************************************************************************************************* ;void VAACalcSadSsdBgd_avx2(const uint8_t *cur_data, const uint8_t *ref_data, int32_t iPicWidth, int32_t iPicHeight, ; int32_t iPicStride, int32_t *psadframe, int32_t *psad8x8, int32_t *psum16x16, int32_t *psqsum16x16, ; int32_t *psqdiff16x16, int32_t *p_sd8x8, uint8_t *p_mad8x8) ;************************************************************************************************************* WELS_EXTERN VAACalcSadSsdBgd_avx2 %define p_sadframe arg6 %define p_sad8x8 arg7 %define p_sum16x16 arg8 %define p_sqsum16x16 arg9 %define p_sqdiff16x16 arg10 %define p_sd8x8 arg11 %define p_mad8x8 arg12 %ifdef X86_32 %define saveregs r5, r6 %else %define saveregs rbx, rbp, r12, r13, r14, r15 %endif %assign push_num 0 LOAD_5_PARA PUSH_XMM 12 SIGN_EXTENSION r2, r2d SIGN_EXTENSION r3, r3d SIGN_EXTENSION r4, r4d PUSHM saveregs %ifdef X86_32 STACK_ALLOC r5, 3 * ymm_width, ymm_width %define mm8 0 %define sadframe_acc_addr r5 %define sadframe_acc [sadframe_acc_addr] %define prev_mad [r5 + ymm_width] %define ymm_zero [r5 + 2 * ymm_width] %define xmm_zero ymm_zero vpxor xmm0, xmm0, xmm0 vmovdqa sadframe_acc, ymm0 vmovdqa ymm_zero, ymm0 %else %define sadframe_acc ymm9 %define xsadframe_acc xmm9 %define prev_mad ymm10 %define ymm_zero ymm11 %define xmm_zero xmm11 vpxor xmm_zero, xmm_zero, xmm_zero vpxor xsadframe_acc, xsadframe_acc, xsadframe_acc %endif and r2, -16 ; iPicWidth &= -16 jle .done ; bail if iPicWidth < 16 sar r3, 4 ; iPicHeight / 16 jle .done ; bail if iPicHeight < 16 shr r2, 2 ; iPicWidth / 4 %define p_cur r0 %define p_ref r1 %define i_xcnt r2 %define i_ycnt ptrword arg4 %define i_stride r4 %define r_tmp r6 %define xcnt_unit 4 %ifdef X86_32 mov i_ycnt, r3 %define i_stride3 r3 %else mov rbp, p_sad8x8 mov r12, p_sum16x16 mov r13, p_sqsum16x16 mov r14, p_sqdiff16x16 mov r15, p_sd8x8 %undef p_sad8x8 %undef p_sum16x16 %undef p_sqsum16x16 %undef p_sqdiff16x16 %undef p_sd8x8 %define p_sad8x8 rbp %define p_sum16x16 r12 %define p_sqsum16x16 r13 %define p_sqdiff16x16 r14 %define p_sd8x8 r15 %define i_stride3 rbx %endif lea i_stride3, [3 * i_stride] ; offset pointers so as to compensate for the i_xcnt offset below. mov r_tmp, i_xcnt and r_tmp, 64 / xcnt_unit - 1 sub p_sum16x16, r_tmp sub p_sqsum16x16, r_tmp sub p_sqdiff16x16, r_tmp sub p_mad8x8, r_tmp shl r_tmp, 2 sub p_sad8x8, r_tmp sub p_sd8x8, r_tmp .height_loop: push i_xcnt %assign push_num push_num + 1 %define i_xcnt_load ptrword [r7] ; use end-of-line pointers so as to enable use of a negative counter as index. lea r_tmp, [xcnt_unit * i_xcnt] add p_sad8x8, r_tmp add p_sum16x16, i_xcnt add p_sqsum16x16, i_xcnt add p_sqdiff16x16, i_xcnt add p_sd8x8, r_tmp add p_mad8x8, i_xcnt and i_xcnt, -(64 / xcnt_unit) jz .width_loop_upper8_64x_end ; use a negative loop counter to enable counting toward zero and indexing with the same counter. neg i_xcnt .width_loop_upper8: AVX2_CalcSadSsdBgd_8Lines y, mm0, mm1, mm2, mm3, mm4, mm5, mm6, mm7, mm8, 0 add i_xcnt, 32 / xcnt_unit jl .width_loop_upper8 jg .width_loop_upper8_32x_end .width_loop_upper8_64x_end: test i_xcnt_load, 32 / xcnt_unit jnz .width_loop_upper8 .width_loop_upper8_32x_end: AVX2_StoreRemainingSingleMad i_xcnt_load, mm1, mm2, 0 test i_xcnt_load, 16 / xcnt_unit jz .width_loop_upper8_end ; remaining 16. AVX2_CalcSadSsdBgd_8Lines x, mm0, mm1, mm2, mm3, mm4, mm5, mm6, mm7, mm8, 0 .width_loop_upper8_end: lea p_cur, [p_cur + 8 * i_stride] lea p_ref, [p_ref + 8 * i_stride] mov i_xcnt, i_xcnt_load lea r_tmp, [xcnt_unit * i_xcnt] sub p_cur, r_tmp sub p_ref, r_tmp and i_xcnt, -(64 / xcnt_unit) jz .width_loop_lower8_64x_end neg i_xcnt .width_loop_lower8: AVX2_CalcSadSsdBgd_8Lines y, mm0, mm1, mm2, mm3, mm4, mm5, mm6, mm7, mm8, 1 add i_xcnt, 32 / xcnt_unit jl .width_loop_lower8 jg .width_loop_lower8_32x_end .width_loop_lower8_64x_end: test i_xcnt_load, 32 / xcnt_unit jnz .width_loop_lower8 .width_loop_lower8_32x_end: AVX2_StoreRemainingSingleMad i_xcnt_load, mm1, mm2, 1 test i_xcnt_load, 16 / xcnt_unit jz .width_loop_lower8_end ; remaining 16. AVX2_CalcSadSsdBgd_8Lines x, mm0, mm1, mm2, mm3, mm4, mm5, mm6, mm7, mm8, 1 .width_loop_lower8_end: lea p_cur, [p_cur + 8 * i_stride] lea p_ref, [p_ref + 8 * i_stride] pop i_xcnt %undef i_xcnt_load %assign push_num push_num - 1 lea r_tmp, [xcnt_unit * i_xcnt] sub p_cur, r_tmp sub p_ref, r_tmp sub i_ycnt, 1 jnz .height_loop .done: mov r_tmp, p_sadframe %ifdef X86_32 vmovdqa xmm2, sadframe_acc vpaddd xmm2, xmm2, [sadframe_acc_addr + xmm_width] %else vextracti128 xmm2, sadframe_acc, 1 vpaddd xmm2, xsadframe_acc, xmm2 %endif vpunpckhqdq xmm1, xmm2, xmm2 vpaddd xmm2, xmm2, xmm1 vmovd [r_tmp], xmm2 vzeroupper %ifdef X86_32 STACK_DEALLOC %endif POPM saveregs POP_XMM LOAD_5_PARA_POP %undef p_cur %undef p_ref %undef i_xcnt %undef i_ycnt %undef i_stride %undef i_stride3 %undef r_tmp %undef xcnt_unit %undef mm8 %undef sadframe_acc %undef sadframe_acc_addr %undef xsadframe_acc %undef prev_mad %undef ymm_zero %undef xmm_zero %undef saveregs %undef p_sadframe %undef p_sad8x8 %undef p_sum16x16 %undef p_sqsum16x16 %undef p_sqdiff16x16 %undef p_sd8x8 %undef p_mad8x8 ret %endif