📄 h264dsp_mmx.c.svn-base
字号:
"pmullw %%mm6, %%mm1 \n\t"\ "movd -2(%0), %%mm2 \n\t"\ "movd 7(%0), %%mm5 \n\t"\ "punpcklbw %%mm7, %%mm2 \n\t"\ "punpcklbw %%mm7, %%mm5 \n\t"\ "paddw %%mm3, %%mm2 \n\t"\ "paddw %%mm5, %%mm4 \n\t"\ "movq %6, %%mm5 \n\t"\ "paddw %%mm5, %%mm2 \n\t"\ "paddw %%mm5, %%mm4 \n\t"\ "paddw %%mm2, %%mm0 \n\t"\ "paddw %%mm4, %%mm1 \n\t"\ "psraw $5, %%mm0 \n\t"\ "psraw $5, %%mm1 \n\t"\ "packuswb %%mm1, %%mm0 \n\t"\ OP(%%mm0, (%1),%%mm5, q)\ "add %3, %0 \n\t"\ "add %4, %1 \n\t"\ "decl %2 \n\t"\ " jnz 1b \n\t"\ : "+a"(src), "+c"(dst), "+g"(h)\ : "d"((long)srcStride), "S"((long)dstStride), "m"(ff_pw_5), "m"(ff_pw_16)\ : "memory"\ );\}\\static av_noinline void OPNAME ## h264_qpel8_h_lowpass_l2_ ## MMX(uint8_t *dst, uint8_t *src, uint8_t *src2, int dstStride, int src2Stride){\ int h=8;\ asm volatile(\ "pxor %%mm7, %%mm7 \n\t"\ "movq %0, %%mm6 \n\t"\ :: "m"(ff_pw_5)\ );\ do{\ asm volatile(\ "movq (%0), %%mm0 \n\t"\ "movq 1(%0), %%mm2 \n\t"\ "movq %%mm0, %%mm1 \n\t"\ "movq %%mm2, %%mm3 \n\t"\ "punpcklbw %%mm7, %%mm0 \n\t"\ "punpckhbw %%mm7, %%mm1 \n\t"\ "punpcklbw %%mm7, %%mm2 \n\t"\ "punpckhbw %%mm7, %%mm3 \n\t"\ "paddw %%mm2, %%mm0 \n\t"\ "paddw %%mm3, %%mm1 \n\t"\ "psllw $2, %%mm0 \n\t"\ "psllw $2, %%mm1 \n\t"\ "movq -1(%0), %%mm2 \n\t"\ "movq 2(%0), %%mm4 \n\t"\ "movq %%mm2, %%mm3 \n\t"\ "movq %%mm4, %%mm5 \n\t"\ "punpcklbw %%mm7, %%mm2 \n\t"\ "punpckhbw %%mm7, %%mm3 \n\t"\ "punpcklbw %%mm7, %%mm4 \n\t"\ "punpckhbw %%mm7, %%mm5 \n\t"\ "paddw %%mm4, %%mm2 \n\t"\ "paddw %%mm3, %%mm5 \n\t"\ "psubw %%mm2, %%mm0 \n\t"\ "psubw %%mm5, %%mm1 \n\t"\ "pmullw %%mm6, %%mm0 \n\t"\ "pmullw %%mm6, %%mm1 \n\t"\ "movd -2(%0), %%mm2 \n\t"\ "movd 7(%0), %%mm5 \n\t"\ "punpcklbw %%mm7, %%mm2 \n\t"\ "punpcklbw %%mm7, %%mm5 \n\t"\ "paddw %%mm3, %%mm2 \n\t"\ "paddw %%mm5, %%mm4 \n\t"\ "movq %5, %%mm5 \n\t"\ "paddw %%mm5, %%mm2 \n\t"\ "paddw %%mm5, %%mm4 \n\t"\ "paddw %%mm2, %%mm0 \n\t"\ "paddw %%mm4, %%mm1 \n\t"\ "psraw $5, %%mm0 \n\t"\ "psraw $5, %%mm1 \n\t"\ "movq (%2), %%mm4 \n\t"\ "packuswb %%mm1, %%mm0 \n\t"\ PAVGB" %%mm4, %%mm0 \n\t"\ OP(%%mm0, (%1),%%mm5, q)\ "add %4, %0 \n\t"\ "add %4, %1 \n\t"\ "add %3, %2 \n\t"\ : "+a"(src), "+c"(dst), "+d"(src2)\ : "D"((long)src2Stride), "S"((long)dstStride),\ "m"(ff_pw_16)\ : "memory"\ );\ }while(--h);\}\\static av_noinline void OPNAME ## h264_qpel8or16_v_lowpass_ ## MMX(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, int h){\ int w= 2;\ src -= 2*srcStride;\ \ while(w--){\ asm volatile(\ "pxor %%mm7, %%mm7 \n\t"\ "movd (%0), %%mm0 \n\t"\ "add %2, %0 \n\t"\ "movd (%0), %%mm1 \n\t"\ "add %2, %0 \n\t"\ "movd (%0), %%mm2 \n\t"\ "add %2, %0 \n\t"\ "movd (%0), %%mm3 \n\t"\ "add %2, %0 \n\t"\ "movd (%0), %%mm4 \n\t"\ "add %2, %0 \n\t"\ "punpcklbw %%mm7, %%mm0 \n\t"\ "punpcklbw %%mm7, %%mm1 \n\t"\ "punpcklbw %%mm7, %%mm2 \n\t"\ "punpcklbw %%mm7, %%mm3 \n\t"\ "punpcklbw %%mm7, %%mm4 \n\t"\ QPEL_H264V(%%mm0, %%mm1, %%mm2, %%mm3, %%mm4, %%mm5, OP)\ QPEL_H264V(%%mm1, %%mm2, %%mm3, %%mm4, %%mm5, %%mm0, OP)\ QPEL_H264V(%%mm2, %%mm3, %%mm4, %%mm5, %%mm0, %%mm1, OP)\ QPEL_H264V(%%mm3, %%mm4, %%mm5, %%mm0, %%mm1, %%mm2, OP)\ QPEL_H264V(%%mm4, %%mm5, %%mm0, %%mm1, %%mm2, %%mm3, OP)\ QPEL_H264V(%%mm5, %%mm0, %%mm1, %%mm2, %%mm3, %%mm4, OP)\ QPEL_H264V(%%mm0, %%mm1, %%mm2, %%mm3, %%mm4, %%mm5, OP)\ QPEL_H264V(%%mm1, %%mm2, %%mm3, %%mm4, %%mm5, %%mm0, OP)\ \ : "+a"(src), "+c"(dst)\ : "S"((long)srcStride), "D"((long)dstStride), "m"(ff_pw_5), "m"(ff_pw_16)\ : "memory"\ );\ if(h==16){\ asm volatile(\ QPEL_H264V(%%mm2, %%mm3, %%mm4, %%mm5, %%mm0, %%mm1, OP)\ QPEL_H264V(%%mm3, %%mm4, %%mm5, %%mm0, %%mm1, %%mm2, OP)\ QPEL_H264V(%%mm4, %%mm5, %%mm0, %%mm1, %%mm2, %%mm3, OP)\ QPEL_H264V(%%mm5, %%mm0, %%mm1, %%mm2, %%mm3, %%mm4, OP)\ QPEL_H264V(%%mm0, %%mm1, %%mm2, %%mm3, %%mm4, %%mm5, OP)\ QPEL_H264V(%%mm1, %%mm2, %%mm3, %%mm4, %%mm5, %%mm0, OP)\ QPEL_H264V(%%mm2, %%mm3, %%mm4, %%mm5, %%mm0, %%mm1, OP)\ QPEL_H264V(%%mm3, %%mm4, %%mm5, %%mm0, %%mm1, %%mm2, OP)\ \ : "+a"(src), "+c"(dst)\ : "S"((long)srcStride), "D"((long)dstStride), "m"(ff_pw_5), "m"(ff_pw_16)\ : "memory"\ );\ }\ src += 4-(h+5)*srcStride;\ dst += 4-h*dstStride;\ }\}\static av_always_inline void OPNAME ## h264_qpel8or16_hv1_lowpass_ ## MMX(int16_t *tmp, uint8_t *src, int tmpStride, int srcStride, int size){\ int w = (size+8)>>2;\ src -= 2*srcStride+2;\ while(w--){\ asm volatile(\ "pxor %%mm7, %%mm7 \n\t"\ "movd (%0), %%mm0 \n\t"\ "add %2, %0 \n\t"\ "movd (%0), %%mm1 \n\t"\ "add %2, %0 \n\t"\ "movd (%0), %%mm2 \n\t"\ "add %2, %0 \n\t"\ "movd (%0), %%mm3 \n\t"\ "add %2, %0 \n\t"\ "movd (%0), %%mm4 \n\t"\ "add %2, %0 \n\t"\ "punpcklbw %%mm7, %%mm0 \n\t"\ "punpcklbw %%mm7, %%mm1 \n\t"\ "punpcklbw %%mm7, %%mm2 \n\t"\ "punpcklbw %%mm7, %%mm3 \n\t"\ "punpcklbw %%mm7, %%mm4 \n\t"\ QPEL_H264HV(%%mm0, %%mm1, %%mm2, %%mm3, %%mm4, %%mm5, 0*48)\ QPEL_H264HV(%%mm1, %%mm2, %%mm3, %%mm4, %%mm5, %%mm0, 1*48)\ QPEL_H264HV(%%mm2, %%mm3, %%mm4, %%mm5, %%mm0, %%mm1, 2*48)\ QPEL_H264HV(%%mm3, %%mm4, %%mm5, %%mm0, %%mm1, %%mm2, 3*48)\ QPEL_H264HV(%%mm4, %%mm5, %%mm0, %%mm1, %%mm2, %%mm3, 4*48)\ QPEL_H264HV(%%mm5, %%mm0, %%mm1, %%mm2, %%mm3, %%mm4, 5*48)\ QPEL_H264HV(%%mm0, %%mm1, %%mm2, %%mm3, %%mm4, %%mm5, 6*48)\ QPEL_H264HV(%%mm1, %%mm2, %%mm3, %%mm4, %%mm5, %%mm0, 7*48)\ : "+a"(src)\ : "c"(tmp), "S"((long)srcStride), "m"(ff_pw_5), "m"(ff_pw_16)\ : "memory"\ );\ if(size==16){\ asm volatile(\ QPEL_H264HV(%%mm2, %%mm3, %%mm4, %%mm5, %%mm0, %%mm1, 8*48)\ QPEL_H264HV(%%mm3, %%mm4, %%mm5, %%mm0, %%mm1, %%mm2, 9*48)\ QPEL_H264HV(%%mm4, %%mm5, %%mm0, %%mm1, %%mm2, %%mm3, 10*48)\ QPEL_H264HV(%%mm5, %%mm0, %%mm1, %%mm2, %%mm3, %%mm4, 11*48)\ QPEL_H264HV(%%mm0, %%mm1, %%mm2, %%mm3, %%mm4, %%mm5, 12*48)\ QPEL_H264HV(%%mm1, %%mm2, %%mm3, %%mm4, %%mm5, %%mm0, 13*48)\ QPEL_H264HV(%%mm2, %%mm3, %%mm4, %%mm5, %%mm0, %%mm1, 14*48)\ QPEL_H264HV(%%mm3, %%mm4, %%mm5, %%mm0, %%mm1, %%mm2, 15*48)\ : "+a"(src)\ : "c"(tmp), "S"((long)srcStride), "m"(ff_pw_5), "m"(ff_pw_16)\ : "memory"\ );\ }\ tmp += 4;\ src += 4 - (size+5)*srcStride;\ }\}\static av_always_inline void OPNAME ## h264_qpel8or16_hv2_lowpass_ ## MMX(uint8_t *dst, int16_t *tmp, int dstStride, int tmpStride, int size){\ int w = size>>4;\ do{\ int h = size;\ asm volatile(\ "1: \n\t"\ "movq (%0), %%mm0 \n\t"\ "movq 8(%0), %%mm3 \n\t"\ "movq 2(%0), %%mm1 \n\t"\ "movq 10(%0), %%mm4 \n\t"\ "paddw %%mm4, %%mm0 \n\t"\ "paddw %%mm3, %%mm1 \n\t"\ "paddw 18(%0), %%mm3 \n\t"\ "paddw 16(%0), %%mm4 \n\t"\ "movq 4(%0), %%mm2 \n\t"\ "movq 12(%0), %%mm5 \n\t"\ "paddw 6(%0), %%mm2 \n\t"\ "paddw 14(%0), %%mm5 \n\t"\ "psubw %%mm1, %%mm0 \n\t"\ "psubw %%mm4, %%mm3 \n\t"\ "psraw $2, %%mm0 \n\t"\ "psraw $2, %%mm3 \n\t"\ "psubw %%mm1, %%mm0 \n\t"\ "psubw %%mm4, %%mm3 \n\t"\ "paddsw %%mm2, %%mm0 \n\t"\ "paddsw %%mm5, %%mm3 \n\t"\ "psraw $2, %%mm0 \n\t"\ "psraw $2, %%mm3 \n\t"\ "paddw %%mm2, %%mm0 \n\t"\ "paddw %%mm5, %%mm3 \n\t"\ "psraw $6, %%mm0 \n\t"\ "psraw $6, %%mm3 \n\t"\ "packuswb %%mm3, %%mm0 \n\t"\ OP(%%mm0, (%1),%%mm7, q)\ "add $48, %0 \n\t"\ "add %3, %1 \n\t"\ "decl %2 \n\t"\ " jnz 1b \n\t"\ : "+a"(tmp), "+c"(dst), "+g"(h)\ : "S"((long)dstStride)\ : "memory"\ );\ tmp += 8 - size*24;\ dst += 8 - size*dstStride;\ }while(w--);\}\\static void OPNAME ## h264_qpel8_v_lowpass_ ## MMX(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\ OPNAME ## h264_qpel8or16_v_lowpass_ ## MMX(dst , src , dstStride, srcStride, 8);\}\static av_noinline void OPNAME ## h264_qpel16_v_lowpass_ ## MMX(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\ OPNAME ## h264_qpel8or16_v_lowpass_ ## MMX(dst , src , dstStride, srcStride, 16);\ OPNAME ## h264_qpel8or16_v_lowpass_ ## MMX(dst+8, src+8, dstStride, srcStride, 16);\}\\static void OPNAME ## h264_qpel16_h_lowpass_ ## MMX(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\ OPNAME ## h264_qpel8_h_lowpass_ ## MMX(dst , src , dstStride, srcStride);\ OPNAME ## h264_qpel8_h_lowpass_ ## MMX(dst+8, src+8, dstStride, srcStride);\ src += 8*srcStride;\ dst += 8*dstStride;\ OPNAME ## h264_qpel8_h_lowpass_ ## MMX(dst , src , dstStride, srcStride);\ OPNAME ## h264_qpel8_h_lowpass_ ## MMX(dst+8, src+8, dstStride, srcStride);\}\\static av_noinline void OPNAME ## h264_qpel16_h_lowpass_l2_ ## MMX(uint8_t *dst, uint8_t *src, uint8_t *src2, int dstStride, int src2Stride){\ OPNAME ## h264_qpel8_h_lowpass_l2_ ## MMX(dst , src , src2 , dstStride, src2Stride);\ OPNAME ## h264_qpel8_h_lowpass_l2_ ## MMX(dst+8, src+8, src2+8, dstStride, src2Stride);\ src += 8*dstStride;\ dst += 8*dstStride;\ src2 += 8*src2Stride;\ OPNAME ## h264_qpel8_h_lowpass_l2_ ## MMX(dst , src , src2 , dstStride, src2Stride);\ OPNAME ## h264_qpel8_h_lowpass_l2_ ## MMX(dst+8, src+8, src2+8, dstStride, src2Stride);\}\\static av_noinline void OPNAME ## h264_qpel8or16_hv_lowpass_ ## MMX(uint8_t *dst, int16_t *tmp, uint8_t *src, int dstStride, int tmpStride, int srcStride, int size){\ put_h264_qpel8or16_hv1_lowpass_ ## MMX(tmp, src, tmpStride, srcStride, size);\ OPNAME ## h264_qpel8or16_hv2_lowpass_ ## MMX(dst, tmp, dstStride, tmpStride, size);\}\static void OPNAME ## h264_qpel8_hv_lowpass_ ## MMX(uint8_t *dst, int16_t *tmp, uint8_t *src, int dstStride, int tmpStride, int srcStride){\ OPNAME ## h264_qpel8or16_hv_lowpass_ ## MMX(dst , tmp , src , dstStride, tmpStride, srcStride, 8);\}\\static void OPNAME ## h264_qpel16_hv_lowpass_ ## MMX(uint8_t *dst, int16_t *tmp, uint8_t *src, int dstStride, int tmpStride, int srcStride){\ OPNAME ## h264_qpel8or16_hv_lowpass_ ## MMX(dst , tmp , src , dstStride, tmpStride, srcStride, 16);\}\\static av_noinline void OPNAME ## pixels4_l2_shift5_ ## MMX(uint8_t *dst, int16_t *src16, uint8_t *src8, int dstStride, int src8Stride, int h)\{\ asm volatile(\ "movq (%1), %%mm0 \n\t"\ "movq 24(%1), %%mm1 \n\t"\ "psraw $5, %%mm0 \n\t"\ "psraw $5, %%mm1 \n\t"\ "packuswb %%mm0, %%mm0 \n\t"\ "packuswb %%mm1, %%mm1 \n\t"\ PAVGB" (%0), %%mm0 \n\t"\ PAVGB" (%0,%3), %%mm1 \n\t"\ OP(%%mm0, (%2), %%mm4, d)\ OP(%%mm1, (%2,%4), %%mm5, d)\ "lea (%0,%3,2), %0 \n\t"\ "lea (%2,%4,2), %2 \n\t"\ "movq 48(%1), %%mm0 \n\t"\ "movq 72(%1), %%mm1 \n\t"\ "psraw $5, %%mm0 \n\t"\ "psraw $5, %%mm1 \n\t"\ "packuswb %%mm0, %%mm0 \n\t"\ "packuswb %%mm1, %%mm1 \n\t"\ PAVGB" (%0), %%mm0 \n\t"\ PAVGB" (%0,%3), %%mm1 \n\t"\ OP(%%mm0, (%2), %%mm4, d)\ OP(%%mm1, (%2,%4), %%mm5, d)\ :"+a"(src8), "+c"(src16), "+d"(dst)\ :"S"((long)src8Stride), "D"((long)dstStride)\ :"memory");\}\static av_noinline void OPNAME ## pixels8_l2_shift5_ ## MMX(uint8_t *dst, int16_t *src16, uint8_t *src8, int dstStride, int src8Stride, int h)\{\ do{\ asm volatile(\ "movq (%1), %%mm0 \n\t"\ "movq 8(%1), %%mm1 \n\t"\ "movq 48(%1), %%mm2 \n\t"\ "movq 8+48(%1), %%mm3 \n\t"\ "psraw $5, %%mm0 \n\t"\ "psraw $5, %%mm1 \n\t"\ "psraw $5, %%mm2 \n\t"\ "psraw $5, %%mm3 \n\t"\⌨️ 快捷键说明
复制代码
Ctrl + C
搜索代码
Ctrl + F
全屏模式
F11
切换主题
Ctrl + Shift + D
显示快捷键
?
增大字号
Ctrl + =
减小字号
Ctrl + -