📄 dsputil_mmx_rnd.h
字号:
}
// avg_pixels
static void av_unused DEF(avg, pixels4)(uint8_t *block, const uint8_t *pixels, int line_size, int h)
{
MOVQ_BFE(mm6);
JUMPALIGN();
do {
__asm __volatile(
"movd %0, %%mm0 \n\t"
"movd %1, %%mm1 \n\t"
PAVGB(%%mm0, %%mm1, %%mm2, %%mm6)
"movd %%mm2, %0 \n\t"
:"+m"(*block)
:"m"(*pixels)
:"memory");
pixels += line_size;
block += line_size;
}
while (--h);
}
// in case more speed is needed - unroling would certainly help
static void DEF(avg, pixels8)(uint8_t *block, const uint8_t *pixels, int line_size, int h)
{
MOVQ_BFE(mm6);
JUMPALIGN();
do {
__asm __volatile(
"movq %0, %%mm0 \n\t"
"movq %1, %%mm1 \n\t"
PAVGB(%%mm0, %%mm1, %%mm2, %%mm6)
"movq %%mm2, %0 \n\t"
:"+m"(*block)
:"m"(*pixels)
:"memory");
pixels += line_size;
block += line_size;
}
while (--h);
}
static void DEF(avg, pixels16)(uint8_t *block, const uint8_t *pixels, int line_size, int h)
{
MOVQ_BFE(mm6);
JUMPALIGN();
do {
__asm __volatile(
"movq %0, %%mm0 \n\t"
"movq %1, %%mm1 \n\t"
PAVGB(%%mm0, %%mm1, %%mm2, %%mm6)
"movq %%mm2, %0 \n\t"
"movq 8%0, %%mm0 \n\t"
"movq 8%1, %%mm1 \n\t"
PAVGB(%%mm0, %%mm1, %%mm2, %%mm6)
"movq %%mm2, 8%0 \n\t"
:"+m"(*block)
:"m"(*pixels)
:"memory");
pixels += line_size;
block += line_size;
}
while (--h);
}
static void DEF(avg, pixels8_x2)(uint8_t *block, const uint8_t *pixels, int line_size, int h)
{
MOVQ_BFE(mm6);
JUMPALIGN();
do {
__asm __volatile(
"movq %1, %%mm0 \n\t"
"movq 1%1, %%mm1 \n\t"
"movq %0, %%mm3 \n\t"
PAVGB(%%mm0, %%mm1, %%mm2, %%mm6)
PAVGB(%%mm3, %%mm2, %%mm0, %%mm6)
"movq %%mm0, %0 \n\t"
:"+m"(*block)
:"m"(*pixels)
:"memory");
pixels += line_size;
block += line_size;
} while (--h);
}
static av_unused void DEF(avg, pixels8_l2)(uint8_t *dst, uint8_t *src1, uint8_t *src2, int dstStride, int src1Stride, int h)
{
MOVQ_BFE(mm6);
JUMPALIGN();
do {
__asm __volatile(
"movq %1, %%mm0 \n\t"
"movq %2, %%mm1 \n\t"
"movq %0, %%mm3 \n\t"
PAVGB(%%mm0, %%mm1, %%mm2, %%mm6)
PAVGB(%%mm3, %%mm2, %%mm0, %%mm6)
"movq %%mm0, %0 \n\t"
:"+m"(*dst)
:"m"(*src1), "m"(*src2)
:"memory");
dst += dstStride;
src1 += src1Stride;
src2 += 8;
} while (--h);
}
static void DEF(avg, pixels16_x2)(uint8_t *block, const uint8_t *pixels, int line_size, int h)
{
MOVQ_BFE(mm6);
JUMPALIGN();
do {
__asm __volatile(
"movq %1, %%mm0 \n\t"
"movq 1%1, %%mm1 \n\t"
"movq %0, %%mm3 \n\t"
PAVGB(%%mm0, %%mm1, %%mm2, %%mm6)
PAVGB(%%mm3, %%mm2, %%mm0, %%mm6)
"movq %%mm0, %0 \n\t"
"movq 8%1, %%mm0 \n\t"
"movq 9%1, %%mm1 \n\t"
"movq 8%0, %%mm3 \n\t"
PAVGB(%%mm0, %%mm1, %%mm2, %%mm6)
PAVGB(%%mm3, %%mm2, %%mm0, %%mm6)
"movq %%mm0, 8%0 \n\t"
:"+m"(*block)
:"m"(*pixels)
:"memory");
pixels += line_size;
block += line_size;
} while (--h);
}
static av_unused void DEF(avg, pixels16_l2)(uint8_t *dst, uint8_t *src1, uint8_t *src2, int dstStride, int src1Stride, int h)
{
MOVQ_BFE(mm6);
JUMPALIGN();
do {
__asm __volatile(
"movq %1, %%mm0 \n\t"
"movq %2, %%mm1 \n\t"
"movq %0, %%mm3 \n\t"
PAVGB(%%mm0, %%mm1, %%mm2, %%mm6)
PAVGB(%%mm3, %%mm2, %%mm0, %%mm6)
"movq %%mm0, %0 \n\t"
"movq 8%1, %%mm0 \n\t"
"movq 8%2, %%mm1 \n\t"
"movq 8%0, %%mm3 \n\t"
PAVGB(%%mm0, %%mm1, %%mm2, %%mm6)
PAVGB(%%mm3, %%mm2, %%mm0, %%mm6)
"movq %%mm0, 8%0 \n\t"
:"+m"(*dst)
:"m"(*src1), "m"(*src2)
:"memory");
dst += dstStride;
src1 += src1Stride;
src2 += 16;
} while (--h);
}
static void DEF(avg, pixels8_y2)(uint8_t *block, const uint8_t *pixels, int line_size, int h)
{
MOVQ_BFE(mm6);
__asm __volatile(
"lea (%3, %3), %%"REG_a" \n\t"
"movq (%1), %%mm0 \n\t"
ASMALIGN(3)
"1: \n\t"
"movq (%1, %3), %%mm1 \n\t"
"movq (%1, %%"REG_a"), %%mm2 \n\t"
PAVGBP(%%mm1, %%mm0, %%mm4, %%mm2, %%mm1, %%mm5)
"movq (%2), %%mm3 \n\t"
PAVGB(%%mm3, %%mm4, %%mm0, %%mm6)
"movq (%2, %3), %%mm3 \n\t"
PAVGB(%%mm3, %%mm5, %%mm1, %%mm6)
"movq %%mm0, (%2) \n\t"
"movq %%mm1, (%2, %3) \n\t"
"add %%"REG_a", %1 \n\t"
"add %%"REG_a", %2 \n\t"
"movq (%1, %3), %%mm1 \n\t"
"movq (%1, %%"REG_a"), %%mm0 \n\t"
PAVGBP(%%mm1, %%mm2, %%mm4, %%mm0, %%mm1, %%mm5)
"movq (%2), %%mm3 \n\t"
PAVGB(%%mm3, %%mm4, %%mm2, %%mm6)
"movq (%2, %3), %%mm3 \n\t"
PAVGB(%%mm3, %%mm5, %%mm1, %%mm6)
"movq %%mm2, (%2) \n\t"
"movq %%mm1, (%2, %3) \n\t"
"add %%"REG_a", %1 \n\t"
"add %%"REG_a", %2 \n\t"
"subl $4, %0 \n\t"
"jnz 1b \n\t"
:"+g"(h), "+S"(pixels), "+D"(block)
:"r"((long)line_size)
:REG_a, "memory");
}
// this routine is 'slightly' suboptimal but mostly unused
static void DEF(avg, pixels8_xy2)(uint8_t *block, const uint8_t *pixels, int line_size, int h)
{
MOVQ_ZERO(mm7);
SET_RND(mm6); // =2 for rnd and =1 for no_rnd version
__asm __volatile(
"movq (%1), %%mm0 \n\t"
"movq 1(%1), %%mm4 \n\t"
"movq %%mm0, %%mm1 \n\t"
"movq %%mm4, %%mm5 \n\t"
"punpcklbw %%mm7, %%mm0 \n\t"
"punpcklbw %%mm7, %%mm4 \n\t"
"punpckhbw %%mm7, %%mm1 \n\t"
"punpckhbw %%mm7, %%mm5 \n\t"
"paddusw %%mm0, %%mm4 \n\t"
"paddusw %%mm1, %%mm5 \n\t"
"xor %%"REG_a", %%"REG_a" \n\t"
"add %3, %1 \n\t"
ASMALIGN(3)
"1: \n\t"
"movq (%1, %%"REG_a"), %%mm0 \n\t"
"movq 1(%1, %%"REG_a"), %%mm2 \n\t"
"movq %%mm0, %%mm1 \n\t"
"movq %%mm2, %%mm3 \n\t"
"punpcklbw %%mm7, %%mm0 \n\t"
"punpcklbw %%mm7, %%mm2 \n\t"
"punpckhbw %%mm7, %%mm1 \n\t"
"punpckhbw %%mm7, %%mm3 \n\t"
"paddusw %%mm2, %%mm0 \n\t"
"paddusw %%mm3, %%mm1 \n\t"
"paddusw %%mm6, %%mm4 \n\t"
"paddusw %%mm6, %%mm5 \n\t"
"paddusw %%mm0, %%mm4 \n\t"
"paddusw %%mm1, %%mm5 \n\t"
"psrlw $2, %%mm4 \n\t"
"psrlw $2, %%mm5 \n\t"
"movq (%2, %%"REG_a"), %%mm3 \n\t"
"packuswb %%mm5, %%mm4 \n\t"
"pcmpeqd %%mm2, %%mm2 \n\t"
"paddb %%mm2, %%mm2 \n\t"
PAVGB(%%mm3, %%mm4, %%mm5, %%mm2)
"movq %%mm5, (%2, %%"REG_a") \n\t"
"add %3, %%"REG_a" \n\t"
"movq (%1, %%"REG_a"), %%mm2 \n\t" // 0 <-> 2 1 <-> 3
"movq 1(%1, %%"REG_a"), %%mm4 \n\t"
"movq %%mm2, %%mm3 \n\t"
"movq %%mm4, %%mm5 \n\t"
"punpcklbw %%mm7, %%mm2 \n\t"
"punpcklbw %%mm7, %%mm4 \n\t"
"punpckhbw %%mm7, %%mm3 \n\t"
"punpckhbw %%mm7, %%mm5 \n\t"
"paddusw %%mm2, %%mm4 \n\t"
"paddusw %%mm3, %%mm5 \n\t"
"paddusw %%mm6, %%mm0 \n\t"
"paddusw %%mm6, %%mm1 \n\t"
"paddusw %%mm4, %%mm0 \n\t"
"paddusw %%mm5, %%mm1 \n\t"
"psrlw $2, %%mm0 \n\t"
"psrlw $2, %%mm1 \n\t"
"movq (%2, %%"REG_a"), %%mm3 \n\t"
"packuswb %%mm1, %%mm0 \n\t"
"pcmpeqd %%mm2, %%mm2 \n\t"
"paddb %%mm2, %%mm2 \n\t"
PAVGB(%%mm3, %%mm0, %%mm1, %%mm2)
"movq %%mm1, (%2, %%"REG_a") \n\t"
"add %3, %%"REG_a" \n\t"
"subl $2, %0 \n\t"
"jnz 1b \n\t"
:"+g"(h), "+S"(pixels)
:"D"(block), "r"((long)line_size)
:REG_a, "memory");
}
//FIXME optimize
static void DEF(put, pixels16_y2)(uint8_t *block, const uint8_t *pixels, int line_size, int h){
DEF(put, pixels8_y2)(block , pixels , line_size, h);
DEF(put, pixels8_y2)(block+8, pixels+8, line_size, h);
}
static void DEF(put, pixels16_xy2)(uint8_t *block, const uint8_t *pixels, int line_size, int h){
DEF(put, pixels8_xy2)(block , pixels , line_size, h);
DEF(put, pixels8_xy2)(block+8, pixels+8, line_size, h);
}
static void DEF(avg, pixels16_y2)(uint8_t *block, const uint8_t *pixels, int line_size, int h){
DEF(avg, pixels8_y2)(block , pixels , line_size, h);
DEF(avg, pixels8_y2)(block+8, pixels+8, line_size, h);
}
static void DEF(avg, pixels16_xy2)(uint8_t *block, const uint8_t *pixels, int line_size, int h){
DEF(avg, pixels8_xy2)(block , pixels , line_size, h);
DEF(avg, pixels8_xy2)(block+8, pixels+8, line_size, h);
}
⌨️ 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 + -