1
0
mirror of https://github.com/FFmpeg/FFmpeg.git synced 2024-11-26 19:01:44 +02:00

dnxhdenc: Optimize get_pixels_8x4_sym for 10-bit

This reverts commit 628e6d0164 and uses
a better fix.

Before:
4483 decicycles in get_pixels_8x4_sym,  131032 runs,     40 skips

After:
2569 decicycles in get_pixels_8x4_sym,  131054 runs,     18 skips
This commit is contained in:
Timothy Gu 2015-09-29 16:50:02 -07:00
parent 00ae5b401b
commit bc22cd244e

View File

@ -87,22 +87,14 @@ void dnxhd_10bit_get_pixels_8x4_sym(int16_t *av_restrict block,
const uint8_t *pixels, const uint8_t *pixels,
ptrdiff_t line_size) ptrdiff_t line_size)
{ {
int i; memcpy(block + 0 * 8, pixels + 0 * line_size, 8 * sizeof(*block));
const uint16_t* pixels16 = (const uint16_t*)pixels; memcpy(block + 7 * 8, pixels + 0 * line_size, 8 * sizeof(*block));
line_size >>= 1; memcpy(block + 1 * 8, pixels + 1 * line_size, 8 * sizeof(*block));
memcpy(block + 6 * 8, pixels + 1 * line_size, 8 * sizeof(*block));
for (i = 0; i < 4; i++) { memcpy(block + 2 * 8, pixels + 2 * line_size, 8 * sizeof(*block));
block[0] = pixels16[0]; block[1] = pixels16[1]; memcpy(block + 5 * 8, pixels + 2 * line_size, 8 * sizeof(*block));
block[2] = pixels16[2]; block[3] = pixels16[3]; memcpy(block + 3 * 8, pixels + 3 * line_size, 8 * sizeof(*block));
block[4] = pixels16[4]; block[5] = pixels16[5]; memcpy(block + 4 * 8, pixels + 3 * line_size, 8 * sizeof(*block));
block[6] = pixels16[6]; block[7] = pixels16[7];
pixels16 += line_size;
block += 8;
}
memcpy(block, block - 8, sizeof(*block) * 8);
memcpy(block + 8, block - 16, sizeof(*block) * 8);
memcpy(block + 16, block - 24, sizeof(*block) * 8);
memcpy(block + 24, block - 32, sizeof(*block) * 8);
} }
static int dnxhd_10bit_dct_quantize(MpegEncContext *ctx, int16_t *block, static int dnxhd_10bit_dct_quantize(MpegEncContext *ctx, int16_t *block,