mirror of
https://github.com/FFmpeg/FFmpeg.git
synced 2024-12-28 20:53:54 +02:00
avcodec/h264: add avx 8-bit 4:2:2 chroma h intra deblock/loop filter
~1.37x faster (147 vs. 108 cycles) compared to mmxext function
This commit is contained in:
parent
0e16b3e2be
commit
cd893b9307
@ -1232,6 +1232,24 @@ cglobal deblock_h_chroma_intra_8, 4, 6, 8, pix_, stride_, alpha_, beta_
|
||||
STORE_8_ROWS PASS8ROWS(pix_q - 2, r4 - 2, stride_q, r5)
|
||||
RET
|
||||
|
||||
cglobal deblock_h_chroma422_intra_8, 4, 6, 8, pix_, stride_, alpha_, beta_
|
||||
CHROMA_H_START_XMM r4, r5
|
||||
LOAD_8_ROWS PASS8ROWS(pix_q - 2, r4 - 2, stride_q, r5)
|
||||
TRANSPOSE_8x4B_XMM
|
||||
CHROMA_INTRA_BODY_XMM
|
||||
TRANSPOSE_4x8B_XMM
|
||||
STORE_8_ROWS PASS8ROWS(pix_q - 2, r4 - 2, stride_q, r5)
|
||||
|
||||
lea pix_q, [pix_q + 8*stride_q]
|
||||
lea r4, [r4 + 8*stride_q]
|
||||
|
||||
LOAD_8_ROWS PASS8ROWS(pix_q - 2, r4 - 2, stride_q, r5)
|
||||
TRANSPOSE_8x4B_XMM
|
||||
CHROMA_INTRA_BODY_XMM
|
||||
TRANSPOSE_4x8B_XMM
|
||||
STORE_8_ROWS PASS8ROWS(pix_q - 2, r4 - 2, stride_q, r5)
|
||||
RET
|
||||
|
||||
%endmacro ; DEBLOCK_CHROMA_XMM
|
||||
|
||||
DEBLOCK_CHROMA_XMM avx
|
||||
|
@ -325,6 +325,7 @@ av_cold void ff_h264dsp_init_x86(H264DSPContext *c, const int bit_depth,
|
||||
c->h264_h_loop_filter_chroma_intra = ff_deblock_h_chroma_intra_8_avx;
|
||||
} else {
|
||||
c->h264_h_loop_filter_chroma = ff_deblock_h_chroma422_8_avx;
|
||||
c->h264_h_loop_filter_chroma_intra = ff_deblock_h_chroma422_intra_8_avx;
|
||||
}
|
||||
}
|
||||
} else if (bit_depth == 10) {
|
||||
|
Loading…
Reference in New Issue
Block a user