Commit 987ffe4b authored by James Darnley's avatar James Darnley

avcodec/h264: add avx 8-bit chroma v intra deblock/loop filter

~1.14x faster (90 vs 78 cycles) compared with mmxext
parent 88307b3e
......@@ -1117,6 +1117,28 @@ ff_chroma_intra_body_mmxext:
DEBLOCK_P0_Q0
%endmacro
%macro CHROMA_INTRA_BODY_XMM 0
LOAD_MASK alpha_d, beta_d
mova m5, m1
mova m6, m2
pxor m4, m1, m3
pand m4, [pb_1]
pavgb m1, m3
psubusb m1, m4
pavgb m1, m0
pxor m4, m2, m0
pand m4, [pb_1]
pavgb m2, m0
psubusb m2, m4
pavgb m2, m3
psubb m1, m5
psubb m2, m6
pand m1, m7
pand m2, m7
paddb m1, m5
paddb m2, m6
%endmacro
%macro CHROMA_V_START_XMM 1
movsxdifnidn stride_q, stride_d
dec alpha_d
......@@ -1190,6 +1212,17 @@ cglobal deblock_h_chroma422_8, 5, 7, 8, 0-16, pix_, stride_, alpha_, beta_, tc0_
STORE_8_ROWS PASS8ROWS(pix_q - 2, r5 - 2, stride_q, r6)
RET
cglobal deblock_v_chroma_intra_8, 4, 5, 8, pix_, stride_, alpha_, beta_
CHROMA_V_START_XMM r4
movq m0, [r4]
movq m1, [r4 + stride_q]
movq m2, [pix_q]
movq m3, [pix_q + stride_q]
CHROMA_INTRA_BODY_XMM
movq [r4 + stride_q], m1
movq [pix_q], m2
RET
%endmacro ; DEBLOCK_CHROMA_XMM
DEBLOCK_CHROMA_XMM avx
......
......@@ -319,6 +319,7 @@ av_cold void ff_h264dsp_init_x86(H264DSPContext *c, const int bit_depth,
#endif
c->h264_v_loop_filter_chroma = ff_deblock_v_chroma_8_avx;
c->h264_v_loop_filter_chroma_intra = ff_deblock_v_chroma_intra_8_avx;
if (chroma_format_idc <= 1) {
c->h264_h_loop_filter_chroma = ff_deblock_h_chroma_8_avx;
} else {
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment