virtualenv/FFmpeg
1
0
Fork 0
mirror of https://github.com/FFmpeg/FFmpeg.git synced 2025-08-15 14:13:16 +02:00
Code Issues Releases Activity

hevc: cleaning up SSE2 and SSSE3 deblocking filters

Browse Source 
Signed-off-by: Mickaël Raulet <mraulet@insa-rennes.fr>

cherry picked from commit b432041d7d1eca38831590f13b4e5baffff8186f
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
This commit is contained in:
Anton Khirnov
2014-07-19 13:53:56 +02:00
committed by Michael Niedermayer
parent 5f08b417b4
commit 7a4cf67117
1 changed files with 50 additions and 44 deletions
Download Patch File Download Diff File Expand all files Collapse all files

94
libavcodec/x86/hevc_deblock.asm
View File

@@ -329,8 +329,10 @@ ALIGN 16
paddw m5, [pw_4]; +4 paddw m5, [pw_4]; +4
psraw m5, 3; >> 3 psraw m5, 3; >> 3
%if %1 > 8
psllw m4, %1-8; << (BIT_DEPTH - 8) psllw m4, %1-8; << (BIT_DEPTH - 8)
psllw m6, %1-8; << (BIT_DEPTH - 8) psllw m6, %1-8; << (BIT_DEPTH - 8)
%endif
pmaxsw m5, m4 pmaxsw m5, m4
pminsw m5, m6 pminsw m5, m6
paddw m1, m5; p0 + delta0 paddw m1, m5; p0 + delta0
@@ -376,7 +378,7 @@ ALIGN 16
pcmpgtw m15, m13, m14; beta0, beta1 pcmpgtw m15, m13, m14; beta0, beta1
movmskps r13, m15 ;filtering mask 0d0 + 0d3 < beta0 (bit 2 or 3) , 1d0 + 1d3 < beta1 (bit 0 or 1) movmskps r13, m15 ;filtering mask 0d0 + 0d3 < beta0 (bit 2 or 3) , 1d0 + 1d3 < beta1 (bit 0 or 1)
cmp r13, 0 cmp r13, 0
je .bypassluma je .bypassluma
;weak / strong decision compare to beta_2 ;weak / strong decision compare to beta_2
psraw m15, m13, 2; beta >> 2 psraw m15, m13, 2; beta >> 2
@@ -420,14 +422,18 @@ ALIGN 16
;decide between strong and weak filtering ;decide between strong and weak filtering
;tc25 calculations ;tc25 calculations
mov r2d, [tcq]; mov r2d, [tcq];
%if %1 > 8
shl r2, %1 - 8 shl r2, %1 - 8
%endif
movd m8, r2d; tc0 movd m8, r2d; tc0
add tcq, 4; add tcq, 4;
mov r3d, [tcq]; mov r3d, [tcq];
%if %1 > 8
shl r3, %1 - 8 shl r3, %1 - 8
%endif
movd m9, r3d; tc1 movd m9, r3d; tc1
add r2d, r3d; tc0 + tc1 add r2d, r3d; tc0 + tc1
jz .bypassluma jz .bypassluma
punpcklwd m8, m8 punpcklwd m8, m8
punpcklwd m9, m9 punpcklwd m9, m9
shufps m8, m9, 0; tc0, tc1 shufps m8, m9, 0; tc0, tc1
@@ -549,7 +555,7 @@ ALIGN 16
paddw m13, m7; q3 + 2 paddw m13, m7; q3 + 2
paddw m13, m6; q3 + q2 + 2 paddw m13, m6; q3 + q2 + 2
psllw m13, 1; 2*q3 + 2*q2 + 4 psllw m13, 1; 2*q3 + 2*q2 + 4
paddw m13, m6; 2*q3 + 3*q2 + 4 paddw m13, m6; 2*q3 + 3*q2 + 4
paddw m13, m10; 2*q3 + 3*q2 + q1 + q0 + p0 + 4 paddw m13, m10; 2*q3 + 3*q2 + q1 + q0 + p0 + 4
psraw m13, 3; (2*q3 + 3*q2 + q1 + q0 + p0 + 4) >> 3 psraw m13, 3; (2*q3 + 3*q2 + q1 + q0 + p0 + 4) >> 3
psubw m13, m6; ((2*q3 + 3*q2 + q1 + q0 + p0 + 4) >> 3) - q2 psubw m13, m6; ((2*q3 + 3*q2 + q1 + q0 + p0 + 4) >> 3) - q2
@@ -565,7 +571,7 @@ ALIGN 16
.weakfilter: .weakfilter:
not r14; strong mask -> weak mask not r14; strong mask -> weak mask
and r14, r13; final weak filtering mask, bits 0 and 1 and r14, r13; final weak filtering mask, bits 0 and 1
jz .store jz .store
; weak filtering mask ; weak filtering mask
mov r2, r14 mov r2, r14
@@ -734,11 +740,11 @@ cglobal hevc_h_loop_filter_chroma_10, 3, 4, 7, pix, stride, tc, pix0
;----------------------------------------------------------------------------- ;-----------------------------------------------------------------------------
cglobal hevc_v_loop_filter_luma_8, 4, 15, 16, pix, stride, beta, tc cglobal hevc_v_loop_filter_luma_8, 4, 15, 16, pix, stride, beta, tc
sub r0, 4 sub r0, 4
lea r5, [3*r1] lea r5, [3 * r1]
mov r6, r0 mov r6, r0
add r0, r5 add r0, r5
TRANSPOSE8x8B_LOAD PASS8ROWS(r6, r0, r1, r5) TRANSPOSE8x8B_LOAD PASS8ROWS(r6, r0, r1, r5)
LUMA_DEBLOCK_BODY 8, v LUMA_DEBLOCK_BODY 8, v
.store: .store:
TRANSPOSE8x8B_STORE PASS8ROWS(r6, r0, r1, r5) TRANSPOSE8x8B_STORE PASS8ROWS(r6, r0, r1, r5)
.bypassluma: .bypassluma:
@@ -746,11 +752,11 @@ cglobal hevc_v_loop_filter_luma_8, 4, 15, 16, pix, stride, beta, tc
cglobal hevc_v_loop_filter_luma_10, 4, 15, 16, pix, stride, beta, tc cglobal hevc_v_loop_filter_luma_10, 4, 15, 16, pix, stride, beta, tc
sub pixq, 8 sub pixq, 8
lea r5, [3*strideq] lea r5, [3 * strideq]
mov r6, pixq mov r6, pixq
add pixq, r5 add pixq, r5
TRANSPOSE8x8W_LOAD PASS8ROWS(r6, pixq, strideq, r5) TRANSPOSE8x8W_LOAD PASS8ROWS(r6, pixq, strideq, r5)
LUMA_DEBLOCK_BODY 10, v LUMA_DEBLOCK_BODY 10, v
.store: .store:
TRANSPOSE8x8W_STORE PASS8ROWS(r6, r0, r1, r5) TRANSPOSE8x8W_STORE PASS8ROWS(r6, r0, r1, r5)
.bypassluma: .bypassluma:
@@ -760,18 +766,18 @@ cglobal hevc_v_loop_filter_luma_10, 4, 15, 16, pix, stride, beta, tc
; void ff_hevc_h_loop_filter_luma(uint8_t *_pix, ptrdiff_t _stride, int *_beta, int *_tc, uint8_t *_no_p, uint8_t *_no_q); ; void ff_hevc_h_loop_filter_luma(uint8_t *_pix, ptrdiff_t _stride, int *_beta, int *_tc, uint8_t *_no_p, uint8_t *_no_q);
;----------------------------------------------------------------------------- ;-----------------------------------------------------------------------------
cglobal hevc_h_loop_filter_luma_8, 4, 15, 16, pix, stride, beta, tc, count, pix0, src3stride cglobal hevc_h_loop_filter_luma_8, 4, 15, 16, pix, stride, beta, tc, count, pix0, src3stride
lea src3strideq, [3*strideq] lea src3strideq, [3 * strideq]
mov pix0q, pixq mov pix0q, pixq
sub pix0q, src3strideq sub pix0q, src3strideq
sub pix0q, strideq sub pix0q, strideq
movdqu m0, [pix0q]; p3 movdqu m0, [pix0q]; p3
movdqu m1, [pix0q+strideq]; p2 movdqu m1, [pix0q + strideq]; p2
movdqu m2, [pix0q+2*strideq]; p1 movdqu m2, [pix0q + 2 * strideq]; p1
movdqu m3, [pix0q+src3strideq]; p0 movdqu m3, [pix0q + src3strideq]; p0
movdqu m4, [pixq]; q0 movdqu m4, [pixq]; q0
movdqu m5, [pixq+strideq]; q1 movdqu m5, [pixq + strideq]; q1
movdqu m6, [pixq+2*strideq]; q2 movdqu m6, [pixq + 2 * strideq]; q2
movdqu m7, [pixq+src3strideq]; q3 movdqu m7, [pixq + src3strideq]; q3
pxor m8, m8 pxor m8, m8
punpcklbw m0, m8 punpcklbw m0, m8
punpcklbw m1, m8 punpcklbw m1, m8
@@ -781,7 +787,7 @@ cglobal hevc_h_loop_filter_luma_8, 4, 15, 16, pix, stride, beta, tc, count, pix0
punpcklbw m5, m8 punpcklbw m5, m8
punpcklbw m6, m8 punpcklbw m6, m8
punpcklbw m7, m8 punpcklbw m7, m8
LUMA_DEBLOCK_BODY 8, h LUMA_DEBLOCK_BODY 8, h
.store: .store:
packuswb m1, m1; p2 packuswb m1, m1; p2
packuswb m2, m2; p1 packuswb m2, m2; p1
@@ -799,33 +805,33 @@ cglobal hevc_h_loop_filter_luma_8, 4, 15, 16, pix, stride, beta, tc, count, pix0
RET RET
cglobal hevc_h_loop_filter_luma_10, 4, 15, 16, pix, stride, beta, tc, count, pix0, src3stride cglobal hevc_h_loop_filter_luma_10, 4, 15, 16, pix, stride, beta, tc, count, pix0, src3stride
lea src3strideq, [3*strideq] lea src3strideq, [3 * strideq]
mov pix0q, pixq mov pix0q, pixq
sub pix0q, src3strideq sub pix0q, src3strideq
sub pix0q, strideq sub pix0q, strideq
movdqu m0, [pix0q]; p3 movdqu m0, [pix0q]; p3
movdqu m1, [pix0q+strideq]; p2 movdqu m1, [pix0q + strideq]; p2
movdqu m2, [pix0q+2*strideq]; p1 movdqu m2, [pix0q + 2 * strideq]; p1
movdqu m3, [pix0q+src3strideq]; p0 movdqu m3, [pix0q + src3strideq]; p0
movdqu m4, [pixq]; q0 movdqu m4, [pixq]; q0
movdqu m5, [pixq+strideq]; q1 movdqu m5, [pixq + strideq]; q1
movdqu m6, [pixq+2*strideq]; q2 movdqu m6, [pixq + 2 * strideq]; q2
movdqu m7, [pixq+src3strideq]; q3 movdqu m7, [pixq + src3strideq]; q3
LUMA_DEBLOCK_BODY 10, h LUMA_DEBLOCK_BODY 10, h
.store: .store:
pxor m8, m8; zeros reg pxor m8, m8; zeros reg
CLIPW m1, m8, [pw_pixel_max] CLIPW m1, m8, [pw_pixel_max]
CLIPW m2, m8, [pw_pixel_max] CLIPW m2, m8, [pw_pixel_max]
CLIPW m3, m8, [pw_pixel_max] CLIPW m3, m8, [pw_pixel_max]
CLIPW m4, m8, [pw_pixel_max] CLIPW m4, m8, [pw_pixel_max]
CLIPW m5, m8, [pw_pixel_max] CLIPW m5, m8, [pw_pixel_max]
CLIPW m6, m8, [pw_pixel_max] CLIPW m6, m8, [pw_pixel_max]
movdqu [pix0q+strideq], m1; p2 movdqu [pix0q + strideq], m1; p2
movdqu [pix0q+2*strideq], m2; p1 movdqu [pix0q + 2 * strideq], m2; p1
movdqu [pix0q+src3strideq], m3; p0 movdqu [pix0q + src3strideq], m3; p0
movdqu [pixq], m4; q0 movdqu [pixq ], m4; q0
movdqu [pixq+strideq], m5; q1 movdqu [pixq + strideq], m5; q1
movdqu [pixq+2*strideq], m6; q2 movdqu [pixq + 2 * strideq], m6; q2
.bypassluma: .bypassluma:
RET RET
%endmacro %endmacro