1 |
;/************************************************************************** |
;/**************************************************************************** |
2 |
; * |
; * |
3 |
; * XVID MPEG-4 VIDEO CODEC |
; * XVID MPEG-4 VIDEO CODEC |
4 |
; * mmx cbp calc |
; * - MMX CBP computation - |
5 |
; * |
; * |
6 |
; * This program is an implementation of a part of one or more MPEG-4 |
; * Copyright (C) 2005 Carlo Bramini <carlo.bramix@libero.it> |
7 |
; * Video tools as specified in ISO/IEC 14496-2 standard. Those intending |
; * 2001-2003 Peter Ross <pross@xvid.org> |
8 |
; * to use this software module in hardware or software products are |
; * 2002-2003 Pascal Massimino <skal@planet-d.net> |
|
; * advised that its use may infringe existing patents or copyrights, and |
|
|
; * any such use would be at such party's own risk. The original |
|
|
; * developer of this software module and his/her company, and subsequent |
|
|
; * editors and their companies, will have no liability for use of this |
|
|
; * software or modifications or derivatives thereof. |
|
9 |
; * |
; * |
10 |
; * This program is free software; you can redistribute it and/or modify |
; * This program is free software; you can redistribute it and/or modify |
11 |
; * it under the terms of the GNU General Public License as published by |
; * it under the terms of the GNU General Public License as published by |
19 |
; * |
; * |
20 |
; * You should have received a copy of the GNU General Public License |
; * You should have received a copy of the GNU General Public License |
21 |
; * along with this program; if not, write to the Free Software |
; * along with this program; if not, write to the Free Software |
22 |
; * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. |
; * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA |
23 |
; * |
; * |
24 |
; *************************************************************************/ |
; * $Id$ |
|
|
|
|
;/************************************************************************** |
|
|
; * |
|
|
; * History: |
|
25 |
; * |
; * |
26 |
; * 14.06.2002 some further tweaks by -Skal- |
; ***************************************************************************/ |
|
; * 24.11.2001 inital version; (c)2001 peter ross <pross@xvid.org> |
|
|
; * |
|
|
; *************************************************************************/ |
|
|
|
|
27 |
|
|
28 |
bits 32 |
BITS 32 |
29 |
|
|
30 |
section .data |
;============================================================================= |
31 |
|
; Macros |
32 |
|
;============================================================================= |
33 |
|
|
34 |
%macro cglobal 1 |
%macro cglobal 1 |
35 |
%ifdef PREFIX |
%ifdef PREFIX |
36 |
|
%ifdef MARK_FUNCS |
37 |
|
global _%1:function %1.endfunc-%1 |
38 |
|
%define %1 _%1:function %1.endfunc-%1 |
39 |
|
%else |
40 |
global _%1 |
global _%1 |
41 |
%define %1 _%1 |
%define %1 _%1 |
42 |
|
%endif |
43 |
|
%else |
44 |
|
%ifdef MARK_FUNCS |
45 |
|
global %1:function %1.endfunc-%1 |
46 |
%else |
%else |
47 |
global %1 |
global %1 |
48 |
%endif |
%endif |
49 |
|
%endif |
50 |
%endmacro |
%endmacro |
51 |
|
|
52 |
align 16 |
;============================================================================= |
53 |
|
; Local data |
54 |
|
;============================================================================= |
55 |
|
|
56 |
|
%ifdef FORMAT_COFF |
57 |
|
SECTION .rodata |
58 |
|
%else |
59 |
|
SECTION .rodata align=16 |
60 |
|
%endif |
61 |
|
|
62 |
ignore_dc dw 0, -1, -1, -1, -1, -1, -1, -1 |
ALIGN 16 |
63 |
|
|
64 |
section .text |
mult_mask: |
65 |
|
db 0x10,0x20,0x04,0x08,0x01,0x02,0x00,0x00 |
66 |
|
ignore_dc: |
67 |
|
dw 0, -1, -1, -1 |
68 |
|
|
69 |
|
;============================================================================= |
70 |
|
; Code |
71 |
|
;============================================================================= |
72 |
|
|
73 |
|
SECTION .text |
74 |
|
|
75 |
cglobal calc_cbp_mmx |
cglobal calc_cbp_mmx |
76 |
|
|
77 |
;=========================================================================== |
;----------------------------------------------------------------------------- |
|
; |
|
78 |
; uint32_t calc_cbp_mmx(const int16_t coeff[6][64]); |
; uint32_t calc_cbp_mmx(const int16_t coeff[6][64]); |
79 |
; |
;----------------------------------------------------------------------------- |
|
;=========================================================================== |
|
80 |
|
|
81 |
align 16 |
%macro MAKE_LOAD 1 |
82 |
calc_cbp_mmx: |
por mm0, [eax-128*1+%1*8] |
83 |
push ebx |
por mm1, [eax+128*0+%1*8] |
84 |
push esi |
por mm2, [eax+128*1+%1*8] |
85 |
|
por mm3, [eax+128*2+%1*8] |
86 |
|
por mm4, [eax+128*3+%1*8] |
87 |
|
por mm5, [eax+128*4+%1*8] |
88 |
|
%endmacro |
89 |
|
|
90 |
mov esi, [esp + 8 + 4] ; coeff |
ALIGN 16 |
91 |
xor eax, eax ; cbp = 0 |
calc_cbp_mmx: |
92 |
mov edx, (1 << 5) |
mov eax, [esp + 4] ; coeff |
93 |
|
|
94 |
movq mm7, [ignore_dc] |
movq mm7, [ignore_dc] |
95 |
|
pxor mm6, mm6 ; used only for comparing |
96 |
.loop |
movq mm0, [eax+128*0] |
97 |
movq mm0, [esi] |
movq mm1, [eax+128*1] |
98 |
movq mm1, [esi+8] |
movq mm2, [eax+128*2] |
99 |
|
movq mm3, [eax+128*3] |
100 |
|
movq mm4, [eax+128*4] |
101 |
|
movq mm5, [eax+128*5] |
102 |
|
add eax, 8+128 |
103 |
pand mm0, mm7 |
pand mm0, mm7 |
104 |
|
pand mm1, mm7 |
105 |
|
pand mm2, mm7 |
106 |
|
pand mm3, mm7 |
107 |
|
pand mm4, mm7 |
108 |
|
pand mm5, mm7 |
109 |
|
|
110 |
|
MAKE_LOAD 0 |
111 |
|
MAKE_LOAD 1 |
112 |
|
MAKE_LOAD 2 |
113 |
|
MAKE_LOAD 3 |
114 |
|
MAKE_LOAD 4 |
115 |
|
MAKE_LOAD 5 |
116 |
|
MAKE_LOAD 6 |
117 |
|
MAKE_LOAD 7 |
118 |
|
MAKE_LOAD 8 |
119 |
|
MAKE_LOAD 9 |
120 |
|
MAKE_LOAD 10 |
121 |
|
MAKE_LOAD 11 |
122 |
|
MAKE_LOAD 12 |
123 |
|
MAKE_LOAD 13 |
124 |
|
MAKE_LOAD 14 |
125 |
|
|
126 |
|
movq mm7, [mult_mask] |
127 |
|
packssdw mm0, mm1 |
128 |
|
packssdw mm2, mm3 |
129 |
|
packssdw mm4, mm5 |
130 |
|
packssdw mm0, mm2 |
131 |
|
packssdw mm4, mm6 |
132 |
|
pcmpgtw mm0, mm6 |
133 |
|
pcmpgtw mm4, mm6 |
134 |
|
psrlw mm0, 15 |
135 |
|
psrlw mm4, 15 |
136 |
|
packuswb mm0, mm4 |
137 |
|
pmaddwd mm0, mm7 |
138 |
|
|
|
por mm0, [esi+16] |
|
|
por mm1, [esi+24] |
|
|
|
|
|
por mm0, [esi+32] |
|
|
por mm1, [esi+40] |
|
|
|
|
|
por mm0, [esi+48] |
|
|
por mm1, [esi+56] |
|
|
|
|
|
por mm0, [esi+64] |
|
|
por mm1, [esi+72] |
|
|
|
|
|
por mm0, [esi+80] |
|
|
por mm1, [esi+88] |
|
|
|
|
|
por mm0, [esi+96] |
|
|
por mm1, [esi+104] |
|
|
|
|
|
por mm0, [esi+112] |
|
|
por mm1, [esi+120] |
|
|
|
|
|
por mm0, mm1 |
|
139 |
movq mm1, mm0 |
movq mm1, mm0 |
140 |
psrlq mm1, 32 |
psrlq mm1, 32 |
141 |
lea esi, [esi + 128] |
paddusb mm0, mm1 |
|
|
|
|
por mm0, mm1 |
|
|
movd ebx, mm0 |
|
|
|
|
|
test ebx, ebx |
|
|
jz .next |
|
|
or eax, edx ; cbp |= 1 << (5-i) |
|
|
|
|
|
.next |
|
|
shr edx,1 |
|
|
jnc .loop |
|
|
|
|
|
pop esi |
|
|
pop ebx |
|
142 |
|
|
143 |
|
movd eax, mm0 |
144 |
|
shr eax, 8 |
145 |
|
and eax, 0x3F |
146 |
ret |
ret |
147 |
|
.endfunc |
148 |
|
|