Statistics
| Branch: | Revision:

ffmpeg / doc / optimization.txt @ 38aca760

History | View | Annotate | Download (5 KB)

1
optimization Tips (for libavcodec):
2

    
3
What to optimize:
4
if you plan to do non-x86 architecture specific optimiztions (SIMD normally) then
5
take a look in the i386/ directory, as most important functions are allready
6
optimized for MMX
7

    
8
if you want to do x86 optimizations then u can either try to finetune the stuff in the
9
i386 directory or find some other functions in the c source to optimize, but there
10
arent many left
11

    
12
Understanding these overoptimized functions:
13
as many functions, like the c ones tend to be a bit unreadable currently becouse
14
of optimizations it is difficult to understand them (and write arichtecture
15
specific versions, or optimize the c functions further) it is recommanded to look
16
at older CVS versions of the interresting files (just use CVSWEB at
17
http://www1.mplayerhq.hu/cgi-bin/cvsweb.cgi/ffmpeg/libavcodec/?cvsroot=FFmpeg)
18
or perhaps look into the other architecture specific versions in i386/, ppc/,
19
alpha/, ...; even if u dont understand the instructions exactly it could help
20
understanding the functions & how they can be optimized
21

    
22
NOTE:!!! if u still dont understand some function then ask at our mailing list!!!
23
(http://www1.mplayerhq.hu/mailman/listinfo/ffmpeg-devel)
24

    
25

    
26

    
27
wtf is that function good for ....:
28
the primary purpose of that list is to avoid wasting time to optimize functions
29
which are rarely used
30

    
31
put(_no_rnd)_pixels{,_x2,_y2,_xy2}
32
    used in motion compensation (en/decoding)
33

    
34
avg_pixels{,_x2,_y2,_xy2}
35
    used in motion compensation of B Frames
36
    these are less important then the put*pixels functions
37

    
38
avg_no_rnd_pixels*
39
    unused
40

    
41
pix_abs16x16{,_x2,_y2,_xy2}
42
    used in motion estimation (encoding) with SAD
43

    
44
pix_abs8x8{,_x2,_y2,_xy2}
45
    used in motion estimation (encoding) with SAD of MPEG4 4MV only
46
    these are less important then the pix_abs16x16* functions
47

    
48
put_mspel8_mc* / wmv2_mspel8*
49
    used only in WMV2
50
    it is not recommanded that u waste ur time with these, as WMV2 is a
51
    ugly and relativly useless codec
52

    
53
mpeg4_qpel* / *qpel_mc*
54
    use in MPEG4 qpel Motion compensation (encoding & decoding)
55
    the qpel8 functions are used only for 4mv
56
    the avg_* functions are used only for b frames
57
    optimizing them should have a significant impact on qpel encoding & decoding
58

    
59
qpel{8,16}_mc??_old_c / *pixels{8,16}_l4
60
    just used to workaround a bug in old libavcodec encoder
61
    dont optimze them
62

    
63
tpel_mc_func {put,avg}_tpel_pixels_tab
64
    used only for SVQ3, so only optimze them if u need fast SVQ3 decoding
65

    
66
add_bytes/diff_bytes
67
    for huffyuv only, optimize if u want a faster ff-huffyuv codec
68

    
69
get_pixels / diff_pixels
70
    used for encoding, easy
71

    
72
clear_blocks
73
    easiest, to optimize
74

    
75
gmc
76
    used for mpeg4 gmc
77
    optimizing this should have a significant effect on the gmc decoding speed but
78
    its very likely impossible to write in SIMD
79

    
80
gmc1
81
    used for chroma blocks in mpeg4 gmc with 1 warp point
82
    (there are 4 luma & 2 chroma blocks per macrobock, so
83
    only 1/3 of the gmc blocks use this, the other 2/3
84
    use the normal put_pixel* code, but only if there is
85
    just 1 warp point)
86
    Note: Divx5 gmc always uses just 1 warp point
87

    
88
pix_sum
89
    used for encoding
90

    
91
hadamard8_diff / sse / sad == pix_norm1 / dct_sad / quant_psnr / rd / bit
92
    specific compare functions used in encoding, it depends upon the command line
93
    switches which of these are used
94
    dont waste ur time with dct_sad & quant_psnr they arent really usefull
95

    
96
put_pixels_clamped / add_pixels_clamped
97
    used for en/decoding in the IDCT, easy
98
    Note, some optimized IDCTs have the add/put clamped code included and then
99
    put_pixels_clamped / add_pixels_clamped will be unused
100

    
101
idct/fdct
102
    idct (encoding & decoding)
103
    fdct (encoding)
104
    difficult to optimize
105

    
106
dct_quantize_trellis
107
    used for encoding with trellis quantization
108
    difficult to optimize
109

    
110
dct_quantize
111
    used for encoding
112

    
113
dct_unquantize_mpeg1
114
    used in mpeg1 en/decoding
115

    
116
dct_unquantize_mpeg2
117
    used in mpeg2 en/decoding
118

    
119
dct_unquantize_h263
120
    used in mpeg4/h263 en/decoding
121

    
122
FIXME remaining functions?
123
btw, most of these are in dsputil.c/.h some are in mpegvideo.c/.h
124

    
125

    
126

    
127
Alignment:
128
some instructions on some architectures have strict alignment restrictions,
129
for example most SSE/SSE2 inctructios on X86
130
the minimum guranteed alignment is writen in the .h files
131
for example:
132
    void (*put_pixels_clamped)(const DCTELEM *block/*align 16*/, UINT8 *pixels/*align 8*/, int line_size);
133

    
134

    
135

    
136
Links:
137
http://www.aggregate.org/MAGIC/
138

    
139
X86 specific:
140
http://developer.intel.com/design/pentium4/manuals/248966.htm
141

    
142
The IA-32 Intel Architecture Software Developer's Manual, Volume 2:
143
Instruction Set Reference
144
http://developer.intel.com/design/pentium4/manuals/245471.htm
145

    
146
http://www.agner.org/assem/
147

    
148
AMD Athlon Processor x86 Code Optimization Guide:
149
http://www.amd.com/us-en/assets/content_type/white_papers_and_tech_docs/22007.pdf
150

    
151
GCC asm links:
152
official doc but quite ugly
153
http://gcc.gnu.org/onlinedocs/gcc/Extended-Asm.html
154

    
155
a bit old (note "+" is valid for input-output, even though the next says its not)
156
http://www.cs.virginia.edu/~clc5q/gcc-inline-asm.pdf
157