• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1optimization Tips (for libavcodec):
2===================================
3
4What to optimize:
5-----------------
6If you plan to do non-x86 architecture specific optimizations (SIMD normally),
7then take a look in the x86/ directory, as most important functions are
8already optimized for MMX.
9
10If you want to do x86 optimizations then you can either try to fine-tune the
11stuff in the x86 directory or find some other functions in the C source to
12optimize, but there aren't many left.
13
14
15Understanding these overoptimized functions:
16--------------------------------------------
17As many functions tend to be a bit difficult to understand because
18of optimizations, it can be hard to optimize them further, or write
19architecture-specific versions. It is recommended to look at older
20revisions of the interesting files (web frontends for the various FFmpeg
21branches are listed at http://ffmpeg.org/download.html).
22Alternatively, look into the other architecture-specific versions in
23the x86/, ppc/, alpha/ subdirectories. Even if you don't exactly
24comprehend the instructions, it could help understanding the functions
25and how they can be optimized.
26
27NOTE: If you still don't understand some function, ask at our mailing list!!!
28(http://lists.ffmpeg.org/mailman/listinfo/ffmpeg-devel)
29
30
31When is an optimization justified?
32----------------------------------
33Normally, clean and simple optimizations for widely used codecs are
34justified even if they only achieve an overall speedup of 0.1%. These
35speedups accumulate and can make a big difference after awhile. Also, if
36none of the following factors get worse due to an optimization -- speed,
37binary code size, source size, source readability -- and at least one
38factor improves, then an optimization is always a good idea even if the
39overall gain is less than 0.1%. For obscure codecs that are not often
40used, the goal is more toward keeping the code clean, small, and
41readable instead of making it 1% faster.
42
43
44WTF is that function good for ....:
45-----------------------------------
46The primary purpose of this list is to avoid wasting time optimizing functions
47which are rarely used.
48
49put(_no_rnd)_pixels{,_x2,_y2,_xy2}
50    Used in motion compensation (en/decoding).
51
52avg_pixels{,_x2,_y2,_xy2}
53    Used in motion compensation of B-frames.
54    These are less important than the put*pixels functions.
55
56avg_no_rnd_pixels*
57    unused
58
59pix_abs16x16{,_x2,_y2,_xy2}
60    Used in motion estimation (encoding) with SAD.
61
62pix_abs8x8{,_x2,_y2,_xy2}
63    Used in motion estimation (encoding) with SAD of MPEG-4 4MV only.
64    These are less important than the pix_abs16x16* functions.
65
66put_mspel8_mc* / wmv2_mspel8*
67    Used only in WMV2.
68    it is not recommended that you waste your time with these, as WMV2
69    is an ugly and relatively useless codec.
70
71mpeg4_qpel* / *qpel_mc*
72    Used in MPEG-4 qpel motion compensation (encoding & decoding).
73    The qpel8 functions are used only for 4mv,
74    the avg_* functions are used only for B-frames.
75    Optimizing them should have a significant impact on qpel
76    encoding & decoding.
77
78qpel{8,16}_mc??_old_c / *pixels{8,16}_l4
79    Just used to work around a bug in an old libavcodec encoder version.
80    Don't optimize them.
81
82add_bytes/diff_bytes
83    For huffyuv only, optimize if you want a faster ffhuffyuv codec.
84
85get_pixels / diff_pixels
86    Used for encoding, easy.
87
88clear_blocks
89    easiest to optimize
90
91gmc
92    Used for MPEG-4 gmc.
93    Optimizing this should have a significant effect on the gmc decoding
94    speed.
95
96gmc1
97    Used for chroma blocks in MPEG-4 gmc with 1 warp point
98    (there are 4 luma & 2 chroma blocks per macroblock, so
99    only 1/3 of the gmc blocks use this, the other 2/3
100    use the normal put_pixel* code, but only if there is
101    just 1 warp point).
102    Note: DivX5 gmc always uses just 1 warp point.
103
104pix_sum
105    Used for encoding.
106
107hadamard8_diff / sse / sad == pix_norm1 / dct_sad / quant_psnr / rd / bit
108    Specific compare functions used in encoding, it depends upon the
109    command line switches which of these are used.
110    Don't waste your time with dct_sad & quant_psnr, they aren't
111    really useful.
112
113put_pixels_clamped / add_pixels_clamped
114    Used for en/decoding in the IDCT, easy.
115    Note, some optimized IDCTs have the add/put clamped code included and
116    then put_pixels_clamped / add_pixels_clamped will be unused.
117
118idct/fdct
119    idct (encoding & decoding)
120    fdct (encoding)
121    difficult to optimize
122
123dct_quantize_trellis
124    Used for encoding with trellis quantization.
125    difficult to optimize
126
127dct_quantize
128    Used for encoding.
129
130dct_unquantize_mpeg1
131    Used in MPEG-1 en/decoding.
132
133dct_unquantize_mpeg2
134    Used in MPEG-2 en/decoding.
135
136dct_unquantize_h263
137    Used in MPEG-4/H.263 en/decoding.
138
139
140
141Alignment:
142Some instructions on some architectures have strict alignment restrictions,
143for example most SSE/SSE2 instructions on x86.
144The minimum guaranteed alignment is written in the .h files, for example:
145    void (*put_pixels_clamped)(const int16_t *block/*align 16*/, uint8_t *pixels/*align 8*/, ptrdiff_t stride);
146
147
148General Tips:
149-------------
150Use asm loops like:
151__asm__(
152    "1: ....
153    ...
154    "jump_instruction ....
155Do not use C loops:
156do{
157    __asm__(
158        ...
159}while()
160
161For x86, mark registers that are clobbered in your asm. This means both
162general x86 registers (e.g. eax) as well as XMM registers. This last one is
163particularly important on Win64, where xmm6-15 are callee-save, and not
164restoring their contents leads to undefined results. In external asm,
165you do this by using:
166cglobal function_name, num_args, num_regs, num_xmm_regs
167In inline asm, you specify clobbered registers at the end of your asm:
168__asm__(".." ::: "%eax").
169If gcc is not set to support sse (-msse) it will not accept xmm registers
170in the clobber list. For that we use two macros to declare the clobbers.
171XMM_CLOBBERS should be used when there are other clobbers, for example:
172__asm__(".." ::: XMM_CLOBBERS("xmm0",) "eax");
173and XMM_CLOBBERS_ONLY should be used when the only clobbers are xmm registers:
174__asm__(".." :: XMM_CLOBBERS_ONLY("xmm0"));
175
176Do not expect a compiler to maintain values in your registers between separate
177(inline) asm code blocks. It is not required to. For example, this is bad:
178__asm__("movdqa %0, %%xmm7" : src);
179/* do something */
180__asm__("movdqa %%xmm7, %1" : dst);
181- first of all, you're assuming that the compiler will not use xmm7 in
182   between the two asm blocks.  It probably won't when you test it, but it's
183   a poor assumption that will break at some point for some --cpu compiler flag
184- secondly, you didn't mark xmm7 as clobbered. If you did, the compiler would
185   have restored the original value of xmm7 after the first asm block, thus
186   rendering the combination of the two blocks of code invalid
187Code that depends on data in registries being untouched, should be written as
188a single __asm__() statement. Ideally, a single function contains only one
189__asm__() block.
190
191Use external asm (nasm/yasm) or inline asm (__asm__()), do not use intrinsics.
192The latter requires a good optimizing compiler which gcc is not.
193
194When debugging a x86 external asm compilation issue, if lost in the macro
195expansions, add DBG=1 to your make command-line: the input file will be
196preprocessed, stripped of the debug/empty lines, then compiled, showing the
197actual lines causing issues.
198
199Inline asm vs. external asm
200---------------------------
201Both inline asm (__asm__("..") in a .c file, handled by a compiler such as gcc)
202and external asm (.s or .asm files, handled by an assembler such as nasm/yasm)
203are accepted in FFmpeg. Which one to use differs per specific case.
204
205- if your code is intended to be inlined in a C function, inline asm is always
206   better, because external asm cannot be inlined
207- if your code calls external functions, external asm is always better
208- if your code takes huge and complex structs as function arguments (e.g.
209   MpegEncContext; note that this is not ideal and is discouraged if there
210   are alternatives), then inline asm is always better, because predicting
211   member offsets in complex structs is almost impossible. It's safest to let
212   the compiler take care of that
213- in many cases, both can be used and it just depends on the preference of the
214   person writing the asm. For new asm, the choice is up to you. For existing
215   asm, you'll likely want to maintain whatever form it is currently in unless
216   there is a good reason to change it.
217- if, for some reason, you believe that a particular chunk of existing external
218   asm could be improved upon further if written in inline asm (or the other
219   way around), then please make the move from external asm <-> inline asm a
220   separate patch before your patches that actually improve the asm.
221
222
223Links:
224======
225http://www.aggregate.org/MAGIC/
226
227x86-specific:
228-------------
229http://developer.intel.com/design/pentium4/manuals/248966.htm
230
231The IA-32 Intel Architecture Software Developer's Manual, Volume 2:
232Instruction Set Reference
233http://developer.intel.com/design/pentium4/manuals/245471.htm
234
235http://www.agner.org/assem/
236
237AMD Athlon Processor x86 Code Optimization Guide:
238http://www.amd.com/us-en/assets/content_type/white_papers_and_tech_docs/22007.pdf
239
240
241ARM-specific:
242-------------
243ARM Architecture Reference Manual (up to ARMv5TE):
244http://www.arm.com/community/university/eulaarmarm.html
245
246Procedure Call Standard for the ARM Architecture:
247http://www.arm.com/pdfs/aapcs.pdf
248
249Optimization guide for ARM9E (used in Nokia 770 Internet Tablet):
250http://infocenter.arm.com/help/topic/com.arm.doc.ddi0240b/DDI0240A.pdf
251Optimization guide for ARM11 (used in Nokia N800 Internet Tablet):
252http://infocenter.arm.com/help/topic/com.arm.doc.ddi0211j/DDI0211J_arm1136_r1p5_trm.pdf
253Optimization guide for Intel XScale (used in Sharp Zaurus PDA):
254http://download.intel.com/design/intelxscale/27347302.pdf
255Intel Wireless MMX 2 Coprocessor: Programmers Reference Manual
256http://download.intel.com/design/intelxscale/31451001.pdf
257
258PowerPC-specific:
259-----------------
260PowerPC32/AltiVec PIM:
261www.freescale.com/files/32bit/doc/ref_manual/ALTIVECPEM.pdf
262
263PowerPC32/AltiVec PEM:
264www.freescale.com/files/32bit/doc/ref_manual/ALTIVECPIM.pdf
265
266CELL/SPU:
267http://www-01.ibm.com/chips/techlib/techlib.nsf/techdocs/30B3520C93F437AB87257060006FFE5E/$file/Language_Extensions_for_CBEA_2.4.pdf
268http://www-01.ibm.com/chips/techlib/techlib.nsf/techdocs/9F820A5FFA3ECE8C8725716A0062585F/$file/CBE_Handbook_v1.1_24APR2007_pub.pdf
269
270GCC asm links:
271--------------
272official doc but quite ugly
273http://gcc.gnu.org/onlinedocs/gcc/Extended-Asm.html
274
275a bit old (note "+" is valid for input-output, even though the next disagrees)
276http://www.cs.virginia.edu/~clc5q/gcc-inline-asm.pdf
277