1 optimization Tips (for libavcodec):
4 If you plan to do non-x86 architecture specific optimizations (SIMD normally),
5 then take a look in the i386/ directory, as most important functions are
6 already optimized for MMX.
8 If you want to do x86 optimizations then you can either try to finetune the
9 stuff in the i386 directory or find some other functions in the C source to
10 optimize, but there aren't many left.
12 Understanding these overoptimized functions:
13 As many functions tend to be a bit difficult to understand because
14 of optimizations, it can be hard to optimize them further, or write
15 architecture-specific versions. It is recommened to look at older
16 revisions of the interesting files (for a web frontend try ViewVC at
17 http://svn.mplayerhq.hu/ffmpeg/trunk/).
18 Alternatively, look into the other architecture-specific versions in
19 the i386/, ppc/, alpha/ subdirectories. Even if you don't exactly
20 comprehend the instructions, it could help understanding the functions
21 and how they can be optimized.
23 NOTE: If you still don't understand some function, ask at our mailing list!!!
24 (http://lists.mplayerhq.hu/mailman/listinfo/ffmpeg-devel)
28 WTF is that function good for ....:
29 The primary purpose of that list is to avoid wasting time to optimize functions
32 put(_no_rnd)_pixels{,_x2,_y2,_xy2}
33 Used in motion compensation (en/decoding).
35 avg_pixels{,_x2,_y2,_xy2}
36 Used in motion compensation of B-frames.
37 These are less important than the put*pixels functions.
42 pix_abs16x16{,_x2,_y2,_xy2}
43 Used in motion estimation (encoding) with SAD.
45 pix_abs8x8{,_x2,_y2,_xy2}
46 Used in motion estimation (encoding) with SAD of MPEG-4 4MV only.
47 These are less important than the pix_abs16x16* functions.
49 put_mspel8_mc* / wmv2_mspel8*
51 it is not recommended that you waste your time with these, as WMV2
52 is an ugly and relatively useless codec.
54 mpeg4_qpel* / *qpel_mc*
55 Used in MPEG-4 qpel motion compensation (encoding & decoding).
56 The qpel8 functions are used only for 4mv,
57 the avg_* functions are used only for B-frames.
58 Optimizing them should have a significant impact on qpel
61 qpel{8,16}_mc??_old_c / *pixels{8,16}_l4
62 Just used to work around a bug in an old libavcodec encoder version.
65 tpel_mc_func {put,avg}_tpel_pixels_tab
66 Used only for SVQ3, so only optimize them if you need fast SVQ3 decoding.
69 For huffyuv only, optimize if you want a faster ffhuffyuv codec.
71 get_pixels / diff_pixels
72 Used for encoding, easy.
79 Optimizing this should have a significant effect on the gmc decoding
83 Used for chroma blocks in MPEG-4 gmc with 1 warp point
84 (there are 4 luma & 2 chroma blocks per macroblock, so
85 only 1/3 of the gmc blocks use this, the other 2/3
86 use the normal put_pixel* code, but only if there is
88 Note: DivX5 gmc always uses just 1 warp point.
93 hadamard8_diff / sse / sad == pix_norm1 / dct_sad / quant_psnr / rd / bit
94 Specific compare functions used in encoding, it depends upon the
95 command line switches which of these are used.
96 Don't waste your time with dct_sad & quant_psnr, they aren't
99 put_pixels_clamped / add_pixels_clamped
100 Used for en/decoding in the IDCT, easy.
101 Note, some optimized IDCTs have the add/put clamped code included and
102 then put_pixels_clamped / add_pixels_clamped will be unused.
105 idct (encoding & decoding)
107 difficult to optimize
110 Used for encoding with trellis quantization.
111 difficult to optimize
117 Used in MPEG-1 en/decoding.
120 Used in MPEG-2 en/decoding.
123 Used in MPEG-4/H.263 en/decoding.
125 FIXME remaining functions?
126 BTW, most of these functions are in dsputil.c/.h, some are in mpegvideo.c/.h.
131 Some instructions on some architectures have strict alignment restrictions,
132 for example most SSE/SSE2 instructions on x86.
133 The minimum guaranteed alignment is written in the .h files, for example:
134 void (*put_pixels_clamped)(const DCTELEM *block/*align 16*/, UINT8 *pixels/*align 8*/, int line_size);
139 http://www.aggregate.org/MAGIC/
142 http://developer.intel.com/design/pentium4/manuals/248966.htm
144 The IA-32 Intel Architecture Software Developer's Manual, Volume 2:
145 Instruction Set Reference
146 http://developer.intel.com/design/pentium4/manuals/245471.htm
148 http://www.agner.org/assem/
150 AMD Athlon Processor x86 Code Optimization Guide:
151 http://www.amd.com/us-en/assets/content_type/white_papers_and_tech_docs/22007.pdf
156 ARM Architecture Reference Manual (up to ARMv5TE):
157 http://www.arm.com/community/university/eulaarmarm.html
159 Procedure Call Standard for the ARM Architecture:
160 http://www.arm.com/pdfs/aapcs.pdf
162 Optimization guide for ARM9E (used in Nokia 770 Internet Tablet):
163 http://infocenter.arm.com/help/topic/com.arm.doc.ddi0240b/DDI0240A.pdf
164 Optimization guide for ARM11 (used in Nokia N800 Internet Tablet):
165 http://infocenter.arm.com/help/topic/com.arm.doc.ddi0211j/DDI0211J_arm1136_r1p5_trm.pdf
166 Optimization guide for Intel XScale (used in Sharp Zaurus PDA):
167 http://download.intel.com/design/intelxscale/27347302.pdf
171 PowerPC32/AltiVec PIM:
172 www.freescale.com/files/32bit/doc/ref_manual/ALTIVECPEM.pdf
174 PowerPC32/AltiVec PEM:
175 www.freescale.com/files/32bit/doc/ref_manual/ALTIVECPIM.pdf
178 http://www-01.ibm.com/chips/techlib/techlib.nsf/techdocs/30B3520C93F437AB87257060006FFE5E/$file/Language_Extensions_for_CBEA_2.4.pdf
179 http://www-01.ibm.com/chips/techlib/techlib.nsf/techdocs/9F820A5FFA3ECE8C8725716A0062585F/$file/CBE_Handbook_v1.1_24APR2007_pub.pdf
182 SPARC Joint Programming Specification (JPS1): Commonality
183 http://www.fujitsu.com/downloads/PRMPWR/JPS1-R1.0.4-Common-pub.pdf
185 UltraSPARC III Processor User's Manual (contains instruction timings)
186 http://www.sun.com/processors/manuals/USIIIv2.pdf
188 VIS Whitepaper (contains optimization guidelines)
189 http://www.sun.com/processors/vis/download/vis/vis_whitepaper.pdf
192 official doc but quite ugly
193 http://gcc.gnu.org/onlinedocs/gcc/Extended-Asm.html
195 a bit old (note "+" is valid for input-output, even though the next disagrees)
196 http://www.cs.virginia.edu/~clc5q/gcc-inline-asm.pdf