generic-library/ffmpeg
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

spelling/grammar/wording/phrasing

Browse Source 
Originally committed as revision 4376 to svn://svn.ffmpeg.org/ffmpeg/trunk
This commit is contained in:
Diego Biurrun 2005-06-11 17:41:23 +00:00
parent 38aca76027
commit 8ea9ce413b
1 changed files with 63 additions and 63 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

126
doc/optimization.txt
View File

@ -1,102 +1,104 @@
optimization Tips (for libavcodec): optimization Tips (for libavcodec):
What to optimize: What to optimize:
if you plan to do non-x86 architecture specific optimiztions (SIMD normally) then If you plan to do non-x86 architecture specific optimiztions (SIMD normally)
take a look in the i386/ directory, as most important functions are allready then take a look in the i386/ directory, as most important functions are
optimized for MMX already optimized for MMX.
if you want to do x86 optimizations then u can either try to finetune the stuff in the If you want to do x86 optimizations then you can either try to finetune the
i386 directory or find some other functions in the c source to optimize, but there stuff in the i386 directory or find some other functions in the C source to
arent many left optimize, but there aren't many left.
Understanding these overoptimized functions: Understanding these overoptimized functions:
as many functions, like the c ones tend to be a bit unreadable currently becouse As many functions, like the C ones tend to be a bit unreadable currently
of optimizations it is difficult to understand them (and write arichtecture because of optimizations it is difficult to understand them (and write
specific versions, or optimize the c functions further) it is recommanded to look architecture specific versions, or optimize the C functions further) it is
at older CVS versions of the interresting files (just use CVSWEB at recommended to look at older CVS versions of the interesting files (just use
http://www1.mplayerhq.hu/cgi-bin/cvsweb.cgi/ffmpeg/libavcodec/?cvsroot=FFmpeg) ViewCVS at http://www1.mplayerhq.hu/cgi-bin/cvsweb.cgi/ffmpeg/?cvsroot=FFMpeg)
or perhaps look into the other architecture specific versions in i386/, ppc/, or perhaps look into the other architecture-specific versions in i386/, ppc/,
alpha/, ...; even if u dont understand the instructions exactly it could help alpha/, ... Even if you don't understand the instructions exactly it could
understanding the functions & how they can be optimized help understanding the functions & how they can be optimized.
NOTE:!!! if u still dont understand some function then ask at our mailing list!!! NOTE: If you still don't understand some function, ask at our mailing list!!!
(http://www1.mplayerhq.hu/mailman/listinfo/ffmpeg-devel) (http://www1.mplayerhq.hu/mailman/listinfo/ffmpeg-devel)
wtf is that function good for ....: WTF is that function good for ....:
the primary purpose of that list is to avoid wasting time to optimize functions The primary purpose of that list is to avoid wasting time to optimize functions
which are rarely used which are rarely used
put(_no_rnd)_pixels{,_x2,_y2,_xy2} put(_no_rnd)_pixels{,_x2,_y2,_xy2}
used in motion compensation (en/decoding) Used in motion compensation (en/decoding).
avg_pixels{,_x2,_y2,_xy2} avg_pixels{,_x2,_y2,_xy2}
used in motion compensation of B Frames Used in motion compensation of B-frames.
these are less important then the put*pixels functions These are less important then the put*pixels functions.
avg_no_rnd_pixels* avg_no_rnd_pixels*
unused unused
pix_abs16x16{,_x2,_y2,_xy2} pix_abs16x16{,_x2,_y2,_xy2}
used in motion estimation (encoding) with SAD Used in motion estimation (encoding) with SAD.
pix_abs8x8{,_x2,_y2,_xy2} pix_abs8x8{,_x2,_y2,_xy2}
used in motion estimation (encoding) with SAD of MPEG4 4MV only Used in motion estimation (encoding) with SAD of MPEG-4 4MV only.
these are less important then the pix_abs16x16* functions These are less important then the pix_abs16x16* functions.
put_mspel8_mc* / wmv2_mspel8* put_mspel8_mc* / wmv2_mspel8*
used only in WMV2 Used only in WMV2.
it is not recommanded that u waste ur time with these, as WMV2 is a it is not recommended that you waste your time with these, as WMV2
ugly and relativly useless codec is an ugly and relatively useless codec.
mpeg4_qpel* / *qpel_mc* mpeg4_qpel* / *qpel_mc*
use in MPEG4 qpel Motion compensation (encoding & decoding) Used in MPEG-4 qpel motion compensation (encoding & decoding).
the qpel8 functions are used only for 4mv The qpel8 functions are used only for 4mv,
the avg_* functions are used only for b frames the avg_* functions are used only for B-frames.
optimizing them should have a significant impact on qpel encoding & decoding Optimizing them should have a significant impact on qpel
encoding & decoding.
qpel{8,16}_mc??_old_c / *pixels{8,16}_l4 qpel{8,16}_mc??_old_c / *pixels{8,16}_l4
just used to workaround a bug in old libavcodec encoder Just used to work around a bug in an old libavcodec encoder version.
dont optimze them Don't optimize them.
tpel_mc_func {put,avg}_tpel_pixels_tab tpel_mc_func {put,avg}_tpel_pixels_tab
used only for SVQ3, so only optimze them if u need fast SVQ3 decoding Used only for SVQ3, so only optimize them if you need fast SVQ3 decoding.
add_bytes/diff_bytes add_bytes/diff_bytes
for huffyuv only, optimize if u want a faster ff-huffyuv codec For huffyuv only, optimize if you want a faster ffhuffyuv codec.
get_pixels / diff_pixels get_pixels / diff_pixels
used for encoding, easy Used for encoding, easy.
clear_blocks clear_blocks
easiest, to optimize easiest to optimize
gmc gmc
used for mpeg4 gmc Used for MPEG-4 gmc.
optimizing this should have a significant effect on the gmc decoding speed but Optimizing this should have a significant effect on the gmc decoding
its very likely impossible to write in SIMD speed but it's very likely impossible to write in SIMD.
gmc1 gmc1
used for chroma blocks in mpeg4 gmc with 1 warp point Used for chroma blocks in MPEG-4 gmc with 1 warp point
(there are 4 luma & 2 chroma blocks per macrobock, so (there are 4 luma & 2 chroma blocks per macroblock, so
only 1/3 of the gmc blocks use this, the other 2/3 only 1/3 of the gmc blocks use this, the other 2/3
use the normal put_pixel* code, but only if there is use the normal put_pixel* code, but only if there is
just 1 warp point) just 1 warp point).
Note: Divx5 gmc always uses just 1 warp point Note: DivX5 gmc always uses just 1 warp point.
pix_sum pix_sum
used for encoding Used for encoding.
hadamard8_diff / sse / sad == pix_norm1 / dct_sad / quant_psnr / rd / bit hadamard8_diff / sse / sad == pix_norm1 / dct_sad / quant_psnr / rd / bit
specific compare functions used in encoding, it depends upon the command line Specific compare functions used in encoding, it depends upon the
switches which of these are used command line switches which of these are used.
dont waste ur time with dct_sad & quant_psnr they arent really usefull Don't waste your time with dct_sad & quant_psnr, they aren't
really useful.
put_pixels_clamped / add_pixels_clamped put_pixels_clamped / add_pixels_clamped
used for en/decoding in the IDCT, easy Used for en/decoding in the IDCT, easy.
Note, some optimized IDCTs have the add/put clamped code included and then Note, some optimized IDCTs have the add/put clamped code included and
put_pixels_clamped / add_pixels_clamped will be unused then put_pixels_clamped / add_pixels_clamped will be unused.
idct/fdct idct/fdct
idct (encoding & decoding) idct (encoding & decoding)
@ -104,31 +106,30 @@ idct/fdct
difficult to optimize difficult to optimize
dct_quantize_trellis dct_quantize_trellis
used for encoding with trellis quantization Used for encoding with trellis quantization.
difficult to optimize difficult to optimize
dct_quantize dct_quantize
used for encoding Used for encoding.
dct_unquantize_mpeg1 dct_unquantize_mpeg1
used in mpeg1 en/decoding Used in MPEG-1 en/decoding.
dct_unquantize_mpeg2 dct_unquantize_mpeg2
used in mpeg2 en/decoding Used in MPEG-2 en/decoding.
dct_unquantize_h263 dct_unquantize_h263
used in mpeg4/h263 en/decoding Used in MPEG-4/H.263 en/decoding.
FIXME remaining functions? FIXME remaining functions?
btw, most of these are in dsputil.c/.h some are in mpegvideo.c/.h BTW, most of these functions are in dsputil.c/.h, some are in mpegvideo.c/.h.
Alignment: Alignment:
some instructions on some architectures have strict alignment restrictions, Some instructions on some architectures have strict alignment restrictions,
for example most SSE/SSE2 inctructios on X86 for example most SSE/SSE2 inctructios on x86.
the minimum guranteed alignment is writen in the .h files The minimum guaranteed alignment is written in the .h files, for example:
for example:
void (*put_pixels_clamped)(const DCTELEM *block/*align 16*/, UINT8 *pixels/*align 8*/, int line_size); void (*put_pixels_clamped)(const DCTELEM *block/*align 16*/, UINT8 *pixels/*align 8*/, int line_size);
@ -136,7 +137,7 @@ for example:
Links: Links:
http://www.aggregate.org/MAGIC/ http://www.aggregate.org/MAGIC/
X86 specific: x86-specific:
http://developer.intel.com/design/pentium4/manuals/248966.htm http://developer.intel.com/design/pentium4/manuals/248966.htm
The IA-32 Intel Architecture Software Developer's Manual, Volume 2: The IA-32 Intel Architecture Software Developer's Manual, Volume 2:
@ -152,6 +153,5 @@ GCC asm links:
official doc but quite ugly official doc but quite ugly
http://gcc.gnu.org/onlinedocs/gcc/Extended-Asm.html http://gcc.gnu.org/onlinedocs/gcc/Extended-Asm.html
a bit old (note "+" is valid for input-output, even though the next says its not) a bit old (note "+" is valid for input-output, even though the next disagrees)
http://www.cs.virginia.edu/~clc5q/gcc-inline-asm.pdf http://www.cs.virginia.edu/~clc5q/gcc-inline-asm.pdf