manual_4.html

高效率的一种通用压缩/解压程序

minimum:

<PRE>
   Machine type.  Operating system version.  
   Exact version of <CODE>bzip2</CODE> (do <CODE>bzip2 -V</CODE>).  
   Exact version of the compiler used.  
   Flags passed to the compiler.
</PRE>

<P>
However, the most important single thing that will help me is
the file that you were trying to compress or decompress at the
time the problem happened.  Without that, my ability to do anything
more than speculate about the cause, is limited.

</P>
<P>
Please remember that I connect to the Internet with a modem, so
you should contact me before mailing me huge files.

</P>



<H2><A NAME="SEC37" HREF="manual_toc.html#TOC37">Did you get the right package?</A></H2>

<P>
<CODE>bzip2</CODE> is a resource hog.  It soaks up large amounts of CPU cycles
and memory.  Also, it gives very large latencies.  In the worst case, you
can feed many megabytes of uncompressed data into the library before
getting any compressed output, so this probably rules out applications
requiring interactive behaviour.

</P>
<P>
These aren't faults of my implementation, I hope, but more
an intrinsic property of the Burrows-Wheeler transform (unfortunately).  
Maybe this isn't what you want.

</P>
<P>
If you want a compressor and/or library which is faster, uses less
memory but gets pretty good compression, and has minimal latency,
consider Jean-loup
Gailly's and Mark Adler's work, <CODE>zlib-1.1.2</CODE> and
<CODE>gzip-1.2.4</CODE>.  Look for them at
<CODE>http://www.cdrom.com/pub/infozip/zlib</CODE> and
<CODE>http://www.gzip.org</CODE> respectively.

</P>
<P>
For something faster and lighter still, you might try Markus F X J
Oberhumer's <CODE>LZO</CODE> real-time compression/decompression library, at
<BR> <CODE>http://wildsau.idv.uni-linz.ac.at/mfx/lzo.html</CODE>.

</P>
<P>
If you want to use the <CODE>bzip2</CODE> algorithms to compress small blocks
of data, 64k bytes or smaller, for example on an on-the-fly disk
compressor, you'd be well advised not to use this library.  Instead,
I've made a special library tuned for that kind of use.  It's part of
<CODE>e2compr-0.40</CODE>, an on-the-fly disk compressor for the Linux
<CODE>ext2</CODE> filesystem.  Look at
<CODE>http://www.netspace.net.au/~reiter/e2compr</CODE>.

</P>



<H2><A NAME="SEC38" HREF="manual_toc.html#TOC38">Testing</A></H2>

<P>
A record of the tests I've done.

</P>
<P>
First, some data sets:

<UL>
<LI>B: a directory containing a 6001 files, one for every length in the

      range 0 to 6000 bytes.  The files contain random lowercase
      letters.  18.7 megabytes.
<LI>H: my home directory tree.  Documents, source code, mail files,

      compressed data.  H contains B, and also a directory of 
      files designed as boundary cases for the sorting; mostly very
      repetitive, nasty files.  445 megabytes.
<LI>A: directory tree holding various applications built from source:

      <CODE>egcs-1.0.2</CODE>, <CODE>gcc-2.8.1</CODE>, KDE Beta 4, GTK, Octave, etc.
      827 megabytes.
<LI>P: directory tree holding large amounts of source code (<CODE>.tar</CODE>

      files) of the entire GNU distribution, plus a couple of
      Linux distributions.  2400 megabytes.
</UL>

<P>
The tests conducted are as follows.  Each test means compressing 
(a copy of) each file in the data set, decompressing it and
comparing it against the original.

</P>
<P>
First, a bunch of tests with block sizes, internal buffer
sizes and randomisation lengths set very small, 
to detect any problems with the
blocking, buffering and randomisation mechanisms.  
This required modifying the source code so as to try to 
break it.

<OL>
<LI>Data set H, with

      buffer size of 1 byte, and block size of 23 bytes.
<LI>Data set B, buffer sizes 1 byte, block size 1 byte.

<LI>As (2) but small-mode decompression (first 1700 files).

<LI>As (2) with block size 2 bytes.

<LI>As (2) with block size 3 bytes.

<LI>As (2) with block size 4 bytes.

<LI>As (2) with block size 5 bytes.

<LI>As (2) with block size 6 bytes and small-mode decompression.

<LI>H with normal buffer sizes (5000 bytes), normal block

      size (up to 900000 bytes), but with randomisation
      mechanism running intensely (randomising approximately every
      third byte).
<LI>As (9) with small-mode decompression.

</OL>

<P>
Then some tests with unmodified source code.

<OL>
<LI>H, all settings normal.

<LI>As (1), with small-mode decompress.

<LI>H, compress with flag <CODE>-1</CODE>.

<LI>H, compress with flag <CODE>-s</CODE>, decompress with flag <CODE>-s</CODE>.

<LI>Forwards compatibility: H, <CODE>bzip2-0.1pl2</CODE> compressing,

      <CODE>bzip2-0.9.0</CODE> decompressing, all settings normal.
<LI>Backwards compatibility:  H, <CODE>bzip2-0.9.0</CODE> compressing,

      <CODE>bzip2-0.1pl2</CODE> decompressing, all settings normal.
<LI>Bigger tests: A, all settings normal.

<LI>P, all settings normal.

<LI>Misc test: about 100 megabytes of <CODE>.tar</CODE> files with

      <CODE>bzip2</CODE> compiled with Purify.
<LI>Misc tests to make sure it builds and runs ok on non-Linux/x86

      platforms.
</OL>

<P>
These tests were conducted on a 205 MHz Cyrix 6x86MX machine, running
Linux 2.0.32.  They represent nearly a week of continuous computation.
All tests completed successfully.

</P>



<H2><A NAME="SEC39" HREF="manual_toc.html#TOC39">Further reading</A></H2>
<P>
<CODE>bzip2</CODE> is not research work, in the sense that it doesn't present
any new ideas.  Rather, it's an engineering exercise based on existing
ideas.

</P>
<P>
Four documents describe essentially all the ideas behind <CODE>bzip2</CODE>:

<PRE>
Michael Burrows and D. J. Wheeler:
  "A block-sorting lossless data compression algorithm"
   10th May 1994. 
   Digital SRC Research Report 124.
   ftp://ftp.digital.com/pub/DEC/SRC/research-reports/SRC-124.ps.gz
   If you have trouble finding it, try searching at the
   New Zealand Digital Library, http://www.nzdl.org.

Daniel S. Hirschberg and Debra A. LeLewer
  "Efficient Decoding of Prefix Codes"
   Communications of the ACM, April 1990, Vol 33, Number 4.
   You might be able to get an electronic copy of this
      from the ACM Digital Library.

David J. Wheeler
   Program bred3.c and accompanying document bred3.ps.
   This contains the idea behind the multi-table Huffman
   coding scheme.
   ftp://ftp.cl.cam.ac.uk/pub/user/djw3/

Jon L. Bentley and Robert Sedgewick
  "Fast Algorithms for Sorting and Searching Strings"
   Available from Sedgewick's web page,
   www.cs.princeton.edu/~rs
</PRE>

<P>
The following paper gives valuable additional insights into the
algorithm, but is not immediately the basis of any code
used in bzip2.

<PRE>
Peter Fenwick:
   Block Sorting Text Compression
   Proceedings of the 19th Australasian Computer Science Conference,
     Melbourne, Australia.  Jan 31 - Feb 2, 1996.
   ftp://ftp.cs.auckland.ac.nz/pub/peter-f/ACSC96paper.ps
</PRE>

<P>
Kunihiko Sadakane's sorting algorithm, mentioned above,
is available from:

<PRE>
http://naomi.is.s.u-tokyo.ac.jp/~sada/papers/Sada98b.ps.gz
</PRE>

<P>
The Manber-Myers suffix array construction
algorithm is described in a paper
available from:

<PRE>
http://www.cs.arizona.edu/people/gene/PAPERS/suffix.ps
</PRE>

<P><HR><P>
Go to the <A HREF="manual_1.html">first</A>, <A HREF="manual_3.html">previous</A>, next, last section, <A HREF="manual_toc.html">table of contents</A>.
</BODY>
</HTML>