rzip64


December 2010

NAME

rzip64 - a large-file compression program

SYNOPSIS

rzip64 [OPTIONS] <files...>

DESCRIPTION

rzip64 is a file compression program designed for very large files. When processing large chunks of input data it takes advantage of repetitiv appearing byte patterns over long distances.

rzip64 can utilize multiple CPU cores. rzip64 can also be interrupted at any time and be restarted later.

OPTIONS SUMMARY

Here is a summary of the options to rzip64.


 -0            fastest (worst) compression
 -6            default compression
 -9            slowest (best) compression
 -d            decompress
 -f            force overwrite of any existing files
 -G            use slices (stop-and-go mode)
 -h            print this help text
 -j n          use n cores in parallel
 -k            keep existing files
 -o filename   specify the output file name
 -P            show compression progress
 -S suffix     specify compressed suffix (default \&'\&.rz\&')
 -V            show version

OPTIONS

-0..9
Set the compression level from 0 to 9. The default is to use level 6, which is a reasonable compromise between speed and compression. The compression level is also strongly related to how much memory rzip64 uses, so if you are running rzip64 on a machine with limited amounts of memory then you will probably want to choose a smaller level.

-d
Decompress. If this option is not used then rzip64 looks at the name used to launch the program. If it contains the string 'runzip' then the -d option is automatically set.

-f
If this option is not specified then rzip64 will not overwrite any existing files. If you set this option then rzip64 will silently overwrite any files as needed.

-G
This option enables the stop-and-go mode of rzip64. If this option is used rzip64 may be interrupted via signal (say kill -15) at any time. When rzip64 is restarted later it will proceed compressing the file. -G always implies -9.

-h
Print an options summary page

-j n
Use n cpu cores in parallel. Using more than a single core autoselects -G, too.

-k
If this option is not specified then rzip64 will delete the source file after successful compression or decompression. When this option is specified then the source files are not deleted.

-o
Set the output file name. If this option is not set then the output file name is chosen based on the input name and the suffix. The -o option cannot be used if more than one file name is specified on the command line.

-P
If this option is specified then rzip64 will show the percentage progress while compressing.

-S
Set the compression suffix. The default is '.rz'.

-V
Print the rzip64 version number

INSTALLATION

Just install rzip64 in your search path.

STOP-AND-GO MODE

Compressing a large file takes time an imposes a considerable load on the machine. It is therefore desireable to run compression tasks only at times where no users are working on the machine.

rzip64 can therefore be interrupted at any time by a simple kill command. When rzip64 is restarted later it will continue the compression task until the job is completed. One may even change the number of cores when restarting.

HISTORY

The ideas behind rzip64 were first implemented in 1998 during the work on rsync. That version was too slow to be practical, and was replaced by version 2.1 in 2003.

The stop-and-go option was added in 09/2010 (version 2.2).

Multicore support was added in 12/2010. The version number was advanced to 3.0 since rzip64 with multiple treads requires a considerable amount of memory address space that can typically be provided by 64 bit environments only. It is not a matter of physical RAM modules but of virtual address space. Each cpu core requires about 900 MByte address space.

One can still compile an run rzip64 in 32 bit environments (while using only a small number of cpu cores) but that is no longer recommended.

COMPATIBILITY

The file format for rzip64 is fully compatible to rzip 2.1 and 2.2 .

NOTE

Unlike most Unix compression programs, rzip64 cannot compress or decompress to or from standard input or standard output. That can possibly be changed but appears to make not much sense when -G is in use.

RESSOURCES

If you wish to report a problem or make a suggestion then please email rzip64-at-GHSi.eu

Source code and binary packages are available here:

Some additional background infos and benchmarking results may be found here.

AUTHOR

Different people have put work in the development of rzip64. Our thanks include

  • Andrew Tridgell for the first implementation of rzip
  • The authors of bzlib for an excellent library
  • Paul Russell for many suggestions and the debian packaging
  • Kay Gorontzi (Homepage) for the stop-and-go mode and the multicore extensions
  • rzip64 is released under the GNU General Public License version 3 or later. Please see the file COPYING for license details.