CSC

Compress Trace Files

Description

All the programs that access trace files can uncompress on-the-fly. This module maybe used to compress existing trace files. No compression programs are supplied with the package, although there are several good public domain compression programs available. Note that using the ZTR trace format will typically yield better compression than using any of the supported compression programs on an SCF file. Attempting to compress a ZTR file using this module will not decrease the size (and may even increase the file size). Also see section Trace Format Conversion.

Option: Compression method

This selects the algorithm used for compressing trace files. The choices are None, Compress (the standard UNIX compression program), Gzip (from GNU) and Bzip versions 1 and 2. These are listed in ascending order of compression ratios, with Bzip (either version) giving the best compression. Generally Gzip is the best supported program and is not too far behind Bzip. The following table provides comparisons with compression sizes on ABI and SCF files. (Note that the ABI file typically holds 12 bit trace data data.) The sizes listed are the average length, in bytes, of the 96 (originally ABI 3700 trace) files that the tests were performed on. The SCF and ZTR files were recalled using phred, so also contain confidence values. For comparison, the ZTR file sizes are also shown.

File type               Size in bytes
abi                     189150
compressed abi          104681
gzipped abi             87789
bzipped abi             62032
16-bit scf              82124
compress 16-bit scf     26574
gzipped 16-bit scf      25957
bzipped 16-bit scf      18877
16-bit ztr              17185
8-bit scf               45967
compress 8-bit scf      15000
gzipped 8-bit scf       14718
bzipped 8-bit scf       12659
8-bit ztr               11155