There is a program called 'Parallel BZIP2 (PBZIP2)' http://compression.ca/pbzip2/
which does exactly what you would think from its name. Detecting if it exists, and if so using it in preference to bzip2 would be very easy. There is an option: -l : load average determines max number processors to use which I think would be sensible to use. There is one problem, which I personally think is worth accepting. The program can not read from stdin, or support pipes. Hence a large tar file first needs to be created, then compressed. That probably means at least 1 GB extra disk space is needed in order to create the binary distribution. I've not checked out the claims of the speedup (in fact, I've not even downloaded its source code yet), but assuming it does work reaonsably well, what are your thoughts on implementing this? Personally, I think if it gives a reasonable speedup, we should use it if available. Dave -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel-unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org
