Re: Gzipping man files
On Tue, 16 Apr 1996, Bruce Perens wrote:
> Have you timed the speed of un-compression? It's 10 times faster than
> compression - it doesn't seem to be a tremendous overhead. Even the
> largest man page uncompressed very quickly on my system. Most of the
> time seemed to be spent in disk I/O, which we'd have anyway.
uncompressing man pages would be probably be faster anyway - less
disk space consumed means less disk I/O...and disk I/O is the real
bottleneck, anyway.
e.g. assuming a compression ratio of 3:1 for man page text, reading 1K
from disk & uncompressing it is almost certain to be faster and less of
a load on the system than reading 3K from disk.
Even with old 486-33 (or even 386-33!) machines the only time CPU
usage becomes a significant concern is when serious number crunching
or graphics processing or program compilation is going on in the
background. Otherwise disk thrashing, memory, and swapping are the real
speed killers.
Craig
Reply to: