> [analog] looks nice, but I am not sure it is suited for my needs
> [because the logs I'm parsing aren't web logs]
Sorry, my brain must have slipped out of gear
while I wasn't looking.
Forget analog.
The file count doesn't seem that much. I suggest
you use opendir and cousins to process files one
at a time.
Will you significantly benefit from running multiple
scripts simultaneously? Using 3 cpus for N seconds
rather than 1 cpu for 3N seconds?
I think you are best off putting your effort into creating
a single threaded version of the task, then profiling
and tweaking that. (I think it will probably boil down to
making the regexes more efficient.)
I suggest you search the archives of this group for
help on using opendir, and on profiling, and references
for making regexes go faster.
