[Shootout-list] Timing resolution? libbsd-resource-perl

Isaac Gouy igouy2@yahoo.com
Fri, 6 Aug 2004 09:30:57 -0700 (PDT)


Please forgive my profound ignorance of Perl and Linux (which will
become obvious shortly).

The old Gentoo Linux setup I've been using to test some Shootout
Language contributions, seems to report cuser csystem to the nearest
hundredth of a second (using Perl times).

Is that the best resolution we have on the Shootout test machine -
0.01s?

On Doug Bagley's older slower test machine, most of the fastest
reported times were > 0.05s and many were > 0.1s; so timing to 0.01s
was not so bad.

On the current test machine only ~half the fastest timings are > 0.05s,
and several are 0.01s or less; so timing to 0.01s is not so good.


Maybe we could get better time resolution: "Also is provided times()
which provides the same functionality as the one in core Perl, only
with better time resolution."
   http://packages.debian.org/testing/perl/libbsd-resource-perl

Maybe we should increase N to get the fastest Shootout tests back into
0.1s 1s range.



		
__________________________________
Do you Yahoo!?
Yahoo! Mail - 50x more storage than other providers!
http://promotions.yahoo.com/new_mail