Perrin Harkins wrote: >>You cannot reliably measure CPU clocks with wallclock on the >>multi-processor machine, unless you are running on Dos :) >> > > Even so, wall time is what most people actually care about, and it's > fine to use if you're the only one doing work on that machine.
Yes, for counting the total run-time of the code that does something. No, for benchmarking nearly-empty subroutines run times. Since the error here can be 1000% and more. You are still running on time-sharing machine and if your sub didn't fit into one CPU timeslice, but the other did, the relative difference can be tremendous, while in fact both subs consume approximately the same number of CPU clocks. This can lead to very wrong conclusions. >>Also search the archives, about a year ago I've posted a subclass to DBI that >measures the SQL execution profiling. I'm not sure if Tim has this unctionality in the latest DBI. > > DBIx::Profile does a good job of this. cool :) -- _____________________________________________________________________ Stas Bekman JAm_pH -- Just Another mod_perl Hacker http://stason.org/ mod_perl Guide http://perl.apache.org/guide mailto:[EMAIL PROTECTED] http://ticketmaster.com http://apacheweek.com http://singlesheaven.com http://perl.apache.org http://perlmonth.com/
