Skip to Content.
Sympa Menu

sm-discuss - Re: [SM-Discuss] Benchmarking Spell?

sm-discuss AT lists.ibiblio.org

Subject: Public SourceMage Discussion List

List archive

Chronological Thread  
  • From: Andrew Stitt <afrayedknot AT thefrayedknot.armory.com>
  • To: sm-discuss AT lists.ibiblio.org
  • Subject: Re: [SM-Discuss] Benchmarking Spell?
  • Date: Sun, 4 Aug 2002 02:34:33 -0700

On Sun, Aug 04, 2002 at 01:02:08AM -0700, Phil/CERisE/KG6MBQ wrote:
> So...I missed something..
>
> Why not just use time?

i think the problem here is a misunderstanding of what the time command
does, a cursory reading of the man page even at 2am says

"These statistics consist of (i) the
----> elapsed real time between invocation and termination, (ii)
the user CPU time (the sum of the tms_utime and tms_cutime
values in a struct tms as returned by times(2)), and (iii)
the system CPU time (the sum of the tms_stime and
tms_cstime values in a struct tms as returned by
times(2))."
<rant>
im not sure why we need to use our scripts to measure the time because
its obviously going to have a larger delta of inacuracy then using
built in system calls, which simply allocate some space in memory, copy
a few data values into place and return. if one reads through times(2)
it does its magic right after the calls to wait and waitpid, which of
course are called once the shell knows that the program has ended. If
we were to run a seperate command before and after the process ends we
have added inacuacies waiting for the program to either a) get swaped or
possibly loaded from fs into physical ram b) have it run almost immediatly
because it wasnt swapped out and remained cached. The time it takes to
load and run date after the compile counts unfairly and unnecesarily in
the benchmark.
</rant>




Archive powered by MHonArc 2.6.24.

Top of Page