Public SourceMage Discussion List

["Reapl Paratorn" <reapl AT reapl.net>]

Hi ppls,

Forgive me for sounding dumb, but couldnt it be really simple.

   Cast a sample benchmarking application
   Once successfully cast excute a run of the benchmarking application
   Just parse a date command before and after the execution run to get
within a second 'wall' time accuracy.

I for one don't give a iota for the user vs system vs whatever time. I care
about the time in my life it takes to perform the operation.

OK there can be all sorts of arguments over how to choose an appropriate
benchmarking tool and whether to compile with the systems optimisations or
to only use 386 ones if you want.

But surely there are some standard linux benchmarking tools we could choose
and give the user the option of which one to install and run during the
cast?

Now maybe it is the late night, or my use of natural substances...but it
seems this simple to me.

Now atm I don't have the time to make this spell myself, but is it this
simple?

Anyhoo have fun

Reapl
-----Original Message-----
From: sm-discuss-admin AT lists.ibiblio.org
[mailto:sm-discuss-admin AT lists.ibiblio.org]On Behalf Of
Phil/CERisE/KG6MBQ
Sent: Wednesday, 31 July 2002 9:13 AM
To: sm-discuss
Subject: Re: [SM-Discuss] Smallest Source Mage


Jonathan Evraire said:

>   A "sorcery-benchmark" spell would be pretty cool.  Anyone could do a
> "cast sorcery-benchmark" and the same repeatable steps would be taken on
> everyone's machine.
>
>   What does everyone think?
>
>   Jonathan Evraire
>   evraire AT tuwg.com

   Not a bad idea for fun, I suppose.
   I'm not sure what the idea is on implementing it (since I can't
think of an accurate way to do it).  I guess we'll call the time it
takes for the benchmark spell a compile-time unit, and assign variables
to the other spells and based on those, figure out the time?
   An example:
      'cast sorcery-benchmark' takes 4 seconds.
      The spell for unixcw is rated as 1.25, therefore the estimate
        would be that it takes 5 seconds to compile.
   I can think of a number of situations where it'll be O(M$)
inaccurate, though.

-Phil/CERisE

_______________________________________________
SM-Discuss mailing list
SM-Discuss AT lists.ibiblio.org
http://lists.ibiblio.org/mailman/listinfo/sm-discuss