Just remarks

Gerard Beekmans gerard at linuxfromscratch.org
Mon May 27 17:18:02 PDT 2002


On Mon, May 27, 2002 at 07:21:13PM +0100, Ian Molton wrote:
> 200MHz Pentium with 8MB RAM - glibc compile: 4 days
> 200MHz Pentium with 128MB RAM - gglibc compile: 12 hours
> 
> therefore a 200MHz Pentium with 68MB RAM should take ~2 days, which is
> patently a load of horsesh*t.
> 
> you CANNOT average different machines and expect to have a meaningful
> result.
> 
> and an average of similar machines wont have enough variance in it to be
> worth doing.

Okay, I totally agree with your points. An averge of 2 days in this case
would be very unrealistic. Now one thing we can do is set limits on what
kind of machines should be used for these stats, ie. a minimum amount of
MHz's and a minimum amount of RAM, but if nothing else it won't give a fair
estimate since other much used machines aren't taken into account.

The other thing is just get the stats from one machine, but then how do we
decide what kind of machine will be used for the stats.

Perhaps the times relative to a static bash installation is the best thing
to do (we'll introduce a SBU for measurement - Static Bash Unit). All we
need to do is how well that works, get some people to calculate their
package's SBUs and see if they are about the same for everybody.

-- 
Gerard Beekmans
www.linuxfromscratch.org

-*- If Linux doesn't have the solution, you have the wrong problem -*-
-- 
Unsubscribe: send email to listar at linuxfromscratch.org
and put 'unsubscribe lfs-dev' in the subject header of the message



More information about the lfs-dev mailing list