[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]

Re: [f-cpu] Running long things (was: Re: Yet Another Upload)



Hi,

On Wed, Apr 23, 2003 at 04:57:20AM +0200, Yann Guidon wrote:
> hi !
> 
> there is some scripting work in perspective ....

I am not sure I understand. "scripting work" to automate such tests?

[snip]
> >
> it could also be used to testbench the software, so we'd need the 
> description of the machines
> (if you go further). today's computer have big differences and can show 
> a 10x runtime difference
> with the same software ....

About 30 dual PIII 500Mhz 256MB, about 30 PIII 733Mhz 256MB, about
20 dual PIII 600Mhz 1-2GB RAM (these are meant for big memory stuff),
about 20 PIII 1000Mhz 256MB, about 10 (our newest) PIV 2000Mhz 512MB
RDRAM (these are usually the most loaded).

Of course we also have some other, non-public (high uptime) machines.
devik said he has a p2-350. If that represents the average machine
of a f-cpu developer, I guess I can help even with few stronger ones.
I guess this might be close to reality, since until a few weeks ago
my strongest home machine was a Celeron 300Mhz 128MB (sometimes OCed
to 450). Now my strongest is a duron 900Mhz, also not such a strong
machine.

> 
> Concerning the installation of Free Software, you can read the 
> VHDL_HOWTO that
> is available in the F-CPU source tarball. you will have to read the 
> chapters about
> Simili and Vanilla, and if you are successful, try to install GHDL.

I Will try.
Where is the latest tarball? Is there already a fixed place for this?

> 
> >It will be even better if such runs can be split to small parts, since
> >most of the machines dual-boot between Linux and Windows (and I do not
> >intend to use them while they are running Windows).
> > 
> >
> small parts ?
> i don't know what this means in this context.
> Some tests last 10 seconds, but Michael's tests are usually extremely 
> long with Vanilla
> (which itself is already slow software).

I guess you won't need my help for the 10 seconds stuff. What I hope
is that the long stuff (e.g. above an hour) can be cut to e.g. 1 hour
parts without too much overhead. This way, if a machine is rebooted
in the middle of such a 1 hour job, we lose at most 1 hour, and with
a sufficiently good batch scheduler (a hand-written script or something
like PBS which we already use) resubmit it without manual intervention.

> 
> thank you for your help,
> 
> >	Didi
> > 
> >
> YG
> 
> *************************************************************
> To unsubscribe, send an e-mail to majordomo@seul.org with
> unsubscribe f-cpu       in the body. http://f-cpu.seul.org/

	Didi

*************************************************************
To unsubscribe, send an e-mail to majordomo@seul.org with
unsubscribe f-cpu       in the body. http://f-cpu.seul.org/