Re: Seti@home.

["Vipul Mathur" <vipul@xxxxxxxxxxxxxxx>]

Sundeep Holani <sundeep@xxxxxxxxxxxxxxx> wrote:

> Lost Vipul's original mail, but anyway. Sounds peculiar to me that two
> units concurrently take the same time as one unit does alone. Try this,
> and see if it gives you some insight --
>
> 1) Run just one instance of SETI, and see how much processor utilisation
> 'top' shows.

CPU=99.8%    MEM=22.2%

> 2) Run two instances of SETI, and now see how much processor is used by
> each process.

CPU=49.7% to 51%    MEM=22.2%  for each process  (Two Processes running)

> If something's fishy... try for three instances, and generalize.

CPU=31% to 35%    MEM=22.2%    for each process  (Three Processes Running)

> 3) Run SETI at nice = -19, see how much processor time is used, and see
> how much time a workunit takes to be processed.

Time Taken = 16hr 30min (approx)

> Maybe all this info will tell you something interesting. Also, do you have
> the latest client? Static or dynamic?

I've tried both versions of the latest client.

> Apart from SETI, do you think the rest of the things work at satisfactory
> speeds? Compiled any kernels lately? How long did that take?
> -- Sundeep.

A compile of Kernel 2.2.16-3 took 3.5min for make bzimage and 5min for make
modules.

I have even confirmed with another person having a similar hardware config
and the time taken is the same as mine. It seems that SETI@home client does
not like the Coppermine processor much. This is indeed quite strange....
Needs further investigation...

Cheerio!

-- Vipul Mathur
                         ,_,
<vipul@xxxxxxxxxxxxxxx> (O,O)  http://www.vipulonline.com/
<vipul@xxxxxxxxxx>      (   )  The Geek shall inherit the earth.-Linus 5:5
-------------------------"-"----------------------------------------------
" #define QUESTION ( (bb) || !(bb) ) " - Shakespeare.