m.marien wrote:
> "Jean-David Beyer" <jdbeyer@exit109.com> wrote in message
 
>
>
> That's a fairly heavy load average. My K6II 400 system went way over 25 wh
en
> I first upgraded but when I added memory it went back down to zero. If I
> understand it correctly, you have 4 threads waiting for processor time at
> any given time ? Are you running some SETI's or do you just type really fa
st
> ?

I type really fast! ;-)

Actually, I am running BOINC that has two setiathome and two
climateprediction running. Luckily at nice level 19. When it really gets
hairy is when I populate an IBM DB2 UDB V8.1.7 database which runs over a
dozen server processes at once. Then the load average can exceed 10. But
the BOINC stuff tends to get out of the way at nice 19.

The BOINC stuff screws up the processor caches, though, and makes the
database slow down a lot, even though, nominally, it gets all the
processor time it wants. But it does not get cache hit ratio it needs, so
the disk IO goes at about half speed or a little less. With BOINC turned
off, I get about 12 Megabytes/second to the hard drives, but it drops to
around 4 when BOINC is running those applications, even though they are at
nice 19.
>
>  21:45:43  up 54 days, 12:53,  1 user,  load average: 0.04, 0.01, 0.00
>
> This is FC1 on a file server, but everybody has gone home. Last shutdown w
as
> to add the memory I think.
>
>


--
.~.  Jean-David Beyer           Registered Linux User 85642.
/V\                             Registered Machine   241939.
/( )\ Shrewsbury, New Jersey     http://counter.li.org
^^-^^ 09:35:00 up 7 days, 20:39, 4 users, load average: 4.30, 4.23, 4.24

[ Post a follow-up to this message ]