[Beowulf] 90nm opteron power usage

[Vincent Diepeveen]

Note sure what the distributed.net client is. Is it doing FFT matrix
multiplications using loads of i/o? Have a link?

The weakest path in the cpu's sometimes you can get close to
when multitasking with some FFT codes, like prime95 is. 
www.mersenne.org. That's not busy with i/o though, nor graphics. 

But if you combine it with some streaming software and some software
that puts pressure on the videocard by streaming opengl/directx data to
videocard from harddisk (in itself that further eats little cpu time);
additional one timing application to put pressure on possible hardwaretimer, 
meanwhile reading data from cdrom and floppydrives,
that all should get you far. 

Not to forget an application that sometimes opens and
closes the cdrom drives automatically (if that's in the nodes)  :)

Please note if i recall well, those opterons have a TDP of 89 watt,
so your 270-300 watt measurement most likely has to do with other
parts in the computer than just the cpu. 

So the improvement you refer to most likely has to do with some harddisk
that put itself to sleep or something.

Do both machines have the same PSU?

Vincent

At 10:59 AM 8/8/2005 -0700, Bill Broadley wrote:
>
>I managed to hook up my kill-a-watt to a dual opteron 2.6 GHz, 4GB pc3200,
>and a 73GB scsi disk (a sun v20z).  The highest power consumption I
>could find was was running 1 distributed.net client per cpu.
>
>The results was 197-198 watts and 200-201 VA.  I'm quite impressed,
>our 2.2GHz version of the same box is in the 270-300 watt range.
>
>Has anyone found a (free/open) code that manages higher power consumption
>than the distributed.net client?
>
>Kudos to AMD for such a big improvement, I thought it might be useful
>for those planning cooling or power for any of the newer 90nm AMD chips.
>
>Anyone have similar numbers for a new dual nacoma/64 bit xeon?
>
>-- 
>Bill Broadley
>Computational Science and Engineering
>UC Davis
>_______________________________________________
>Beowulf mailing list, Beowulf at beowulf.org
>To change your subscription (digest mode or unsubscribe) visit
http://www.beowulf.org/mailman/listinfo/beowulf
>
>