Dual AMD systems in rackmount cases

Mark Hahn hahn at physics.mcmaster.ca
Mon Jun 24 11:19:09 PDT 2002


> > Does anybody have any experience with Dual AMD systems (MP2000) in 2 RU
> > rackmount cases?  Is heat generation a problem?  I am planning on

there seem to be quite a lot of vendors who seem confident enough
to sell dual AMD 1U's.  I don't see a lot of 2U's; my impression
is that the engineering for such boxes is strictly based on flow,
and that the increased dead-air capacity of 2U doesn't buy you much.

> > building a system with Tyan 2466 motherboards and 400 W power
> > supplies.  Should I use 3RU cases?

if you can afford the lower density, go for it!  is 3U high enough
to handle normal upright PCI cards?  risers for low-profile cases
appear to be a source of some difficulty...

racksaver, for instance, seems comfortable letting you configure 
a dual athMP/2100 right now.  apparently the thoroughbred version
will save just a handful of wats.

> As important as these two components is being able to provide ambient
> air at around 65F (or cooler, of course:-) on the air intake side of the
> cases and to get rid of the 75F air coming out of the air outflow side

75 seems a little conservative on exhaust temp, though I suppose it 
depends hugely on the volume of air you're talking about.  my machineroom
(currently 30 KW of alphas, but soon to have 5-10KW new ia32's)
registers around 60-65 in the cold ducts, and around 85 in return.
and, importantly I think, there's a pretty clear and strong airflow
pattern.

> Dual AMD's will just plain crash when they get hot, and they get hot
> very, very easily as they draw about 150 W sustained under load.  In our

AMD's pushing vendors to implement thermal management, but I'm not 
clear on how well it all works.

> new server room, our 2U cases were not originally getting enough, cold
> enough, air and the temperature at the case rears was up in the 80's.

it depends on how "dilluted" that 80F is - by itself, 80F is still
quite cool: suppose you have a CPU dissipating 70W, and your HS/fan
will need 20-25C of thermal gradient.  since you don't want your CPU
at its rated max (usually 90C), but say 60C, you need fan ambient
to be 95F or less.

> KW/meter^3 seems unwise -- any sort of bobble in your cooling system
> will rapidly cause your stack to reach furnace temperatures, and even

ah, well, I think you have to start with assuming *something* stays up.
I'm assuming that I can shut down my space heaters if (when) our chilled
water goes out (again)...

regards, mark hahn.




More information about the Beowulf mailing list