[Beowulf] Servers Too Hot? Intel Recommends a Luxurious Oil Bath

Jeff Johnson jeff.johnson at aeoncomputing.com
Wed Sep 5 11:40:56 PDT 2012


On 9/5/12 11:33 AM, Robert G. Brown wrote:
> On Wed, 5 Sep 2012, Jeff Johnson wrote:
>
>> for nuke reactors) escapes me. If you combined the GreenRevolution
>> approach with looping sea water to exchange the heat you would get
>> pretty close to a PUE of 1. Especially if servers and their components
>> were redesigned for submerged operation as the original article 
>> mentioned.
>
> Personally, designing for immersion in salt water is, um, not a sane
> choice.  Salt water has all of these annoying properties -- it's a good
> conductor of electricity, it's corrosive as all hell, it is filled with
> all of these really annoying animals and plants that like to grow on
> warm surfaces especially, and the pressure increases as you descend in
> it at 1 atmosphere per 10 meters) (to name four, two of which I have to
> deal with on a regular basis just maintaining a boat that sits on the
> surface in salt water for a week.  I'm guessing barnacles would
> interfere with heat transport...;-)
>
> However, using seawater in the pacific to dump heat into carried there
> by cooling fluid (e.g.) is quite reasonable.
>
>    rgb
I wasn't suggesting sea water as an immersion medium for active 
components. I was using sea water as a flow through medium for a 
radiator cooling the mineral oil used as an immersion medium. Use sea 
water instead of air to cool the oil. No sea water actually touching 
components. Exchange the heat from the oil to the water.

--Jeff



-- 
------------------------------
Jeff Johnson
Co-Founder
Aeon Computing

jeff.johnson at aeoncomputing.com
www.aeoncomputing.com
t: 858-412-3810 x101   f: 858-412-3845
m: 619-204-9061

/* New Address */
4170 Morena Boulevard, Suite D - San Diego, CA 92117



More information about the Beowulf mailing list