[Beowulf] Re: Estimating cluster power consumption
Many of your questions may have already been answered in earlier discussions or in the FAQ. The search results page will indicate current discussions as well as past list serves, articles, and papers.
steve_heaton at iinet.net.au steve_heaton at iinet.net.auThu Dec 15 22:15:08 PST 2005
- Previous message: [Beowulf] Re: Estimating cluster power consumption (r reynoldson)
- Next message: [Beowulf] RE: Compare and contrast MPI implementations
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
G'day all I'd like to throw in my little bit of 1st hand experience on the topic. I've got a cluster of moboz. A new, "smart" 380W power supply runs one dual P3 and all 4x IDE HDDs. The other 3 nodes each have their own, old AT stype, 200W power supply (A$10 each). I was lucky enough to get a loan of the Oz equiv of a Kill-A-Watt and made an interesting discovery... The old 200W units, just powering the mobo pulled 75.5W / 0.5A. The newer 380W unit powering a mobo and the hard drives pulled 58W / 0.4A. Yep, the newer power supply, supporting the same load *and* the hard drives pulled considerably less power! Seems the age of the power supplies (ie. the associated technology) seems to make a considerable difference! In case it's of interest, including the 3GHz P4 head node: Idle: Power: 460W (552 Max) / Current: 2.57A Work (Nbody sims): Power: 510-585W (Max 654) / Current 3A (Max 4.18) Rating: 0.86 kWHr : $0.09 / hr Call it A$15 a week =) Cheers Stevo
- Previous message: [Beowulf] Re: Estimating cluster power consumption (r reynoldson)
- Next message: [Beowulf] RE: Compare and contrast MPI implementations
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
More information about the Beowulf mailing list
