Folding efficiency. KWH:WU
Posted: Sun May 11, 2008 2:32 am
I've brought up this concept before a couple of years ago and I would like to visit it again. After recently
reading an article on the newer servers they mentioned that the power consumption counts for 50%
of the total user costs (the rest is hardware purchasing and maintenance). It behooves many a server
farm to upgrade simply because the reduced power consumption pays for the purchase of new hardware
That said, I would like to think there is a standard of efficiency, like Points/Watt Hour. I've seen the
Kill A Watt plug in power meter and I now have one. Should we collectively make some stats and
compare the overall efficiency of individual rigs to massive clusters. Time standards are also an
idea. 24 hours, 7 days and 30 days seem to be good ideas. Maybe we can make a standard or
benchmark updated yearly (or every 6 months) to give us a hint when to upgrade.
Just playing around, I was testing my beater and test rig (and old Epox 8KTA3 with 2400+ @ 2400mhz)
and it (not including monitor) is pulling 170 watts while my laptop (Vostro 1500) is pulling only 35watts.
And the Laptop is folding at around 6 times the rate (give or take).
I'd still like a more accurate way of measuring PPD and I'm open to suggestions. I can start and stop
the Kill A Watt like a stopwatch but the PPD is just an estimate.
Just an idea. Is anyone willing to participate?
CE
reading an article on the newer servers they mentioned that the power consumption counts for 50%
of the total user costs (the rest is hardware purchasing and maintenance). It behooves many a server
farm to upgrade simply because the reduced power consumption pays for the purchase of new hardware
That said, I would like to think there is a standard of efficiency, like Points/Watt Hour. I've seen the
Kill A Watt plug in power meter and I now have one. Should we collectively make some stats and
compare the overall efficiency of individual rigs to massive clusters. Time standards are also an
idea. 24 hours, 7 days and 30 days seem to be good ideas. Maybe we can make a standard or
benchmark updated yearly (or every 6 months) to give us a hint when to upgrade.
Just playing around, I was testing my beater and test rig (and old Epox 8KTA3 with 2400+ @ 2400mhz)
and it (not including monitor) is pulling 170 watts while my laptop (Vostro 1500) is pulling only 35watts.
And the Laptop is folding at around 6 times the rate (give or take).
I'd still like a more accurate way of measuring PPD and I'm open to suggestions. I can start and stop
the Kill A Watt like a stopwatch but the PPD is just an estimate.
Just an idea. Is anyone willing to participate?
CE