F@H on University Lab Comps

[bjh22]

I was thinking about approaching the University of Pittsburgh Tech Department director about installing F@H on the various computers around campus, most of which run P4 HT or C2D Intel processors. The only thing I was unsure about is how effective this might be, as the logged in time on many computers will be brief. I was looking for some confirmation about how F@H runs. As I understand it the F@H client assigns a work unit, and transfers the completed work unit back to the main server. Assuming this is correct, will any brief amount of work done while the computers are logged on actually be contributed, or would it be a waste because whole WU's will not often be completed?

[John Naylor]

So long as the computers remain switched on but not logged on you can set the CPU clients to run as a service, which means that processing continues while the computer is sat at the log in screen as well as while logged in... is this the case?

Welcome to the forums

[bjh22]

Thanks for the input, didn't know it would run continuously as a service. Hopefully I'll have time to approach the director tomorrow about it. If it's a success I'll see if they can set it up as a single team, see what kind of numbers so many computers can put out.

Thanks for the welcome, and any advice as to which client to run, I'm assuming regular client for the P4 and SMP for the C2D machines. I believe most of the comps run onboard video, so GPU client is probably not an option.

[John Naylor]

I personally would run the regular client on both (just run 2 copies on the C2Ds... ask if you do not know how ) - the SMP client is less temperamental than it used to be but I (and most other people on here) would not recommend rolling it out to computers that cannot be monitored regularly, because it is not a finished product .

[toTOW]

Plus SMP doesn't (officially) support service mode ... so the good old uniprocessor client is the right choice (two copies on C2D, one copy on P4 HT)

[ppetrone]

Hi bjh22,

From the science side of Folding@Home, we are really excited about this perspective. It sounds awesome.
Here in the lab, we keep our fingers crossed. Even if it does not work out, thank you for giving it a chance.

Paula

PS: Please keep us posted!