Page 1 of 1

New to FAH; general/COVID-19 questions

Posted: Thu Mar 19, 2020 8:32 pm
by VoiceCoil
Hi -

I used to participate in Seti@Home years ago and when I learned FAH was being used for COVID-19 I decided I'd start burning some cycles on FAH.

At the moment I have one instance running on Gentoo Linux running the in-repo FAH 7.5.1 client and I have several more I'm able to stand up. The one instance I'm running is on a machine with an nVidia GeForce GT 610 GPU; I indicated I have a GPU at service configuration time and I have "Any disease" selected and "Full" selected for Speed. I've been watching this one instance's web control page and it's now on its third WU since last night; all three have been for Project 13824, which has to do with myosins. Some questions:

1) Will my machine(s) be getting COVID-19-related work units eventually?
2) The web control page shows that I have a GPU but its indicator is yellow and its percentage stays fixed at 0.00%. Is this because this project doesn't use GPU or is something broken on my end? This machine does have the correct proprietary nVidia driver and is using it for display.
3) i notice that while these WUs run, the machine is lighting up three of four threads of its dual-core CPU. Is t-1 how this typically goes?

Thanks!

Re: New to FAH; general/COVID-19 questions

Posted: Thu Mar 19, 2020 9:22 pm
by Joe_H
1 - They should eventually, a number of CPU related projects are in the process of being rolled out to add to the ones already started

2 - The servers for the GPU work are getting particularly heavy traffic, so connections and downloads of WU's from them is still a bit of a chance. New Work Servers just came online yesterday evening and they have been moving some projects to them. The assign rate has gone up by a lot just recently.

That said, your GT610 may not be able to process work within the deadlines. It is hard to say, there were several variants sold by nVidia and the other manufacturers using their chips, and there is some variation in processing power.

3 - The client software reserves one CPU core/thread for GPU processing, it is needed to prepare data and transfer it to and back from the GPU. So 3 of 4 threads in use for Full is normal. Medium would use one less thread.