Is this linux or windows ? In linux you have to do this to get the GPU to be recognized:Crunching for the TeAm again after 15 years idle!
Need a little help with the Folding client, I set it up with my i5, shows slot ID 00 4 cores, fine. I added a GT640, it's recognized in System Info, but no additional slots from it. Is the card inadequate to crunch WUs? And, does Folding just run one WU with all 4 cores? BOINC client shows different WUs running on each core.
Also, I'm running BOINC as well (Rosetta, LHC), the manager shows 4 threads running, but no additional threads when I installed the GT640. Again, card not good enough? I also have an i7 box with a 1050, it is definitely using the GPU.
Wish I had an i7 back when I was crunching RC5
Thanks.
JC
Is this linux or windows ? In linux you have to do this to get the GPU to be recognized:
cd /var/lib/fahclient
sudo wget http://fah-web.stanford.edu/file-releases/public/GPUs.txt
I think windows does that automatically.
Not sure if there is more to it, but GeForce GT 640 is definitely whitelisted. (Looking at the file GPU.txt in the client's data directory, to which Mark referred to.)Need a little help with the Folding client, I set it up with my i5, shows slot ID 00 4 cores, fine. I added a GT640, it's recognized in System Info, but no additional slots from it. Is the card inadequate to crunch WUs?
Yes, the science application for folding on CPUs, fahcore_a7, is multithreaded. (There are also some multithreaded applications in the BOINC world, e.g. LHC's ATLAS subproject.)And, does Folding just run one WU with all 4 cores? BOINC client shows different WUs running on each core.
@Markfw, link fixed.
I was just reading an article in Wired, an interview with the science consultant for the movie Contagion, that I'd like to quote from:
That's what we're working on here.
It is possible to configure more than one CPU slot on the same machine, and to give them a fraction of the total of available processor threads each. But this would give fewer PPD in total than a single CPU slot which takes all processor threads, due to the Quick Return Bonus effect.
If you have one or more GPUs, reduce the CPU slot to fewer processor threads, since the GPU application occupies a processor thread too. But with just four cores in a CPU, just pause or remove the CPU slot altogether; it's performance is two magnitudes below a typical GPU, and a busy CPU folding slot may even degrade performance of the GPU folding slot more than the small amount of PPD it adds.
Does any one else have a problem looking at the stats page? I get a "Bad Gateway" error every time I try to look at it.
Last night the F@H site stats that had been just barely working stopped working. So I implemented a backup plan that gets stats for my stats graph from https://folding.extremeoverclocking.com/ which apparently gets its stats from large flat files.
I hope F@H gets the resources they need set up soon.
Found some nice stats that work here: https://folding.extremeoverclocking.com/team_summary.php?s=&t=198Folding for Team 198, Is there a scoreboard or something? Where can i get some recognition? lol
Yes. VA not watts.Is it 675W, or 675VA? When talking about AC loads, the two are NOT the same. I would expect a 675VA UPS to have around 350-360W load capacity. Actual runtime dependent upon battery size. (Wattage capacity depends on inverter.)
For a "folding rig", I don't recommend anything below 1350VA or 800W actual load capacity. (High-end gaming rig too.)
How do you know you're getting Covid-19 work and not something else ?
Yeah, you're going to need a "bigger" UPS. (Larger inverter wattage capacity. Bigger battery means longer runtime, too, they generally go hand-in-hand.)Yes. VA not watts.
If you have one or more GPUs, reduce the CPU slot to fewer processor threads, since the GPU application occupies a processor thread too. But with just four cores in a CPU, just pause or remove the CPU slot altogether; it's performance is two magnitudes below a typical GPU, and a busy CPU folding slot may even degrade performance of the GPU folding slot more than the small amount of PPD it adds.
Regarding the GPU slot, leave gpu-index at the default of -1.Not sure on this....you mean reduce the CPU threads to say, 2, and set gpu-index to 0?