| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Active Research
Forum: OpenPandemics - COVID-19 Project
Thread: OpenPandemics - GPU Stress Test |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 781
|
|
| Author |
|
|
spRocket
Senior Cruncher Joined: Mar 25, 2020 Post Count: 277 Status: Offline Project Badges: 
|
Some are adjusting the app_config.xml to run 10 GPU tasks concurrently on one computer. If we go back to not many, I might as leave this alone at 1 GPU per computer running (default). I'm only running two tasks (that's enough to keep my GPU humming with few pauses), but I'm with you there. If we return to the scarce situation, there's no sense in tying up CPU slots. |
||
|
|
spRocket
Senior Cruncher Joined: Mar 25, 2020 Post Count: 277 Status: Offline Project Badges:
|
Still getting lots of stuck transfers, sometimes threatening to enter backoff hell. So far, it hasn't gotten to that yet. Those servers must be emitting smoke by now.
|
||
|
|
Drago75
Cruncher Joined: May 17, 2020 Post Count: 25 Status: Offline Project Badges: 
|
From my side it looks like the main issue at the moment is lack of bandwith to and from the server. My completed WU's don't go out anymore and also the download is impaired. I bet the server is steaming right now!
 |
||
|
|
Jim1348
Veteran Cruncher USA Joined: Jul 13, 2009 Post Count: 1066 Status: Offline Project Badges:
|
Yes, I knew I shouldn't have added that last GTX 1060. It is taking down the whole system.
|
||
|
|
Chooka
Cruncher Australia Joined: Jan 25, 2017 Post Count: 48 Status: Offline Project Badges:
|
I just got a lot of new workunits for my GPU, so the problems seem to be resolved. that's what I had though before, too; but then the download problems started again.I think it is because everyone ran out of WUs and requesting them all at the same time, should ease out soon. I ran out of WUs on my fastest machine too, full disclosure: I succumbed to editing the app_config.xml and run 16 concurrent GPU jobs on 2 NVIDIA cards and that does not even 'fill' them totally, but my main CPU does not have more threads so it will have to do, I also stopped processing iGPU jobs, it doesn't seem worth it in comparison to the NVIDIA cards (if iGPU is all you got or circumstances differ, go for it...). The embarrasing thing is I'm now racking up so many points in just a few days it's a big percentage of the points I've been gathering during about 15 years of contributing to WCG... It may be hard to go back to 87 point / WU for a CPU job compared to 1600 for a GPU job that also completes within a fraction of the time. I've wondered about this too. Will people who chase points give up on the cpu work and just fill their systems with copious amounts of OpenPandemics GPU work leaving less or no cores for CPU task. Some will...those who chase credits, other won't or they'll mix the tasks. Interesting to see I'm not the only one who's seeing bug issues with some Radeon cards.   |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Is this stress test affecting CPU work units also as on one of my computers that is only doing CPU work units i ran out of work to do and from time to time i am having trouble uploading some CPU work units or is this a different issue that is not being affected by the stress test.
|
||
|
|
Grumpy Swede
Master Cruncher Svíþjóð Joined: Apr 10, 2020 Post Count: 2224 Status: Recently Active Project Badges:
|
The stress test is stressing the entire system, so yes it also affects all other WCG projects, and CPU work.
----------------------------------------[Edit 1 times, last edit by Grumpy Swede at May 3, 2021 12:41:05 AM] |
||
|
|
AgrFan
Senior Cruncher USA Joined: Apr 17, 2008 Post Count: 376 Status: Offline Project Badges: 
|
Getting download/upload timeouts again tonight. I only run CPU work. Website is slow again also.
Can someone explain why a "stress test" is being conducted in a production environment? This would be totally unexceptable where I work. We'd all be fired if our production systems were impacted due to testing. |
||
|
|
DennyInDurham
Cruncher USA Joined: Aug 4, 2020 Post Count: 23 Status: Offline Project Badges: 
|
Getting download/upload timeouts again tonight. I only run CPU work. Website is slow again also. Can someone explain why a "stress test" is being conducted in a production environment? This would be totally unexceptable where I work. We'd all be fired if our production systems were impacted due to testing. Maybe because the "stress test". is doing real work, quite possibly more, computationally, than all the previous years of WCG combined. The occasional hiccups aren't hurting anything. |
||
|
|
biini
Senior Cruncher Finland Joined: Jan 25, 2007 Post Count: 334 Status: Offline Project Badges: 
|
I observe a 200 task limit. is your CPU+GPU equating to 200 tasks? Please define your test scenario. I see a limit of 100 - on a dual GPU machine. I also remember Uplinger writing that he was enforcing a 50 task limit.I surmise that limit was actually '50 tasks per GPU'. I have a dual GPU laptop, which has a cache of 400 WUs for some reason  I have not succesfully downloaded any GPU packets for Intel/AMD/Nvidia for couple of horus, is the stress test over yet? ---------------------------------------- [Edit 2 times, last edit by biini at May 3, 2021 4:28:29 AM] |
||
|
|
|