Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
World Community Grid Forums
Category: Active Research Forum: OpenPandemics - COVID-19 Project Thread: Work unit availability |
No member browsing this thread |
Thread Status: Active Total posts in this thread: 822
|
Author |
|
El_Pinguino
Cruncher Joined: May 15, 2020 Post Count: 8 Status: Offline Project Badges: |
After 3 days of waiting, I finally got one GPU work unit. Once the GPU started work the computer promptly crashed. Hard. I just happened to be viewing boinctui when this happened. Since the computer is co-located 1/2 way around the world from me I had to wait for the onsite tech to restart the computer.
When the GPU is in the computer it crashes upon reboot. With the GPU out the computer runs fine. So with all of the logistics of trying to resolve the problem from so far away so I can run 1 work unit every three days I decided to take the GPU out and stick with CPU for now. I'll save some money until this is a more mature rollout as the Quadro RTX 4000 was $98 USD per month. |
||
|
Martin Schnellinger
Advanced Cruncher Joined: Apr 29, 2007 Post Count: 123 Status: Offline Project Badges: |
Hello El_Pinguino,
if you post the content of BOINC's log just before the crash, some helpful cruncher could maybe find out the technical reason for the crash and give you some advice how to fix the problem. It would be positive for the project if your GPU could be used, as GPUs are so efficient Maybe try posting the content of the log just before the crash. Greetings Martin |
||
|
El_Pinguino
Cruncher Joined: May 15, 2020 Post Count: 8 Status: Offline Project Badges: |
I would post the logs .... if I knew what to post. Of course it is important for developers to know what is happening but I just don't know what to post.
Personally, I am not interested in getting this GPU running right now as there is no work available to keep it busy and justify a $98 rental fee. That will probably change in the future and at that point, I'll have it reinstalled. But even then, as the computer is located 1/2 way around the world from me I will wait until the bugs are pretty well fixed. If this box was in my home it would be a different story. |
||
|
biini
Senior Cruncher Finland Joined: Jan 25, 2007 Post Count: 334 Status: Offline Project Badges: |
After 3 days of waiting, I finally got one GPU work unit. Once the GPU started work the computer promptly crashed. Hard. I just happened to be viewing boinctui when this happened. Since the computer is co-located 1/2 way around the world from me I had to wait for the onsite tech to restart the computer. When the GPU is in the computer it crashes upon reboot. With the GPU out the computer runs fine. So the whole OS crashed? Any idea if the boot will complete with GPU and crashes then or crash in the middle of the boot? FWIW For me it does not sound like boinc/wcg bug but thermal/power issue (if the boot completes and the boinc client starts) But yes, for that kind of money, there's no point. I've been buying a lot of second hand rigs to keep my house warm for pennies. If I was up for a rental, I'd try some cheap multiCPU like amazon spot instances. ---------------------------------------- [Edit 3 times, last edit by biini at Jul 26, 2021 1:10:58 PM] |
||
|
El_Pinguino
Cruncher Joined: May 15, 2020 Post Count: 8 Status: Offline Project Badges: |
The computer required a person to physically turn it on. Tech support tried to use KVM but could not reach it. I have an interface where I can power on / power off. Neither worked. This is one reason not to pursue this problem. If the box were in my living room I'd be accepting the challenge.
The box completely crashed needing a physical reboot. The GPU was taken out and the box booted fine. The GPU was reinstalled and would not complete the boot. BOINC runs at boot time so the crash comes when the GPU starts working on a work unit. They took the GPU out and the box booted fine. I requested they leave the GPU out. The power issues were resolved. Initially the GPU was installed with inadequate power. I had seen a post where this happened to someone else though I can't find the link now. It would have been a nice box. AMD Ryzen 9 3900X with a nVidia Quadro RTX 4000. I'll keep the CPU working for a while longer. Corona virus is far from being resolved. |
||
|
kittyman
Advanced Cruncher Joined: May 14, 2020 Post Count: 140 Status: Offline Project Badges: |
Having a nice long run with the kitties' kibble bowl keeping full.
----------------------------------------Meow! |
||
|
Jim1348
Veteran Cruncher USA Joined: Jul 13, 2009 Post Count: 1066 Status: Offline Project Badges: |
You must be on an approved list. They are partial to kitties.
I try every day, and haven't gotten any in a couple of weeks. That is really no problem, I can just do Folding for biomedical, or MLC for AI research. There is no point banging away on useless requests. I am glad they have more than enough support. |
||
|
kittyman
Advanced Cruncher Joined: May 14, 2020 Post Count: 140 Status: Offline Project Badges: |
Oooo, the kitties are quite happy to be on an approved list!!!! LOL.
----------------------------------------Keep in mind that they are playing with only a single old GTX980, so it doesn't take as much to keep them busy as somebody with more or more powerful GPUs. But the kitties are pleased to be able to contribute at their slow steady pace. The real hope is that researchers can use the crunched information to possibly come up with a 'magic bullet' that will kill not only the current version of Covid, but also the dang variants that are propagating like mad now. |
||
|
erich56
Senior Cruncher Austria Joined: Feb 24, 2007 Post Count: 294 Status: Offline Project Badges: |
From what I see, the availability of GPU tasks has gone back markedly since yesterday :-(
|
||
|
Richard Haselgrove
Senior Cruncher United Kingdom Joined: Feb 19, 2021 Post Count: 360 Status: Offline Project Badges: |
They're coming in now! The time of the 'big morning launch' (UK time) seems to vary by an hour or two.
|
||
|
|