Advanced search

Message boards : Server and website : no tasks available for short/long runs (according to the event log)

Author Message
Nethlek
Send message
Joined: 4 Jul 18
Posts: 8
Credit: 55,542,350
RAC: 479,183
Level
Thr
Scientific publications
wat
Message 49826 - Posted: 10 Jul 2018 | 13:09:31 UTC

im fairly new to gpugrid. did the project really run out of tasks or is there a problem on my end? if it really did run out, when can we expect new tasks and is this a frequent occurrence? cheers

Jim1348
Send message
Joined: 28 Jul 12
Posts: 640
Credit: 1,207,282,694
RAC: 96,054
Level
Met
Scientific publications
watwatwatwatwatwatwatwatwatwat
Message 49827 - Posted: 10 Jul 2018 | 14:40:13 UTC - in response to Message 49826.
Last modified: 10 Jul 2018 | 14:41:05 UTC

Yes, they really run out frequently. There are usually more longs than shorts, but they run out too.

If you run Linux, they have plenty of CPU Quantum Chemistry jobs though. Eventually, they will have a Windows version.

JoergF
Avatar
Send message
Joined: 20 Apr 15
Posts: 273
Credit: 836,357,554
RAC: 1,455,835
Level
Glu
Scientific publications
watwat
Message 49829 - Posted: 10 Jul 2018 | 14:47:37 UTC - in response to Message 49826.
Last modified: 10 Jul 2018 | 14:50:16 UTC

im fairly new to gpugrid. did the project really run out of tasks or is there a problem on my end? if it really did run out, when can we expect new tasks and is this a frequent occurrence? cheers


see also this thread
https://www.gpugrid.net/forum_thread.php?id=4443

Nothing wrong on your side... sometimes they run out of jobs indeed, but there are increased efforts made to fix this. The problem is that all this data needs to be reviewed and therefore it comes to staff shortages time and again. Having said this, GPUGRID is really one of the projects that produces high quality papers, so we should have patience here.

GPU jobs are rare sometimes, but there are plenty of CPU jobs (Linux only) at present waiting for someone to crunch them.

Edit: Jim already touched on this ... our messages must have crossed paths. :)
____________
I would love to see HCF1 protein folding and interaction simulations to help my little boy... someday.

Nethlek
Send message
Joined: 4 Jul 18
Posts: 8
Credit: 55,542,350
RAC: 479,183
Level
Thr
Scientific publications
wat
Message 49832 - Posted: 10 Jul 2018 | 15:37:45 UTC

alright, thank you. how long does it usually take for new tasks to be added?
(im on windows and im not interested in cpu tasks anyway, i run cosmology@home for that)

Jim1348
Send message
Joined: 28 Jul 12
Posts: 640
Credit: 1,207,282,694
RAC: 96,054
Level
Met
Scientific publications
watwatwatwatwatwatwatwatwatwat
Message 49833 - Posted: 10 Jul 2018 | 16:50:36 UTC - in response to Message 49832.

alright, thank you. how long does it usually take for new tasks to be added?

They are doing pretty well for the longs; usually less than a day before new ones. As for the shorts, they come and go at more irregular intervals.

Nethlek
Send message
Joined: 4 Jul 18
Posts: 8
Credit: 55,542,350
RAC: 479,183
Level
Thr
Scientific publications
wat
Message 49835 - Posted: 10 Jul 2018 | 17:42:31 UTC

ok then, thank you.

PappaLitto
Send message
Joined: 21 Mar 16
Posts: 413
Credit: 2,924,744,691
RAC: 4,432,185
Level
Phe
Scientific publications
watwat
Message 49837 - Posted: 10 Jul 2018 | 18:53:35 UTC - in response to Message 49835.

ok then, thank you.

As you have a 1080, I would suggest in your profile to go to GPUGRID preferences and uncheck "ACEMD short runs" and check "If no work for selected applications is available, accept work from other applications?" This will provide you with a default path of requesting long WUs because you have a fast GPU and if there is work for short but no work from long you can still receive some work.

Nethlek
Send message
Joined: 4 Jul 18
Posts: 8
Credit: 55,542,350
RAC: 479,183
Level
Thr
Scientific publications
wat
Message 49838 - Posted: 10 Jul 2018 | 19:05:56 UTC

but if neither short nor long runs are available, wouldnt i end up getting cpu tasks instead (which im trying to avoid) or is unchecking the "use cpu" box enough to prevent that? also im curious, where did you see what gpu im using?

Profile Chilean
Avatar
Send message
Joined: 8 Oct 12
Posts: 94
Credit: 322,799,073
RAC: 1,792,901
Level
Asp
Scientific publications
watwatwatwatwatwatwatwatwat
Message 49841 - Posted: 10 Jul 2018 | 21:28:10 UTC - in response to Message 49838.
Last modified: 10 Jul 2018 | 21:28:35 UTC

but if neither short nor long runs are available, wouldnt i end up getting cpu tasks instead (which im trying to avoid) or is unchecking the "use cpu" box enough to prevent that? also im curious, where did you see what gpu im using?


If I click on your name, I can see your stats and your computers. You can see my computers as well if you click on my name. You can hide this if you want under GPUGRID preferences on your profile. No IP is shows. Just technical information.

If you uncheck the "Use CPU" box, you won't get CPU jobs even if you check the "If no work for selected applications is available, accept work from other applications?" box.
____________

Nethlek
Send message
Joined: 4 Jul 18
Posts: 8
Credit: 55,542,350
RAC: 479,183
Level
Thr
Scientific publications
wat
Message 49850 - Posted: 12 Jul 2018 | 13:23:25 UTC - in response to Message 49841.

i see, thanks.

Post to thread

Message boards : Server and website : no tasks available for short/long runs (according to the event log)