World Community Grid - View Thread - GFAM: Overclocking caused compute error. O/C now off, how do I fetch new work?

World Community Grid Forums

Category: Completed Research

Forum: GO Fight Against Malaria

Thread: GFAM: Overclocking caused compute error. O/C now off, how do I fetch new work?

Quick Go »

No member browsing this thread

Thread Status: Active
Total posts in this thread: 6

[ ]

Author

This topic has been viewed 2598 times and has 5 replies

Former Member
Cruncher
Joined: May 22, 2018
Post Count: 0
Status: Offline


GFAM: Overclocking caused compute error. O/C now off, how do I fetch new work?

Thought I'd try running BOINC/GFAM with my AMD Phenom II X4 940 CPU at 3.45Ghz instead of 3.00.
It's got a fairly hefty heat sink, core never tops 40 deg (limit is 59, IIRC), and it's never barfed over some pretty hefty video encodes.

But after 5 mins, it gave a "compute error" or similar, and stopped. So I reset the project, turn off overclocking, and requested an update over and over and over again - now, looking in the logs, it says: (reached daily quota of 4 results)

17/11/2011 10:30:21 | World Community Grid | Requesting new tasks for CPU and ATI GPU
17/11/2011 10:30:24 | World Community Grid | Scheduler request completed: got 0 new tasks
17/11/2011 10:30:24 | World Community Grid | No work sent
17/11/2011 10:30:24 | World Community Grid | No work is available for GO Fight Against Malaria
17/11/2011 10:30:24 | World Community Grid | (reached daily quota of 4 results)
17/11/2011 10:30:36 | World Community Grid | Resetting project
17/11/2011 10:30:40 | World Community Grid | Sending scheduler request: To fetch work.
17/11/2011 10:30:40 | World Community Grid | Requesting new tasks for ATI GPU
17/11/2011 10:30:42 | World Community Grid | Scheduler request completed: got 0 new tasks

I've looked in my BOINC, GFAM and WCG prefs and I can't find any way to collect new work today, one of 4 days I'll be able to dedicate my PC solely to this while I do some massive video uploads. I don't even want new work, I just want to restarted the work which bombed out.

The quicker I can find a solution, the quicker I can get working. Thanks!

[Nov 17, 2011 11:31:37 AM]

Former Member
Cruncher
Joined: May 22, 2018
Post Count: 0
Status: Offline


Re: GFAM: Overclocking caused compute error. O/C now off, how do I fetch new work?

Hi,

Look in My Grid > Result Status page. How many are listed with an "error" state?

At 00:00:01 UTC you'll be allowed another WU to test if your machine is reliable again. If that's completes fine, more are allowed.

Workaround: Utterly and completely uninstall, remove BOINC and all register traces. Then install clean. WCG Servers might recognize your device as a new one. Then you get 5 tasks per core to proof that your in Robert De Niro's circle of trust.

--//--

[Nov 17, 2011 11:37:43 AM]

Former Member
Cruncher
Joined: May 22, 2018
Post Count: 0
Status: Offline


Re: GFAM: Overclocking caused compute error. O/C now off, how do I fetch new work?

P.S. I've found some heavy duty work on my Linux box to cause the CEP2 to bomb (heartbeat), so I've actually set it to pause if the non-BOINC use is > than 20%. DSFL would just resume from previous checkpoint if the device is very busy, but the message and result log would tell what the error cause was. GFAM uses the same research engine as DSFL, so I'm considering the same error susceptibility.

--//--

----------------------------------------
[Edit 1 times, last edit by Former Member at Nov 17, 2011 11:41:45 AM]

[Nov 17, 2011 11:40:42 AM]

Former Member
Cruncher
Joined: May 22, 2018
Post Count: 0
Status: Offline


Re: GFAM: Overclocking caused compute error. O/C now off, how do I fetch new work?

Look in My Grid > Result Status page. How many are listed with an "error" state?

None - GFAM isn't even listed there yet.

At 00:00:01 UTC you'll be allowed another WU to test if your machine is reliable again. If that's completes fine, more are allowed.

Workaround: Utterly and completely uninstall, remove BOINC and all register traces. Then install clean. WCG Servers might recognize your device as a new one. Then you get 5 tasks per core to proof that your in Robert De Niro's circle of trust.

Hmmm, you know, I'll wait! GFAM will lose 1/4 of the days I can give, but between them making it easier for me to restart failed work, and me messing about AND giving my time (the latter I don't mind, the former I can't do), I'll not worry about waiting :)

Thanks for you help.

[Nov 17, 2011 12:05:13 PM]

Former Member
Cruncher
Joined: May 22, 2018
Post Count: 0
Status: Offline


Re: GFAM: Overclocking caused compute error. O/C now off, how do I fetch new work?

Waiting will do fine, when also participating at other DCs. The client will compensate later, long as you do not suspend the WCG project, meaning that 1/4 day missed will be made up in the near future, when your WCG results compute fine.

As for not seeing any GFAM on the result status page under digitaltoast... happens to those who've got their client attached under a different account (it seems to be a sport at WCG to have multiple accounts and then get a bald spot). Look in the project tab of BOINC manager. Even a space in the name will still make it a different account.

--//--

edit: insert "suspend".

----------------------------------------
[Edit 1 times, last edit by Former Member at Nov 17, 2011 12:18:22 PM]

[Nov 17, 2011 12:17:26 PM]

Former Member
Cruncher
Joined: May 22, 2018
Post Count: 0
Status: Offline


Re: GFAM: Overclocking caused compute error. O/C now off, how do I fetch new work?

OK, I'll wait. Just to clarify something...

meaning that 1/4 day missed will be made up in the near future

In case you thought I meant 1/4 as in quarter of a day, I mean, I can dedicate a 4 core CPU PC to this, 100%, for solid days and nights this year. I've just lost one of them. If there was a "retry WU" option, that would have helped. Possible suggestion for the future :)

[Nov 17, 2011 12:42:50 PM]

[ ]