0 Seconds and Verified?

B.I.G
B.I.G
Joined: 26 Oct 07
Posts: 116
Credit: 1,126,744,861
RAC: 1,130,099
Topic 193681

Something is odd with this WU:
http://einsteinathome.org/workunit/39431299

I would like such a fast CPU too ;)
Seems like the client of my wingman forgot to count the seconds?

Michael Karlinsky
Michael Karlinsky
Joined: 22 Jan 05
Posts: 888
Credit: 23,502,182
RAC: 0

0 Seconds and Verified?

Quote:

Something is odd with this WU:
http://einsteinathome.org/workunit/39431299

I would like such a fast CPU too ;)
Seems like the client of my wingman forgot to count the seconds?

Very old core client (Is it still called that?) bug.

4.43

Michael

Jord
Joined: 26 Jan 05
Posts: 2,952
Credit: 5,878,109
RAC: 7,006

More probably a problem with

More probably a problem with the 4.43 version of BOINC that he is still using.

Gary Roberts
Gary Roberts
Moderator
Joined: 9 Feb 05
Posts: 5,870
Credit: 115,823,015,255
RAC: 35,384,016

RE: More probably a problem

Message 81868 in response to message 81867

Quote:
More probably a problem with the 4.43 version of BOINC that he is still using.

You and Michael are close but, I'm sorry, no cigar :-).

The problem with the old BOINC clients is not that they misreport time but rather that they claim very low credit. This is not a problem because the correct server side credit is awarded automatically.

If you actually look through the complete results list for the wingman's computer, you will find that most of the results listed there show zero time but that on the second page there are the two oldest which do show the correct time. All tasks were done with BOINC 4.43 but there was an app change that seems related to the zero time issue.

The two oldest visible results which show the correct time (but have the low credit claim) were done with app version 4.03 whilst all the zero time results (which also show a zero credit claim) were done with app version 4.43. Perhaps the issue is the combination of the old core client with the latest app.

If you actually look through the stderr.txt output of any of the zero crunch time results, you can see the correct starting and finishing time stamps listed there so you can see an estimation of crunch time by subtracting one from the other. This would be the "wall clock" time that elapsed during crunching.

I'll bring this thread to Bernd's attention in case he might be interested. Because it's such an old core client, he may not be :-).

Cheers,
Gary.

Bernd Machenschalk
Bernd Machenschalk
Moderator
Administrator
Joined: 15 Oct 04
Posts: 4,305
Credit: 248,799,596
RAC: 33,436

Interesting question is: why

Interesting question is: why did it stop counting (or reporting) CPU seconds on May 14? The tasks before look ok, apparently with the same client version.

BM

BM

Jord
Joined: 26 Jan 05
Posts: 2,952
Credit: 5,878,109
RAC: 7,006

Interaction between the new

Message 81870 in response to message 81869

Interaction between the new application and the old client? Something in the application that depends on a 5.x.x client?

Here's the 4.43 client for Windows if anyone wants to test it. In case it's something specific with a Mac, 4.43 for MacOSX, or separate for PPC.

Gary Roberts
Gary Roberts
Moderator
Joined: 9 Feb 05
Posts: 5,870
Credit: 115,823,015,255
RAC: 35,384,016

RE: Interesting question

Message 81871 in response to message 81869

Quote:
Interesting question is: why did it stop counting (or reporting) CPU seconds on May 14?

Are you talking about the result returned on May 13 that had just 45.86 seconds recorded?

I think you will find the source of those 45.86 seconds if you look at the time stamps right at the end of stderr.txt.

Cheers,
Gary.

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.