High Scores Anyone?

Message boards : Number crunching : High Scores Anyone?

To post messages, you must log in.

AuthorMessage
Profile dcdc

Send message
Joined: 3 Nov 05
Posts: 1829
Credit: 115,423,876
RAC: 52,998
Message 30240 - Posted: 29 Oct 2006, 19:27:29 UTC
Last modified: 29 Oct 2006, 19:29:15 UTC

Some of the credit given for these results is slightly high! It's a Celeron 2.8 with a RAC of over 5k:
https://boinc.bakerlab.org/rosetta/show_host_detail.php?hostid=260872
It looks like the decoys are being considered complete and so are getting full credit even though some are only running for a few seconds.

This one too:
https://boinc.bakerlab.org/rosetta/show_host_detail.php?hostid=319789
ID: 30240 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile scsimodo

Send message
Joined: 17 Sep 05
Posts: 93
Credit: 946,359
RAC: 0
Message 30248 - Posted: 29 Oct 2006, 20:57:14 UTC - in response to Message 30240.  

Unfortunately no high credits for me :)
But my last 6 or 8 results are all credited granted = claimed. Never seen this in the last 6 or 8 weeks. Something is really strange here...

ID: 30248 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile River~~
Avatar

Send message
Joined: 15 Dec 05
Posts: 761
Credit: 285,578
RAC: 0
Message 30250 - Posted: 29 Oct 2006, 21:38:52 UTC - in response to Message 30248.  

Unfortunately no high credits for me :)
But my last 6 or 8 results are all credited granted = claimed. Never seen this in the last 6 or 8 weeks. Something is really strange here...


If you look at your results on the website after they have been returned, are they failing in some way?

If so the fact that grant = claim could simply reflect the fact that the project team are giving you credit for time spent on broken software, which has been Rosetta policy since late 2005.

R~~
ID: 30250 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
BennyRop

Send message
Joined: 17 Dec 05
Posts: 555
Credit: 140,800
RAC: 0
Message 30252 - Posted: 29 Oct 2006, 22:00:48 UTC

ID: 30252 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile scsimodo

Send message
Joined: 17 Sep 05
Posts: 93
Credit: 946,359
RAC: 0
Message 30253 - Posted: 29 Oct 2006, 22:01:39 UTC - in response to Message 30250.  
Last modified: 29 Oct 2006, 22:04:26 UTC

Unfortunately no high credits for me :)
But my last 6 or 8 results are all credited granted = claimed. Never seen this in the last 6 or 8 weeks. Something is really strange here...


If you look at your results on the website after they have been returned, are they failing in some way?


No, all results seem to be OK. Claimed = Granted only happened with FRA_t368-WUs, DOC-WUs are granted different to claimed.
ID: 30253 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Jose

Send message
Joined: 28 Mar 06
Posts: 820
Credit: 48,297
RAC: 0
Message 30288 - Posted: 30 Oct 2006, 14:28:47 UTC - in response to Message 30240.  

Some of the credit given for these results is slightly high! It's a Celeron 2.8 with a RAC of over 5k:
https://boinc.bakerlab.org/rosetta/show_host_detail.php?hostid=260872
It looks like the decoys are being considered complete and so are getting full credit even though some are only running for a few seconds.

This one too:
https://boinc.bakerlab.org/rosetta/show_host_detail.php?hostid=319789


Interesting numbers for his Celerons. In many cases some of the computers identified (recognized by the Rosetta Server) as Celerons are getting more credits than the Pentiums.

¿Could it be a case of a bad id, that the comps are not Celeries?

Or a case of bad validation? A quirk at the validator level?


This and no other is the root from which a Tyrant springs; when he first appears he is a protector.”
Plato
ID: 30288 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Mats Petersson

Send message
Joined: 29 Sep 05
Posts: 225
Credit: 951,788
RAC: 0
Message 30289 - Posted: 30 Oct 2006, 14:53:41 UTC

I had a look, and it seems like several decoys are attempted with very short runtime before the watchdog kills the process - I'm not sure what happens, but one had about 10 attempts to get going before it was ultimately killed, all within 10 minutes (600 seconds). That could account for a quick accumulation of points...

--
Mats
ID: 30289 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
AMD_is_logical

Send message
Joined: 20 Dec 05
Posts: 299
Credit: 31,460,681
RAC: 0
Message 30301 - Posted: 30 Oct 2006, 19:02:25 UTC

Perhaps this person is sneakernetting several machines and is using this machine to upload?
ID: 30301 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile dcdc

Send message
Joined: 3 Nov 05
Posts: 1829
Credit: 115,423,876
RAC: 52,998
Message 30302 - Posted: 30 Oct 2006, 19:07:36 UTC

the problem is the decoys are being counted and credited as valid even though they only run for a few seconds.
ID: 30302 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
AMD_is_logical

Send message
Joined: 20 Dec 05
Posts: 299
Credit: 31,460,681
RAC: 0
Message 30317 - Posted: 30 Oct 2006, 21:05:25 UTC - in response to Message 30302.  

the problem is the decoys are being counted and credited as valid even though they only run for a few seconds.


It's true those decoys couldn't have been generated in just a few seconds. However, If he's using some strange method of uploading sneakernetted results from other machines, then the reported run times might not reflect the actual run times.
ID: 30317 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Chu

Send message
Joined: 23 Feb 06
Posts: 120
Credit: 112,439
RAC: 0
Message 30377 - Posted: 31 Oct 2006, 17:07:23 UTC

We are looking into this issue right now. For the stderr output, it looks like a file clean-up issue as it constains many copies of "normal exit". Hopefully we will be able to reproduce it locally to confirm that. More to come...
ID: 30377 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote

Message boards : Number crunching : High Scores Anyone?



©2024 University of Washington
https://www.bakerlab.org