WARNING: This website is obsolete! Please follow this link to get to the new Albert@Home website!

Work availability

Message boards : News : Work availability
Message board moderation

To post messages, you must log in.

AuthorMessage
Profile Oliver Behnke
Volunteer moderator
Project administrator
Project developer

Send message
Joined: 4 Sep 07
Posts: 130
Credit: 8,545,955
RAC: 0
Message 112261 - Posted: 5 Nov 2012, 14:24:25 UTC

Hi,

You may have noticed that we didn't produce new work units over the past few days. This was due to an infrastructure upgrade in our data center. While we made sure that our production project wasn't affected, our test project ran out of work. Oh well, it's a test project after all.

Anyway, we began producing work again, at least for the binary radio pulsar search, so you should be able to fetch new work units soon.

FYI, we're also going to update the BRP apps themselves over the next days. They include improvements that should make them more robust against external termination, in particular the GPU apps.


Thanks for your patience!

Oliver
ID: 112261 · Report as offensive     Reply Quote
Profile Oliver Behnke
Volunteer moderator
Project administrator
Project developer

Send message
Joined: 4 Sep 07
Posts: 130
Credit: 8,545,955
RAC: 0
Message 112262 - Posted: 6 Nov 2012, 11:17:53 UTC - in response to Message 112261.  

Update: new CUDA and OpenCL applications v1.31 released (see above for details).

Cheers,
Oliver
ID: 112262 · Report as offensive     Reply Quote
Profile Stephan Volkmann

Send message
Joined: 5 Jun 12
Posts: 4
Credit: 169,531
RAC: 0
Message 112263 - Posted: 6 Nov 2012, 15:48:01 UTC - in response to Message 112262.  

Update: new CUDA and OpenCL applications v1.31 released (see above for details).

Cheers,
Oliver


thx, three of them arrived on my PC suggesting 20 h,
but in real they will be done >1h i think

cu, hf, stephan


ID: 112263 · Report as offensive     Reply Quote
Profile Stephan Volkmann

Send message
Joined: 5 Jun 12
Posts: 4
Credit: 169,531
RAC: 0
Message 112264 - Posted: 6 Nov 2012, 16:33:07 UTC - in response to Message 112263.  

[16:47:35][10121][INFO ] Checkpoint committed!
[16:47:50][10121][ERROR] Error during CUDA device->host time series length transfer (error: 700)
[16:47:50][10121][ERROR] Demodulation failed (error: 1008)!
16:47:50 (10121): called boinc_finish

grmml, whats happend there?

ID: 112264 · Report as offensive     Reply Quote
Profile Oliver Behnke
Volunteer moderator
Project administrator
Project developer

Send message
Joined: 4 Sep 07
Posts: 130
Credit: 8,545,955
RAC: 0
Message 112268 - Posted: 9 Nov 2012, 13:06:00 UTC - in response to Message 112264.  

This means an error occurred during GPU operations. You seem to be the only one with this particular error so it's not a general issue. Could very well be a driver issue...

Oliver
ID: 112268 · Report as offensive     Reply Quote
Profile Oliver Behnke
Volunteer moderator
Project administrator
Project developer

Send message
Joined: 4 Sep 07
Posts: 130
Credit: 8,545,955
RAC: 0
Message 112269 - Posted: 9 Nov 2012, 13:08:01 UTC - in response to Message 112263.  
Last modified: 9 Nov 2012, 13:08:14 UTC


thx, three of them arrived on my PC suggesting 20 h,
but in real they will be done >1h i think


Yes, that's an effect of BOINC's new runtime estimation scheme we test here at Albert@Home. BOINC needs to gather statistics each time we release a new app version until the numbers converge on the actual/real/correct value...

Sorry,
Oliver
ID: 112269 · Report as offensive     Reply Quote
Profile Stephan Volkmann

Send message
Joined: 5 Jun 12
Posts: 4
Credit: 169,531
RAC: 0
Message 112270 - Posted: 9 Nov 2012, 15:02:35 UTC - in response to Message 112268.  

ok, thinking the wu is died by a lost network connection, grmml, shit happend..
ID: 112270 · Report as offensive     Reply Quote
gm102gameailatrieuphu

Send message
Joined: 8 Mar 13
Posts: 1
Credit: 0
RAC: 0
Message 112365 - Posted: 8 Mar 2013, 6:55:49 UTC - in response to Message 112270.  

it's still good man!
ID: 112365 · Report as offensive     Reply Quote

Message boards : News : Work availability



This material is based upon work supported by the National Science Foundation (NSF) under Grant PHY-0555655 and by the Max Planck Gesellschaft (MPG). Any opinions, findings, and conclusions or recommendations expressed in this material are those of the investigators and do not necessarily reflect the views of the NSF or the MPG.

Copyright © 2019 Bruce Allen for the LIGO Scientific Collaboration