rpi_logo
Best Paper
Best Paper
log in

Advanced search

Message boards : Number crunching : Best Paper

Author Message
Emanuel
Send message
Joined: 18 Nov 07
Posts: 280
Credit: 2,442,757
RAC: 0

Message 1125 - Posted: 20 Dec 2007, 1:11:16 UTC

Congrats on the award, guys! By the way, how are the convolution code included searches progressing?

On a personal note, I'm leaving for England tomorrow until the end of the year, so no more crunching for me (though I'll finish the WUs I have). Thought I'd mention it, as I seem to be in the top 20 O-O
____________

Profile Travis
Volunteer moderator
Project administrator
Project developer
Project tester
Project scientist
Send message
Joined: 30 Aug 07
Posts: 2046
Credit: 26,480
RAC: 0

Message 1126 - Posted: 20 Dec 2007, 1:17:31 UTC - in response to Message 1125.

Congrats on the award, guys! By the way, how are the convolution code included searches progressing?

On a personal note, I'm leaving for England tomorrow until the end of the year, so no more crunching for me (though I'll finish the WUs I have). Thought I'd mention it, as I seem to be in the top 20 O-O


Yeah i'm assuming things are going to be slowing down because of the holidays :) I'm going to be away from the 26th to the 1st so i hope nothing breaks while i'm gone!

As to the convolution code searches, with the new genetic search/simplex hybrid i've been using, the results are extremely promising. We're getting the search to converge in less evaluations than we were getting using the double-shot approach on the bluegene. This is really good, so what i'm doing right now is re-running the double shot for the convolved and unconvolved data sets (gs_60-64 are unconvolved, gs_70-74 are convolved) to get a side to side comparison of the two on BOINC. I'm also updating our code to run the GS/Simplex method on the BlueGene. I'm hoping to submit these results to the upcoming GECCO conference (http://www.sigevo.org/gecco-2008/) if everything keeps looking as well as it has :)

Profile Travis
Volunteer moderator
Project administrator
Project developer
Project tester
Project scientist
Send message
Joined: 30 Aug 07
Posts: 2046
Credit: 26,480
RAC: 0

Message 1127 - Posted: 20 Dec 2007, 1:18:48 UTC - in response to Message 1126.

Another note about the convolved results. I've sent the results of the 5 searches we did to Nate (with the likelihoods and parameters used to generate them), so hopefully he'll be able to verify that everything is working like it should be. The likelihoods seem right (if not better than what we were getting on the RPI Grid), but i need Nate to make sure that the parameters that generated them are looking good.


Post to thread

Message boards : Number crunching : Best Paper


Main page · Your account · Message boards


Copyright © 2018 AstroInformatics Group