Advanced search

Message boards : Number crunching : Welcome back SDOERR !

Author Message
tomba
Send message
Joined: 21 Feb 09
Posts: 497
Credit: 700,690,702
RAC: 0
Level
Lys
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 41683 - Posted: 20 Aug 2015 | 18:13:33 UTC

Got my first SDOERR WU since April. I hope more are in the pipeline. My GTX 750Ti is overloaded with GERARDs..

lycnus
Send message
Joined: 18 Jan 09
Posts: 8
Credit: 196,775,113
RAC: 0
Level
Ile
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwat
Message 41958 - Posted: 6 Oct 2015 | 14:38:52 UTC

Got one today. It would take 350 hours to finish on my 780Ti I abort it and hope the next would be below 20 hours :)

Profile Beyond
Avatar
Send message
Joined: 23 Nov 08
Posts: 1112
Credit: 6,162,416,256
RAC: 0
Level
Tyr
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 44457 - Posted: 7 Sep 2016 | 14:09:27 UTC

And they're back AGAIN. Just got my 1st SDOERR in probably a year. Hope it's not 350 hours like the one Perox received above.

Bedrich Hajek
Send message
Joined: 28 Mar 09
Posts: 485
Credit: 11,233,181,238
RAC: 12,828,835
Level
Trp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 44478 - Posted: 11 Sep 2016 | 14:26:43 UTC

I received a few of these tasks, myself, and I noticed that they are initially sent out to 2 hosts. I think it would be a good idea to do this for all the tasks, because you would get results back faster, and it's another way to verify the accuracy of the results, which gives you piece of mind.

If you do end up with different results from the same task from different hosts, that may tell you the simulation is faulty, or one of the hosts has given you faulty results and you get a big headache.

Overall, I believe this would be positive.






Tomas Brada
Send message
Joined: 3 Nov 15
Posts: 38
Credit: 6,768,093
RAC: 0
Level
Ser
Scientific publications
wat
Message 44498 - Posted: 12 Sep 2016 | 20:35:58 UTC - in response to Message 44478.

I received a few of these tasks, myself, and I noticed that they are initially sent out to 2 hosts. I think it would be a good idea to do this for all the tasks, because you would get results back faster, and it's another way to verify the accuracy of the results, which gives you piece of mind.

That would be bad. Because half of the work done by volunteers will be wasted.

____________

Profile skgiven
Volunteer moderator
Volunteer tester
Avatar
Send message
Joined: 23 Apr 09
Posts: 3968
Credit: 1,995,359,260
RAC: 0
Level
His
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 44503 - Posted: 12 Sep 2016 | 22:13:31 UTC - in response to Message 44498.

Based on what I know about this project and their publication record, which requires vetting and validation, I'm happy for the researchers to work out what's best for their research, including validation methodology.
____________
FAQ's

HOW TO:
- Opt out of Beta Tests
- Ask for Help

Profile caffeineyellow5
Avatar
Send message
Joined: 30 Jul 14
Posts: 225
Credit: 2,658,976,345
RAC: 0
Level
Phe
Scientific publications
watwatwatwatwatwatwatwat
Message 44509 - Posted: 13 Sep 2016 | 6:36:03 UTC - in response to Message 44478.

I received a few of these tasks, myself, and I noticed that they are initially sent out to 2 hosts. I think it would be a good idea to do this for all the tasks, because you would get results back faster, and it's another way to verify the accuracy of the results, which gives you piece of mind.
I brought this up before based on other projects that did this method and was convinced that the method for GPUGRID is sufficient for the work and not wasting any of it. Back in my UD days they would send out 3 per WU and the environment they used was subject to 'jitter' in the CPU. But this project apparently does not have any jitter between units and the validation/verification process is good for the work being done.
____________
1 Corinthians 9:16 "For though I preach the gospel, I have nothing to glory of: for necessity is laid upon me; yea, woe is unto me, if I preach not the gospel!"
Ephesians 6:18-20, please ;-)
http://tbc-pa.org

Stefan
Project administrator
Project developer
Project tester
Project scientist
Send message
Joined: 5 Mar 13
Posts: 348
Credit: 0
RAC: 0
Level

Scientific publications
wat
Message 44519 - Posted: 14 Sep 2016 | 10:23:57 UTC

Thanks! Happy to run some sims again :)
I do indeed duplicate WUs for speed and consistency but the number of WUs is so low (max 40 at a single time) I don't feel like I'm really wasting resources.

350hours sounds like a bug in time calculation. These are pretty tested and proven simulations that we have run in the past. This time I'm just testing a different algorithm for sampling but the setup should be identical.

I will also be sending out some OPM simulations again soon which were quite the subject of discussion last time, haha. But they are a one-off deal of around 700 simulations.

Profile Beyond
Avatar
Send message
Joined: 23 Nov 08
Posts: 1112
Credit: 6,162,416,256
RAC: 0
Level
Tyr
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 44533 - Posted: 14 Sep 2016 | 23:21:36 UTC - in response to Message 44519.

350hours sounds like a bug in time calculation. These are pretty tested and proven simulations that we have run in the past. This time I'm just testing a different algorithm for sampling but the setup should be identical.

Actually they're running refreshingly fast (in about 2/3 the time of a Gerard WU and many times faster than a long Gianni). They bring back fond memories of the typical long WUs of (less than) a year ago. Thumbs up!

Post to thread

Message boards : Number crunching : Welcome back SDOERR !

//