Validation inconclusive with V0.38g installer

Message boards : Number crunching : Validation inconclusive with V0.38g installer
Message board moderation

To post messages, you must log in.

1 · 2 · 3 · 4 · Next

AuthorMessage
Profile perryjay
Volunteer tester
Avatar

Send message
Joined: 20 Aug 02
Posts: 3377
Credit: 20,676,751
RAC: 0
United States
Message 1119343 - Posted: 20 Jun 2011, 13:23:50 UTC

Now that the validators are running again I thought this would be worth another mention.

With the new installer you will] see many more inconclusive results. This is okay because the new app is a lot more accurate than either stock or older opt-apps. Because of this you will give a result that is probably only weakly similar to those other results. They will validate, you just have to wait for an extra wingman to report.

I thought I would put this here in a new thread so we wouldn't clutter up the 0.38 problem threads since this is really not a problem.


PROUD MEMBER OF Team Starfire World BOINC
ID: 1119343 · Report as offensive
Profile Helli_retiered
Volunteer tester
Avatar

Send message
Joined: 15 Dec 99
Posts: 707
Credit: 108,785,585
RAC: 0
Germany
Message 1119354 - Posted: 20 Jun 2011, 13:44:44 UTC - in response to Message 1119343.  

Does that mean, if i have a result done with the stock client and
my wingman has used the lunatcs app, the chance that my stock
client result gets valitated are only 50% ?

Helli
ID: 1119354 · Report as offensive
Profile Sutaru Tsureku
Volunteer tester

Send message
Joined: 6 Apr 07
Posts: 7105
Credit: 147,663,825
RAC: 5
Germany
Message 1119356 - Posted: 20 Jun 2011, 13:50:55 UTC - in response to Message 1119343.  
Last modified: 20 Jun 2011, 13:54:37 UTC

Now that the validators are running again I thought this would be worth another mention.

With the new installer you will] see many more inconclusive results. This is okay because the new app is a lot more accurate than either stock or older opt-apps. Because of this you will give a result that is probably only weakly similar to those other results. They will validate, you just have to wait for an extra wingman to report.

I thought I would put this here in a new thread so we wouldn't clutter up the 0.38 problem threads since this is really not a problem.


Don't mix Windows Installer V0.38 with x38g_cuda32 app.. ;-)

AFAIK, only the new x38g_cuda32 app is now more accurate..

If this new CUDA app is now the wingman of the opt. CPU apps of V0.38 installer & stock CPU apps (and all other stock/opt. CUDA apps out there) it could happen that a 3rd result is needed ((and all 3 results get the Cr.) except, of course, an error happened during calculation).


- Best regards! - Sutaru Tsureku, team seti.international founder. - Optimize your PC for higher RAC. - SETI@home needs your help. -
ID: 1119356 · Report as offensive
Profile perryjay
Volunteer tester
Avatar

Send message
Joined: 20 Aug 02
Posts: 3377
Credit: 20,676,751
RAC: 0
United States
Message 1119357 - Posted: 20 Jun 2011, 13:55:17 UTC - in response to Message 1119354.  
Last modified: 20 Jun 2011, 13:59:21 UTC

No Helli, it means that you will have to wait for another wingman to complete before you get credit.

I should also mention this inconclusive doesn't happen on all tasks just a lot more than we were seeing before. Many do validate on the first try. Again, so long as there is no problem with your rig and you have been turning in good results, you will get your credit. It's just that the new App is giving a bit different result from the other apps. Not enough to cause it to not validate, just enough to cause it to give an inconclusive.

Sorry Sutaru, it's early here and I've been thinking about this all night. I'm sure everyone will get the idea though.


PROUD MEMBER OF Team Starfire World BOINC
ID: 1119357 · Report as offensive
Profile Helli_retiered
Volunteer tester
Avatar

Send message
Joined: 15 Dec 99
Posts: 707
Credit: 108,785,585
RAC: 0
Germany
Message 1119360 - Posted: 20 Jun 2011, 14:08:02 UTC - in response to Message 1119357.  

No Helli, it means that you will have to wait for another wingman to complete before you get credit.
...


Thanks Perry for the quick reply.

If, in this example, the 2nd wingman (the 3rd Result) has also a Lunatics App, do
i get credit, even the two Lunatics Results are equal and gets validated?


I wanted to know if there is a chance, that you dont't get credits for
your hard work, even with the Stock Client, only because two results
from 3rd Party Clients gets validated...

Helli
ID: 1119360 · Report as offensive
JohnDK Crowdfunding Project Donor*Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 28 May 00
Posts: 1222
Credit: 451,243,443
RAC: 1,127
Denmark
Message 1119363 - Posted: 20 Jun 2011, 14:15:54 UTC

Guess it wont help the servers with the extra wingmen needed, o/c depends on how many there actually uses the new app.
ID: 1119363 · Report as offensive
Profile Helli_retiered
Volunteer tester
Avatar

Send message
Joined: 15 Dec 99
Posts: 707
Credit: 108,785,585
RAC: 0
Germany
Message 1119364 - Posted: 20 Jun 2011, 14:25:32 UTC

Yes, i don't say we will see this rare Situation often. But could this
Situation happen?

Helli
ID: 1119364 · Report as offensive
Profile perryjay
Volunteer tester
Avatar

Send message
Joined: 20 Aug 02
Posts: 3377
Credit: 20,676,751
RAC: 0
United States
Message 1119366 - Posted: 20 Jun 2011, 14:42:55 UTC - in response to Message 1119364.  
Last modified: 20 Jun 2011, 14:44:25 UTC

So far, on all mine that have validated with extra wingmen, all three (or four) have received credit except of course for those that would not have validated anyway. Like people running the old V12 app on Fermi cards or just plain bad results. The new app is not that far off from the others, just enough to cause another look at it.

I was teamed with Jason G in a situation like you describe. It took four results before it validated but all four got credit.


PROUD MEMBER OF Team Starfire World BOINC
ID: 1119366 · Report as offensive
Profile James Sotherden
Avatar

Send message
Joined: 16 May 99
Posts: 10436
Credit: 110,373,059
RAC: 54
United States
Message 1119367 - Posted: 20 Jun 2011, 14:44:58 UTC

Probally happen as often as you get stuck with a wingman running a V 12 ap with a fermi card, and then it gets sent to another wingman running a fermi card and v12 to. It validates. And you get squat. Until things get sorted out, It could happen.
[/quote]

Old James
ID: 1119367 · Report as offensive
Profile jason_gee
Volunteer developer
Volunteer tester
Avatar

Send message
Joined: 24 Nov 06
Posts: 7489
Credit: 91,093,184
RAC: 0
Australia
Message 1119369 - Posted: 20 Jun 2011, 14:49:22 UTC - in response to Message 1119364.  
Last modified: 20 Jun 2011, 14:49:31 UTC

Yes, i don't say we will see this rare Situation often. But could this
Situation happen?

Helli


Hi Helli,
While theoretically possible, I have only seen so far legacy clients 'gang-up' on the new boy in a couple of (rare) cases, resulting in no credit:
- stock GPU's that get '-12's' while we process fine finding nice looking triplets. The server gets 'too many errors' and throws out the task with 'Can't validate', granting no-one credit.
- Some spikes or other low level signals possibly in a noisy workunit, sometimes with -9 overflow results where the wingmen were running CPU apps. That possibility actually applies to legacy Cuda apps anyway, is actually a precision issue in the CPU apps, and usually results in all three getting credit anyway, usually (but not always) uses hardly any runtime, and is addressed in coming V7 spike & autocorrelation CPU app code.

HTH,
Jason
"Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions.
ID: 1119369 · Report as offensive
Profile jason_gee
Volunteer developer
Volunteer tester
Avatar

Send message
Joined: 24 Nov 06
Posts: 7489
Credit: 91,093,184
RAC: 0
Australia
Message 1119371 - Posted: 20 Jun 2011, 14:51:51 UTC - in response to Message 1119366.  

...I was teamed with Jason G in a situation like you describe. It took four results before it validated but all four got credit.


LoL, I was amused when you pointed out that my 480 got canonical too...

"Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions.
ID: 1119371 · Report as offensive
Profile perryjay
Volunteer tester
Avatar

Send message
Joined: 20 Aug 02
Posts: 3377
Credit: 20,676,751
RAC: 0
United States
Message 1119382 - Posted: 20 Jun 2011, 15:03:38 UTC - in response to Message 1119371.  
Last modified: 20 Jun 2011, 15:06:04 UTC

Thought you'd like that. I guess we matched and the other two were just far enough off to get stuck with the weakly similar and you were the first in line. Maybe next time I'll get the canonical. :-)

Oh, everything is still getting validated. No errors except one that is still hanging around from way back. It was a -12 that looks like everyone is going to error out on.


PROUD MEMBER OF Team Starfire World BOINC
ID: 1119382 · Report as offensive
Profile Helli_retiered
Volunteer tester
Avatar

Send message
Joined: 15 Dec 99
Posts: 707
Credit: 108,785,585
RAC: 0
Germany
Message 1119393 - Posted: 20 Jun 2011, 15:22:52 UTC

Thanks guys for clarifying this situation and Jason for sharing his deep inside view.

Helli
ID: 1119393 · Report as offensive
Profile Sutaru Tsureku
Volunteer tester

Send message
Joined: 6 Apr 07
Posts: 7105
Credit: 147,663,825
RAC: 5
Germany
Message 1119473 - Posted: 20 Jun 2011, 19:42:26 UTC - in response to Message 1119357.  
Last modified: 20 Jun 2011, 19:46:54 UTC

Sorry Sutaru, it's early here and I've been thinking about this all night. I'm sure everyone will get the idea though.


Yes, the advanced S@h user.

If someone is new and hear installer V0.38g, he will know also that it's meant the installer V0.38 have the x38g_cuda32 app?

This will confuse people, no?

They will ask here in the forum, where I can get the installer V0.38g - although they found already the installer V0.38 DL site..

JMHO.


- Best regards! - Sutaru Tsureku, team seti.international founder. - Optimize your PC for higher RAC. - SETI@home needs your help. -
ID: 1119473 · Report as offensive
Profile perryjay
Volunteer tester
Avatar

Send message
Joined: 20 Aug 02
Posts: 3377
Credit: 20,676,751
RAC: 0
United States
Message 1119508 - Posted: 20 Jun 2011, 22:23:15 UTC

Well, I finally got mine uploaded and reported. Still no errors but still getting a bunch of inconclusives. I figure out of a couple of hundred reported I was about half and half as to valids and to pendings. Those that went to pending had a whole bunch of inconclusives. Hope the next guys are fast.

doncha just love these exact figures? I'm too lazy to try to count them all up. This new app is fast!! :-)


PROUD MEMBER OF Team Starfire World BOINC
ID: 1119508 · Report as offensive
Profile Lint trap

Send message
Joined: 30 May 03
Posts: 871
Credit: 28,092,319
RAC: 0
United States
Message 1119512 - Posted: 20 Jun 2011, 22:49:49 UTC
Last modified: 20 Jun 2011, 22:52:23 UTC

I have 2 validate errors so far, both with the same x32f wingmate.

http://setiathome.berkeley.edu/results.php?userid=108980&offset=0&show_names=0&state=4&appid=

Martin


/edited/ They both have disappeared! That was quick.
ID: 1119512 · Report as offensive
Profile Tazz
Volunteer tester
Avatar

Send message
Joined: 5 Oct 99
Posts: 137
Credit: 34,342,390
RAC: 0
Canada
Message 1119517 - Posted: 20 Jun 2011, 23:17:32 UTC

Quick question, what does " Legacy setiathome_enhanced V6 mode." mean? Or does it mean anything?
</Tazz>
ID: 1119517 · Report as offensive
Profile arkayn
Volunteer tester
Avatar

Send message
Joined: 14 May 99
Posts: 4438
Credit: 55,006,323
RAC: 0
United States
Message 1119528 - Posted: 20 Jun 2011, 23:47:27 UTC - in response to Message 1119517.  

Quick question, what does " Legacy setiathome_enhanced V6 mode." mean? Or does it mean anything?


The new x38 version has the new v7 code that they are testing over at beta in it as well. It will work on both main and beta.

ID: 1119528 · Report as offensive
Profile jason_gee
Volunteer developer
Volunteer tester
Avatar

Send message
Joined: 24 Nov 06
Posts: 7489
Credit: 91,093,184
RAC: 0
Australia
Message 1119530 - Posted: 20 Jun 2011, 23:49:00 UTC - in response to Message 1119517.  

Quick question, what does " Legacy setiathome_enhanced V6 mode." mean? Or does it mean anything?


It means that the application detected that the task is a 'normal' seti@home enhanced mutlibeam version 6 task, rather than the upcoming V7 variety which involves extra searches using autocorrelations for looking for periodic patterns in the data.
"Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions.
ID: 1119530 · Report as offensive
Profile perryjay
Volunteer tester
Avatar

Send message
Joined: 20 Aug 02
Posts: 3377
Credit: 20,676,751
RAC: 0
United States
Message 1119531 - Posted: 20 Jun 2011, 23:49:06 UTC - in response to Message 1119512.  

Martin, there was a post on Lunatics about some validate errors. He just posted again that they had been changed to valid. Might have been the same for you.


PROUD MEMBER OF Team Starfire World BOINC
ID: 1119531 · Report as offensive
1 · 2 · 3 · 4 · Next

Message boards : Number crunching : Validation inconclusive with V0.38g installer


 
©2024 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.