Message boards : Number crunching : Strange!
Message board moderation

To post messages, you must log in.

AuthorMessage
Profile littleBouncer
Avatar

Send message
Joined: 23 Oct 04
Posts: 358
Credit: 1,439,205
RAC: 0
Message 8792 - Posted: 22 Jul 2005, 12:00:02 UTC
Last modified: 22 Jul 2005, 12:11:00 UTC

I have just reported a WU and noticed that this WU was only sent to me. In this case I can wait for a long time that it will be validated.

see WU 330152!

I recept it yesterday (jul. 21.) and reported it today (jul. 22.)

or is something wrong with distributing of WU's?

greetz littleBouncer

[EDIT]This touches the same problem as in thread Where is everyone. Sorry for duplicate!
[/EDIT]

ID: 8792 · Report as offensive     Reply Quote
Richard M
Avatar

Send message
Joined: 18 Sep 04
Posts: 2
Credit: 32,631
RAC: 0
Message 8794 - Posted: 22 Jul 2005, 12:33:42 UTC

Hi littleBouncer,
This is not unusual when you have a fast machine.
It happens all the time with one of mine at Einstein.
The work unit will usually be sent out to other hosts in a day or two.
This however, sometimes causes me to wait more than one week for the work unit to be validated.

:)
Richard

Click the Sig!
Click the Sig!

BOINC Wiki
ID: 8794 · Report as offensive     Reply Quote
Antjest

Send message
Joined: 30 Sep 04
Posts: 21
Credit: 1,442,034
RAC: 0
Message 8795 - Posted: 22 Jul 2005, 12:47:35 UTC
Last modified: 22 Jul 2005, 12:48:14 UTC

I kind of like it this way. Personaly I believe that sending out WU for 5 times is overkill and waste of resources and energy. This way WU's are sent out less times if crunched fast enough.

Whenever I have time to spare I check and abort already validated WU's that are in my cache. My believe is that everything crunched over the validated trio is unnecessary. I do that too in Einstein and Seti .

Tony
ID: 8795 · Report as offensive     Reply Quote
Travis DJ

Send message
Joined: 29 Sep 04
Posts: 196
Credit: 207,040
RAC: 0
Message 8802 - Posted: 22 Jul 2005, 14:50:19 UTC - in response to Message 8795.  

> Personaly I believe that sending out WU for 5
> times is overkill and waste of resources and energy. This way WU's are sent
> out less times if crunched fast enough.

The key words are "if crunched fast enough." They do this because not every host is going to respond within the 2 week window presently prescribed. Presently 3 identical results are required to validate a given WU. If they only send out 3 results per WU and then one additional result after the WU expiration date occurs (assuming 2 of 3 are complete and valid and 1 never replies before expiry), then it could take potentially up to 4 weeks or more for a given WU to be successfully processed & valided. So by initially sending out 5 results they reduce the chance a WU will take several weeks to complete. If all 5 respond and are validated successfully, all 5 hosts recieve credit and the WU has more valid responses to back the work.

Hope that answers your question.

ID: 8802 · Report as offensive     Reply Quote
Profile littleBouncer
Avatar

Send message
Joined: 23 Oct 04
Posts: 358
Credit: 1,439,205
RAC: 0
Message 8807 - Posted: 22 Jul 2005, 15:32:39 UTC
Last modified: 22 Jul 2005, 15:33:09 UTC

THX to all for any replies...

To complicate a little more. At the time when I reported this 'no more sent out WU ', I downloaded a result to crunch which doesn't appear on my resultspages of the database!
Otherhands: this host's resultspage show my a WU (from jul.20.) which I don't have on this host to crunch (in the cache).
That's it, what is suspicious.

greetz littleBouncer
BTW (@ Richard M.): I don't think this host (Notebook 1.5 GHz) is soo fast ...
ID: 8807 · Report as offensive     Reply Quote
Antjest

Send message
Joined: 30 Sep 04
Posts: 21
Credit: 1,442,034
RAC: 0
Message 8813 - Posted: 22 Jul 2005, 18:55:56 UTC - in response to Message 8802.  

I wasn't asking anything and what you did is just speculate.
We can speculate further.
If I would want faster crunching I would also consider shorter deadlines
(like Einstein or even shorter) or combination of number of sent WU with deadlines to achieve most effective throughput.

Tony

> The key words are "if crunched fast enough." They do this because not every
> host is going to respond within the 2 week window presently prescribed.
> Presently 3 identical results are required to validate a given WU. If they
> only send out 3 results per WU and then one additional result after the WU
> expiration date occurs (assuming 2 of 3 are complete and valid and 1 never
> replies before expiry), then it could take potentially up to 4 weeks or more
> for a given WU to be successfully processed & valided. So by initially
> sending out 5 results they reduce the chance a WU will take several weeks to
> complete. If all 5 respond and are validated successfully, all 5 hosts
> recieve credit and the WU has more valid responses to back the work.
>
> Hope that answers your question.
>
ID: 8813 · Report as offensive     Reply Quote
[B@H] Ray

Send message
Joined: 13 Jul 05
Posts: 82
Credit: 6,336
RAC: 0
Message 8827 - Posted: 22 Jul 2005, 23:40:51 UTC

I have the same with this one:

http://lhcathome.cern.ch/workunit.php?wuid=325652

I recieved it 20 July, returned it 21 July, other 4 copies are unsent as of today 22 July.

At least I know that they will go out sometime and the credit applied (at least I hope).

Ray

Pizza@Home - Rays Place - Rays place Forums
ID: 8827 · Report as offensive     Reply Quote
Vedran Brnjetic
Avatar

Send message
Joined: 16 Jul 05
Posts: 24
Credit: 6,549
RAC: 0
Message 8841 - Posted: 23 Jul 2005, 8:58:50 UTC - in response to Message 8827.  

> I have the same with this one:
>
> http://lhcathome.cern.ch/workunit.php?wuid=325652
>
> I recieved it 20 July, returned it 21 July, other 4 copies are unsent as of
> today 22 July.
>
> At least I know that they will go out sometime and the credit applied (at
> least I hope).
>
> Ray
>

I posted the same question on the other forum thread (Where is everyone). Meanwhile one of the workunits was sent to one other host which also finished it before it was sent to the others.
<img border="0" src="http://boinc.mundayweb.com/one/stats.php?userID=2328&amp;trans=off" />
<img border="0" src="http://boinc.mundayweb.com/one/stats.php?userID=2328&amp;prj=5&amp;trans=off" />
ID: 8841 · Report as offensive     Reply Quote
Profile Keck_Komputers

Send message
Joined: 1 Sep 04
Posts: 275
Credit: 2,652,452
RAC: 0
Message 8843 - Posted: 23 Jul 2005, 9:19:47 UTC - in response to Message 8807.  

> THX to all for any replies...
>
> To complicate a little more. At the time when I reported this <a> href="http://lhcathome.cern.ch/workunit.php?wuid=330152">'no more sent out WU
> '[/url], I downloaded a result to crunch which doesn't appear on my resultspages
> of the database!
> Otherhands: this host's resultspage show my a WU (from jul.20.) which I don't
> have on this host to crunch (in the cache).
> That's it, what is suspicious.
>
> greetz littleBouncer
> BTW (@ Richard M.): I don't think this host (Notebook 1.5 GHz) is soo fast ...
>
Search for ghost workunits in the fora. This is a known problem with older versions of the server and client software. It is believed to be fixed in the lastest source (about 4.45 actually if I remember correctly). At any rate it is now a rare problem on the projects that it was a big problem on before.
BOINC WIKI

BOINCing since 2002/12/8
ID: 8843 · Report as offensive     Reply Quote
Jayargh

Send message
Joined: 24 Oct 04
Posts: 79
Credit: 257,762
RAC: 0
Message 8852 - Posted: 23 Jul 2005, 13:35:07 UTC - in response to Message 8843.  
Last modified: 23 Jul 2005, 13:40:23 UTC

At any rate it
> is now a rare problem on the projects that it was a big problem on before.
>
Yes it was paticularly bad over at Einstein with ghost units..... and this sending out 1 unit then more in a day or 2 has been a common experience since LHC got the new server,I have been observing this but never made mention , didn't think it was important enough... guess with more people crunching more things get noticed and posted.....@ Colt... LHC had a 3 validate system and quorum at 2 but this got out of control when 0 was reported by some users due to a problem with sixtrack program and many paticipants got 0 credit when crunching a 100 credit workunit,remember?.... sometimes a good reason to over validate... Beta was a lot of fun...many of us lost a LOT of Boinc credit but thats not my primary concern or the others who have crunched from the beginning. The science is the X factor :)
ID: 8852 · Report as offensive     Reply Quote
Antjest

Send message
Joined: 30 Sep 04
Posts: 21
Credit: 1,442,034
RAC: 0
Message 8885 - Posted: 24 Jul 2005, 8:46:43 UTC - in response to Message 8852.  

I remember that times. I got many 0 credits. And many resets to make things work again. But now in production mode we can do more science work if 5th result would not be sent or sent later or just in case of errors.

All those 0 credit errors are gone now, at least I haven't seen any lately and since LHC crunching results must be exactly the same, they can return to the first WU sending policy, send 3 quorum 2 and even more WU's can be crunched.
They have many in the pipeline as they said.

At the end, I like to do more science, and credit comes second. At least here in LHC.
Or maybe I am sentimental cause this is European project or that I have a cousin in Switzerland :)

Tony

> @Colt... LHC had a 3 validate system and
> quorum at 2 but this got out of control when 0 was reported by some users due
> to a problem with sixtrack program and many paticipants got 0 credit when
> crunching a 100 credit workunit,remember?.... sometimes a good reason to over
> validate... Beta was a lot of fun...many of us lost a LOT of Boinc credit but
> thats not my primary concern or the others who have crunched from the
> beginning. The science is the X factor :)
>
ID: 8885 · Report as offensive     Reply Quote
Profile Juerschi

Send message
Joined: 26 Oct 04
Posts: 12
Credit: 8,909
RAC: 0
Message 9028 - Posted: 29 Jul 2005, 18:02:49 UTC

Another strange thing here: http://lhcathome.cern.ch/result.php?resultid=1864418

It looks like result of a BURP WU
ID: 9028 · Report as offensive     Reply Quote
tekwyzrd
Avatar

Send message
Joined: 15 Jul 05
Posts: 6
Credit: 2,710
RAC: 0
Message 9069 - Posted: 1 Aug 2005, 3:43:41 UTC - in response to Message 8807.  


> Otherhands: this host's resultspage show my a WU (from jul.20.) which I don't
> have on this host to crunch (in the cache).


I noticed a unit on my result page that's I never received.

http://lhcathome.cern.ch/workunit.php?wuid=353216

Nothing travels faster than the speed of light with the possible exception of bad news, which obeys its own special laws.
Douglas Adams (1952 - 2001)
ID: 9069 · Report as offensive     Reply Quote
Profile Alex

Send message
Joined: 2 Sep 04
Posts: 378
Credit: 10,765
RAC: 0
Message 9070 - Posted: 1 Aug 2005, 4:14:40 UTC - in response to Message 9028.  
Last modified: 1 Aug 2005, 4:14:55 UTC

> Another strange thing here:
> http://lhcathome.cern.ch/result.php?resultid=1864418
>
> It looks like result of a BURP WU
>

Looks like their machine is sending LHC mostly BURP wu's.

http://lhcathome.cern.ch/results.php?hostid=34918


I'm not the LHC Alex. Just a number cruncher like everyone else here.
ID: 9070 · Report as offensive     Reply Quote
tekwyzrd
Avatar

Send message
Joined: 15 Jul 05
Posts: 6
Credit: 2,710
RAC: 0
Message 9077 - Posted: 1 Aug 2005, 8:37:07 UTC - in response to Message 9070.  
Last modified: 1 Aug 2005, 9:02:02 UTC

> > Another strange thing here:
> > http://lhcathome.cern.ch/result.php?resultid=1864418
> >
> > It looks like result of a BURP WU
> >
>
> Looks like their machine is sending LHC mostly BURP wu's.
>
> http://lhcathome.cern.ch/results.php?hostid=34918
>
>
>

Re: Host ID above
Comparing the wu times of the machine in question and the wu times of other machines running the same units it looks like there's a major problem. Someone might want to look into the result details. There's an awful lot of invalid units there.

EDIT: I just saw reference to this machine in another message.
Nothing travels faster than the speed of light with the possible exception of bad news, which obeys its own special laws.
Douglas Adams (1952 - 2001)
ID: 9077 · Report as offensive     Reply Quote
tekwyzrd
Avatar

Send message
Joined: 15 Jul 05
Posts: 6
Credit: 2,710
RAC: 0
Message 9080 - Posted: 1 Aug 2005, 20:02:56 UTC
Last modified: 1 Aug 2005, 20:03:55 UTC

I'm running BOINC version 4.72. I just noticed that the Number of times client has contacted server entry for my computer is 0. I've returned 25 completed work units. This number is updated on the seti site but not here. The free disk space info isn't updated on either site.
Nothing travels faster than the speed of light with the possible exception of bad news, which obeys its own special laws.
Douglas Adams (1952 - 2001)
ID: 9080 · Report as offensive     Reply Quote

Message boards : Number crunching : Strange!


©2024 CERN