Message boards : Number crunching : Results page
Message board moderation

To post messages, you must log in.

1 · 2 · Next

AuthorMessage
Profile B-Roy

Send message
Joined: 1 Sep 04
Posts: 55
Credit: 20,907
RAC: 0
Message 5746 - Posted: 21 Feb 2005, 21:48:50 UTC

I know that the server is under heavy load at the moment, but I would invite you to enable the results page as soon as possible. This page is crucial to see the progress of the wus and gives information that can be used for bug-tracking.

and thanks for the nicest screen saver once again.
ID: 5746 · Report as offensive     Reply Quote
PCHome Denmark

Send message
Joined: 17 Sep 04
Posts: 14
Credit: 12,301
RAC: 0
Message 5749 - Posted: 21 Feb 2005, 22:11:52 UTC

Yes. It would be nice if they could enable the Result page.
ID: 5749 · Report as offensive     Reply Quote
Profile sysfried

Send message
Joined: 27 Sep 04
Posts: 282
Credit: 1,415,417
RAC: 0
Message 5760 - Posted: 21 Feb 2005, 22:52:42 UTC - in response to Message 5749.  

> Yes. It would be nice if they could enable the Result page.
>
voting for that too!!!! ;-)
ID: 5760 · Report as offensive     Reply Quote
Profile The Gas Giant

Send message
Joined: 2 Sep 04
Posts: 309
Credit: 715,258
RAC: 0
Message 5772 - Posted: 22 Feb 2005, 1:50:58 UTC

Me too....I'd love to know if my machine at home has picked up wu's since I attached to the public project on the way out this morning. I didn't have time to sit and watch it download.

Paul
(S@H1 8888)
BOINC/SAH BETA
ID: 5772 · Report as offensive     Reply Quote
Profile littleBouncer
Avatar

Send message
Joined: 23 Oct 04
Posts: 358
Credit: 1,439,205
RAC: 0
Message 5805 - Posted: 22 Feb 2005, 10:00:30 UTC - in response to Message 5772.  
Last modified: 22 Feb 2005, 10:03:22 UTC

> Me too....I'd love to know if my machine at home has picked up wu's since I
> attached to the public project on the way out this morning. I didn't have
> time to sit and watch it download.
>
>
Maybe they (LHC-team) want not show the next problem (what you have posted in the old Forum about claiming credits with linux-crunchers and windows).

-no offense, with this reply!!-

I vote to activate the result-page too:-)
greetz from Switzerland
littleBouncer


ID: 5805 · Report as offensive     Reply Quote
Profile Logan5@SETI.USA
Avatar

Send message
Joined: 30 Sep 04
Posts: 112
Credit: 104,059
RAC: 0
Message 5808 - Posted: 22 Feb 2005, 11:20:33 UTC

They had to reduce the amount of RPC calls until the initial flood of returning users subsides... LOL Stats are likely low on their priority list ATM...

I was downloading the new 6 track overlay at a speedy 1.3 kbps....My 56k Dialup Modem does 4.1...sheesh!!

ID: 5808 · Report as offensive     Reply Quote
PCHome Denmark

Send message
Joined: 17 Sep 04
Posts: 14
Credit: 12,301
RAC: 0
Message 5810 - Posted: 22 Feb 2005, 11:38:44 UTC
Last modified: 22 Feb 2005, 11:41:51 UTC

Are we still in a test phase of some kind? Or is there a good explanition, why the result page is disable. People don't have any clue about there returned results, if they are good or bad. Just wondering.
ID: 5810 · Report as offensive     Reply Quote
Profile FalconFly
Avatar

Send message
Joined: 2 Sep 04
Posts: 121
Credit: 592,214
RAC: 0
Message 5812 - Posted: 22 Feb 2005, 12:03:39 UTC - in response to Message 5810.  
Last modified: 22 Feb 2005, 12:06:35 UTC

Yep, sure would be nice to have it back ASAP :)

Looks like they're trying to manage their Server Loads until they stabilize and/or optimize their Systems to accept a higher load.

(I can't imagine the relatively low amount of Users imposing a severe Problem, SETI handles 15x the Userbase with "just" a Dual Opteron Server by now; although arguably, Berkeley split various Project Tasks onto smaller ones as well for better workload distribution)
Scientific Network : 45000 MHz - 77824 MB - 1970 GB
ID: 5812 · Report as offensive     Reply Quote
Borged by MGP

Send message
Joined: 22 Oct 04
Posts: 2
Credit: 1,381,520
RAC: 0
Message 5875 - Posted: 22 Feb 2005, 18:04:00 UTC
Last modified: 22 Feb 2005, 18:04:28 UTC

It would be useful to see at least the results pages active, althouhg I can understand the server balancing issues.

One of My crunchers has already reached it's daily quota of WUs. Yet it is about to run out of LHC work. It's only a P4 1800 and also does Einstein and Predictor all equally resourced. I can't beleive the PC has correclty crunched all the LHC stuff it had, yet I can't see anything abnormal in the boinc messages.

The Results page would help me verify all is in order, or, if there was an issue, perhaps help troubleshooting, at this early stage, before loads of WUs are invalidated or wasted.
ID: 5875 · Report as offensive     Reply Quote
Profile Markku Degerholm

Send message
Joined: 3 Sep 04
Posts: 212
Credit: 4,545
RAC: 0
Message 5908 - Posted: 22 Feb 2005, 20:35:16 UTC
Last modified: 22 Feb 2005, 20:36:18 UTC

We will enable the results page soon.

But even before we do, I can tell you that we do have a new problem with credit granting. For some reason the client may report the used CPU time as zero, even if the computed result is valid. We are not sure if the problem is with the core client, the sixtrack, or both. It happens on both windows and linux. And it's random, meaning than some results reported by a host have zero CPU time while others don't. You will find this problem from Known Problems page too.

Most of the time this is not a big problem because the granted credit is an average of claimed credits. If two results have nonzero CPU time and one does not, the result with zero CPU time will receive credit too. But if too many results have zero CPU time, none will get credit.

This is heard to happen with Einstein@home too, but I have heard of no good solution.

Anyway, even the results with zero CPU time are 100% usable for the project and you are not wasting your CPU time. And we are trying to fix the thing.

Markku Degerholm
LHC@home admin
ID: 5908 · Report as offensive     Reply Quote
Profile FalconFly
Avatar

Send message
Joined: 2 Sep 04
Posts: 121
Credit: 592,214
RAC: 0
Message 5930 - Posted: 22 Feb 2005, 22:10:00 UTC - in response to Message 5908.  
Last modified: 22 Feb 2005, 22:10:52 UTC

Thanks for the feedback :)

In fact, right now I'm seeing one of my systems (Host 21915) having a finished WorkUnit "Ready to report" with 0 CPU Time indicated both in the BOINC GUI, as well as monitored via GUI_RPC (BOINCview).

I tried "Suspend - Resume", as well as physically exiting the BOINC V4.19 Client and restarting it; all to no effect.
Unfortunately I didn't monitor it all the time, but I'm confident the WorkUnit did count CPU time during processing (nothing abnormal observed).

IMHO (best educated guess, based on my limited observations in this isolated case) the error most probably occured at the point where the Client took/finalized and wrote the total CPU time into the Result File on the System.
Scientific Network : 45000 MHz - 77824 MB - 1970 GB
ID: 5930 · Report as offensive     Reply Quote
Profile littleBouncer
Avatar

Send message
Joined: 23 Oct 04
Posts: 358
Credit: 1,439,205
RAC: 0
Message 5987 - Posted: 23 Feb 2005, 17:24:25 UTC - in response to Message 5908.  
Last modified: 23 Feb 2005, 17:25:20 UTC

> This is heard to happen with Einstein@home too, but I have heard of no good
> solution.
>
@ Markku

Even on PAH: Some very few linux-WU-returns had negativ CPU-time!
(They even don't know, that they have a time-machine,LOL-that's a joke)

negative CPU-time gives negative claimed credit -> negativ granted credits.
the scenario discribed was given -0.12 credits; I posted that on PAH's - forum and Chahm (from the PAH-team) tries to solve it, as he has replied.

It seems : it is the same "bug".

This post is only for information

greetz from Switzerland
littleBouncer


ID: 5987 · Report as offensive     Reply Quote
John McLeod VII
Avatar

Send message
Joined: 2 Sep 04
Posts: 165
Credit: 146,925
RAC: 0
Message 5990 - Posted: 23 Feb 2005, 17:46:16 UTC - in response to Message 5908.  

> We will enable the results page soon.
>
> But even before we do, I can tell you that we do have a new problem with
> credit granting. For some reason the client may report the used CPU time as
> zero, even if the computed result is valid. We are not sure if the problem is
> with the core client, the sixtrack, or both. It happens on both windows and
> linux. And it's random, meaning than some results reported by a host have zero
> CPU time while others don't. You will find this problem from Known Problems
> page too.
>
> Most of the time this is not a big problem because the granted credit is an
> average of claimed credits. If two results have nonzero CPU time and one does
> not, the result with zero CPU time will receive credit too. But if too many
> results have zero CPU time, none will get credit.
>
> This is heard to happen with Einstein@home too, but I have heard of no good
> solution.
>

I watched the CPU time switch to 0 once (I was watching when it happened in the GUI). That time, it happened at the very end of processing when it switched from Running to Uploading to Ready to Report. (I have a high speed connection (T1 on that macine), and the upload took significantly less than a second).
> Anyway, even the results with zero CPU time are 100% usable for the project
> and you are not wasting your CPU time. And we are trying to fix the
> thing.
>
>


BOINC WIKI
ID: 5990 · Report as offensive     Reply Quote
shady

Send message
Joined: 1 Sep 04
Posts: 2
Credit: 578,977
RAC: 0
Message 6019 - Posted: 24 Feb 2005, 7:19:12 UTC

It certainly would be usefull to be able to see if any given pc is returning results that are valid or not.

Shady
<img src='http://www.boincsynergy.com/images/stats/comb-1527.jpg'>
ID: 6019 · Report as offensive     Reply Quote
Profile littleBouncer
Avatar

Send message
Joined: 23 Oct 04
Posts: 358
Credit: 1,439,205
RAC: 0
Message 6022 - Posted: 24 Feb 2005, 8:29:15 UTC - in response to Message 5908.  

@ Markku

It seems at Einstein they have discovered something what accords to this "0 CPU-time" Problem (IMO).

[QUOTE.from their Homep.]
Feb 23, 2005
A small bug in the validator has been fixed. This should reduce the number of correct results that were being marked as invalid. Thanks to Steffen Grunewald for spotting and fixing this.

[/QUOTE]

greetz littleBouncer


ID: 6022 · Report as offensive     Reply Quote
Profile Markku Degerholm

Send message
Joined: 3 Sep 04
Posts: 212
Credit: 4,545
RAC: 0
Message 6031 - Posted: 24 Feb 2005, 10:55:57 UTC - in response to Message 6022.  

>
> It seems at Einstein they have discovered something what accords to this "0
> CPU-time" Problem (IMO).
>
> [QUOTE.from their Homep.]
> Feb 23, 2005
> A small bug in the validator has been fixed. This should reduce the number of
> correct results that were being marked as invalid. Thanks to Steffen Grunewald
> for spotting and fixing this.
>
> [/QUOTE]
>
> greetz littleBouncer

Thanks, but this can't be what's happening with us. The CPU time is reported as zero by the core client already, so there is nothing validator can do about it.

Markku Degerholm
LHC@home admin
ID: 6031 · Report as offensive     Reply Quote
Profile littleBouncer
Avatar

Send message
Joined: 23 Oct 04
Posts: 358
Credit: 1,439,205
RAC: 0
Message 6032 - Posted: 24 Feb 2005, 11:09:24 UTC - in response to Message 6031.  
Last modified: 24 Feb 2005, 11:20:04 UTC

> Thanks, but this can't be what's happening with us. The CPU time is reported
> as zero by the core client already, so there is nothing validator can do about
> it.
>
>

First I thought that too (because the statement has to do with the validator), but then I read the forums post "Problems and bugs" and in some thread(*) it seemed to me , that it has to do with our 0 CPU-time-problem.
(*) where linux-cruncher have 0 CPU-time and for that 0 granted credits.

THX anyway for your reply
(shows me that my seeking (or searching for a solution) isn't worthless :-))

[EDIT: @ Markku]
Did you read this note from me? Maybe it is a bit complicate discribed.
[/EDIT]

greetz littleBouncer
ID: 6032 · Report as offensive     Reply Quote
Profile Markku Degerholm

Send message
Joined: 3 Sep 04
Posts: 212
Credit: 4,545
RAC: 0
Message 6049 - Posted: 24 Feb 2005, 15:32:56 UTC - in response to Message 6032.  

> First I thought that too (because the statement has to do with the validator),
> but then I read the forums post "Problems and bugs" and in some thread(*) it
> seemed to me , that it has to do with our 0 CPU-time-problem.
> (*) where linux-cruncher have 0 CPU-time and for that 0 granted credits.

Couldn't find it - I assume you meant this forum: http://einstein.phys.uwm.edu/forum_thread.php?id=893

> [EDIT: @ Markku]
> Did you read <a> href="http://lhcathome.cern.ch/forum_reply.php?thread=1177#6025">this note[/url]
> from me? Maybe it is a bit complicate discribed.
> [/EDIT]

Yes.

Markku Degerholm
LHC@home admin
ID: 6049 · Report as offensive     Reply Quote
Profile littleBouncer
Avatar

Send message
Joined: 23 Oct 04
Posts: 358
Credit: 1,439,205
RAC: 0
Message 6054 - Posted: 24 Feb 2005, 17:14:45 UTC - in response to Message 6049.  


> Couldn't find it - I assume you meant this forum:
> http://einstein.phys.uwm.edu/forum_thread.php?id=893
>
Yes!


greetz littleBouncer
ID: 6054 · Report as offensive     Reply Quote
RichaG
Avatar

Send message
Joined: 1 Sep 04
Posts: 1
Credit: 406,739
RAC: 0
Message 6065 - Posted: 24 Feb 2005, 23:56:40 UTC

At first I thought it was my computer. When I merged two hosts, the results page by host disappeared.

After reading this thread I can see that it wasn't me.

I hope the results page is up soon. I have no idea if there is credit being granted for the new software package.

<br>
RichardG
ID: 6065 · Report as offensive     Reply Quote
1 · 2 · Next

Message boards : Number crunching : Results page


©2024 CERN