> The error message in the first post is extremely useful to us -- this is a
> *reproducable* bug in our analysis code, which only appears under Windows and
> not under Linux or Mac. Two people are currently (as I type) working on
> this.
>
> So please don't give up on E@H. While you may be losing a few credits, you
> are helping us to track down and fix things in the science part of the code,
> which is extremely valuable to us.
>
> Cheers, Bruce
=========
Since this morning none of my other PC's have had this Bug so far Bruce, even the 1 PC that had it 3 WU straight WU's finally Reported the 4'th one ok. All my PC's run Window's XP Pro, several have the SP2 Update & the rest just the SP1 update ...
I had another Computation Error at the end of the WU Run during the night off the same Computer, so thats 4 out of 6 WU's to give me the Error on that Computer. So far it's the only Computer to give me the Error, all the rest of my PC's are not giving me the error up till now.
I have been running BOINC Client v4.66 on all my Computers for the last 3 or 4 days, so I decided to un-install that Version from that Computer & try the newly released v4.20 on it to see if I still get the Computation Errors on it. If that don't work the I'm going to Reset the Project on that Computer & get a fresh set of WU's for it to see if I still get the Error on it.
I'm beginning to wonder if it could be the CPU on that particular Computer. All my Computers are Northwood Intel HT P4's with HT Capability except for that one. That Computer has a Prescott Intel P4 with HT Capability ... Just a thought ... ???
Below are listed the 4 WU's that Errored at the end of the run in the order that I got them, their pretty much the same error it seems like ...
==========
4.66 The system cannot find the file specified. (0x2) - exit code 2 (0x2)
Line 24281 of file Fstats.Ha is too long or has no NEWLINE. First 255 chars are:
368.598000000000 0.20047987 0.06920367 70 12185575844078799000000000000000000000000000000000000000000.00000 37877301887353856000000000000000000000000000000000000000000.00000 199084999298019520000000000000000000000000000000000000000000.00000000000000000
==========
4.66 The system cannot find the file specified. (0x2) - exit code 2 (0x2)
Line 23295 of file Fstats.Ha is too long or has no NEWLINE. First 255 chars are:
368.698000000000 0.20047987 0.06920367 73 22689416281690087000000000000000000000000000000000000000000.00000 63742735344024206000000000000000000000000000000000000000000.00000 313839249739090890000000000000000000000000000000000000000000.00000000000000000
===========
4.66 The system cannot find the file specified. (0x2) - exit code 2 (0x2)
Line 3907 of file Fstats.Ha is too long or has no NEWLINE. First 255 chars are:
368.798000000000 0.22026658 0.04920367 71 15839407291719011000000000000000000000000000000000000000000.00000 46799844498547902000000000000000000000000000000000000000000.00000 237975688506274870000000000000000000000000000000000000000000.00000000000000000
==========
4.66 The system cannot find the file specified. (0x2) - exit code 2 (0x2)
Line 16118 of file Fstats.Ha is too long or has no NEWLINE. First 255 chars are:
368.598000000000 6.24036369 -0.01079633 67 5432106848858945400000000000000000000000000000000000000000.00000 19412325260896811000000000000000000000000000000000000000000.00000 111568029230606070000000000000000000000000000000000000000000.00000000000000000
Of course we sort out the bugs and problems of our Apps that happen all the time before putting them on the server for the larger public. So the bugs that remain to our test users are ones that occur only on certain types of machines or with certain data.
The scheduler we set up prefers to send work to you according to the files you already have to possibly avoid downloading new large data chunks. However this also means that the workunits you get are similar to the ones you already have crunched.
These two facts lead to the effect you are seing now: having stumbled over a data-driven bug you will likely get another Result that shows the same problem. The way to avoid this would be to reset the project, causing a new download of a (most likey) different data set.
However we are very glad that this problem showed up, it is reproducable and I myself and some other people around me too are working hard to solve it, so that the users (and of course our results) are not affected by it in the public case.
These two facts lead to the effect you are seing now: having stumbled over a data-driven bug you will likely get another Result that shows the same problem. The way to avoid this would be to reset the project, causing a new download of a (most likey) different data set.
==========
I had already thought of that "Resetting the Project on that Computer" but I wanted to try a different version of BOINC first to see what happened. I'm within a few Hours of the next WU Completion so I'll see what happens.
If I continue to get the Error then I'll just Reset & get a fresh load of WU's for that Computer. Like I said already so far it's the only PC I have that has showed any errors and it only started with the v4.75 client so I figured it was a bad batch of WU's & not my Computer ...
Well I Reset the Project on that Computer because it just showed another Computation Error at the end of the run. But I don't think it did me any good to do that either because the Server just sent me the same # WU's that were in line as the ones I just Reset from.
Twice I reset it and twice it sent close to what I just got rid of. In fact I'm almost sure it sent the same ones back the second time I Reset the Project. So now I won't know for 10-11 hours if these are any good or not ... But I'll go ahead and run a couple of them to see what happens.
If that Computer still gets the Computation Errors then I'm just going to run another Project with it for awhile until this matter is straightened out ...
Einstein@Home - 2005-02-11 12:59:15 - Resuming result H1_0417.9__0418.4_0.1_T03_Test02_2 using einstein version 4.75
Einstein@Home - 2005-02-11 13:09:22 - Unrecoverable error for result H1_0417.9__0418.4_0.1_T03_Test02_2 (Das System kann die angegebene Datei nicht finden. (0x2) - exit code 2 (0x2))
Einstein@Home - 2005-02-11 13:09:22 - Computation for result H1_0417.9__0418.4_0.1_T03_Test02 finished
Ok i try to use the new Apllication, but i can not make any update, if i try to get the application by using the "update" Function in BOINc nothing will happen.
How can i get the new Application ?
Greetings from Wetzlar in Germany
Sascha Bickel
Admin, Teamleader CPDN & Einstein
Team Science and Research Hessen (SaR Hessen) http://www.sar-hessen.de
> The error message in the
)
> The error message in the first post is extremely useful to us -- this is a
> *reproducable* bug in our analysis code, which only appears under Windows and
> not under Linux or Mac. Two people are currently (as I type) working on
> this.
>
> So please don't give up on E@H. While you may be losing a few credits, you
> are helping us to track down and fix things in the science part of the code,
> which is extremely valuable to us.
>
> Cheers, Bruce
=========
Since this morning none of my other PC's have had this Bug so far Bruce, even the 1 PC that had it 3 WU straight WU's finally Reported the 4'th one ok. All my PC's run Window's XP Pro, several have the SP2 Update & the rest just the SP1 update ...
I had another Computation
)
I had another Computation Error at the end of the WU Run during the night off the same Computer, so thats 4 out of 6 WU's to give me the Error on that Computer. So far it's the only Computer to give me the Error, all the rest of my PC's are not giving me the error up till now.
I have been running BOINC Client v4.66 on all my Computers for the last 3 or 4 days, so I decided to un-install that Version from that Computer & try the newly released v4.20 on it to see if I still get the Computation Errors on it. If that don't work the I'm going to Reset the Project on that Computer & get a fresh set of WU's for it to see if I still get the Error on it.
I'm beginning to wonder if it could be the CPU on that particular Computer. All my Computers are Northwood Intel HT P4's with HT Capability except for that one. That Computer has a Prescott Intel P4 with HT Capability ... Just a thought ... ???
Below are listed the 4 WU's that Errored at the end of the run in the order that I got them, their pretty much the same error it seems like ...
==========
4.66 The system cannot find the file specified. (0x2) - exit code 2 (0x2)
Line 24281 of file Fstats.Ha is too long or has no NEWLINE. First 255 chars are:
368.598000000000 0.20047987 0.06920367 70 12185575844078799000000000000000000000000000000000000000000.00000 37877301887353856000000000000000000000000000000000000000000.00000 199084999298019520000000000000000000000000000000000000000000.00000000000000000
==========
4.66 The system cannot find the file specified. (0x2) - exit code 2 (0x2)
Line 23295 of file Fstats.Ha is too long or has no NEWLINE. First 255 chars are:
368.698000000000 0.20047987 0.06920367 73 22689416281690087000000000000000000000000000000000000000000.00000 63742735344024206000000000000000000000000000000000000000000.00000 313839249739090890000000000000000000000000000000000000000000.00000000000000000
===========
4.66 The system cannot find the file specified. (0x2) - exit code 2 (0x2)
Line 3907 of file Fstats.Ha is too long or has no NEWLINE. First 255 chars are:
368.798000000000 0.22026658 0.04920367 71 15839407291719011000000000000000000000000000000000000000000.00000 46799844498547902000000000000000000000000000000000000000000.00000 237975688506274870000000000000000000000000000000000000000000.00000000000000000
==========
4.66 The system cannot find the file specified. (0x2) - exit code 2 (0x2)
Line 16118 of file Fstats.Ha is too long or has no NEWLINE. First 255 chars are:
368.598000000000 6.24036369 -0.01079633 67 5432106848858945400000000000000000000000000000000000000000.00000 19412325260896811000000000000000000000000000000000000000000.00000 111568029230606070000000000000000000000000000000000000000000.00000000000000000
Poor Boy, Of course we
)
Poor Boy,
Of course we sort out the bugs and problems of our Apps that happen all the time before putting them on the server for the larger public. So the bugs that remain to our test users are ones that occur only on certain types of machines or with certain data.
The scheduler we set up prefers to send work to you according to the files you already have to possibly avoid downloading new large data chunks. However this also means that the workunits you get are similar to the ones you already have crunched.
These two facts lead to the effect you are seing now: having stumbled over a data-driven bug you will likely get another Result that shows the same problem. The way to avoid this would be to reset the project, causing a new download of a (most likey) different data set.
However we are very glad that this problem showed up, it is reproducable and I myself and some other people around me too are working hard to solve it, so that the users (and of course our results) are not affected by it in the public case.
At least, that's what testing is for.
Thanks a lot for your cooperation!
BM
BM
These two facts lead to the
)
These two facts lead to the effect you are seing now: having stumbled over a data-driven bug you will likely get another Result that shows the same problem. The way to avoid this would be to reset the project, causing a new download of a (most likey) different data set.
==========
I had already thought of that "Resetting the Project on that Computer" but I wanted to try a different version of BOINC first to see what happened. I'm within a few Hours of the next WU Completion so I'll see what happens.
If I continue to get the Error then I'll just Reset & get a fresh load of WU's for that Computer. Like I said already so far it's the only PC I have that has showed any errors and it only started with the v4.75 client so I figured it was a bad batch of WU's & not my Computer ...
Well I Reset the Project on
)
Well I Reset the Project on that Computer because it just showed another Computation Error at the end of the run. But I don't think it did me any good to do that either because the Server just sent me the same # WU's that were in line as the ones I just Reset from.
Twice I reset it and twice it sent close to what I just got rid of. In fact I'm almost sure it sent the same ones back the second time I Reset the Project. So now I won't know for 10-11 hours if these are any good or not ... But I'll go ahead and run a couple of them to see what happens.
If that Computer still gets the Computation Errors then I'm just going to run another Project with it for awhile until this matter is straightened out ...
We have just found the bug in
)
We have just found the bug in the code and are about building new apps now. Shouldn't take too long...
BM
BM
Take me in for
)
Take me in for that:
Einstein@Home - 2005-02-11 12:59:15 - Resuming result H1_0417.9__0418.4_0.1_T03_Test02_2 using einstein version 4.75
Einstein@Home - 2005-02-11 13:09:22 - Unrecoverable error for result H1_0417.9__0418.4_0.1_T03_Test02_2 (Das System kann die angegebene Datei nicht finden. (0x2) - exit code 2 (0x2))
Einstein@Home - 2005-02-11 13:09:22 - Computation for result H1_0417.9__0418.4_0.1_T03_Test02 finished
:{
Aloha, Uli
> We have just found the bug
)
> We have just found the bug in the code and are about building new apps now.
> Shouldn't take too long...
Out now.
BM
BM
Ok i try to use the new
)
Ok i try to use the new Apllication, but i can not make any update, if i try to get the application by using the "update" Function in BOINc nothing will happen.
How can i get the new Application ?
Greetings from Wetzlar in Germany
Sascha Bickel
Admin, Teamleader CPDN & Einstein
Team Science and Research Hessen (SaR Hessen)
http://www.sar-hessen.de
When Boinc downloads a new
)
When Boinc downloads a new workunit you will automatically get the new application with that WU. The only thing you have to do is wait. ;-)
(or, if you have one of the "problem WU's" now, you can reset the project)