No Credit For A Clean Run?


Advanced search

Message boards : SZTAKI Desktop Grid : No Credit For A Clean Run?

AuthorMessage
[B^S] DAMedek
Send message
Joined: Nov 23 06
Posts: 3
Credit: 23,716
RAC: 0
Message 5942 - Posted 19 Mar 2007 15:28:36 UTC

    I uploaded results for several SZTAKI WUs yesterday, and when I checked on their status today I noticed that I got 0.0 credit for <a href=\"http://szdg.lpds.sztaki.hu/szdg/workunit.php?wuid=114334\">WU 114334</a>. This concerns me because I know this WU had a clean run . . . It ran through from start to finish with no interruption . . . and the stderr_txt output (<a href=\"http://szdg.lpds.sztaki.hu/szdg/result.php?resultid=647457\">Result ID 647457</a>) shows no problems.

    I\'ve had problems getting credit at Leiden Classical because my operating system is Windows XP Home Edition, but this doesn\'t seem to be the problem here as I get credit for other SZTAKI WUs. Also, for this WU, one of the other results that got credit was generated on an XP Home machine.

    Can anyone enlighten me on the cause of this glitch?


    ____________

    Profile [B^S] Gamma^Ray
    Send message
    Joined: Jan 3 07
    Posts: 18
    Credit: 5,977
    RAC: 0
    Message 5944 - Posted 19 Mar 2007 18:40:23 UTC

      Last modified: 19 Mar 2007 18:41:49 UTC

      Well, I doubt it. And don\'t expect any kind of answers from the Adm. on it either. You can read this thread to get a better idea why :

      Output Errors

      G^R
      ____________

      Stick
      Send message
      Joined: Jun 12 06
      Posts: 193
      Credit: 66,271
      RAC: 0
      Message 5945 - Posted 19 Mar 2007 18:55:21 UTC - in response to Message 5942.

        Last modified: 19 Mar 2007 19:39:00 UTC

        I just noticed a similar problem with one of my WU\'s. Mine is the result with 0.00 credit (and, I believe it was validated today). You may want to read the Any update on the Validator server? thread. Obviously, the project is having problems with the validator and our WU\'s may have been impacted by it. Note that the SZTAKI Desktop Grid Status Page shows the validator as \"Running\" right now. It has not been running for most of the last two weeks.

        EDIT: I disagree with Gamma^Ray. These do not appear to be related to the \"Output is Empty\" problem.

        EDIT2: Two more WU\'s with similar 0.00 results: 134697 and 130460.

        I uploaded results for several SZTAKI WUs yesterday, and when I checked on their status today I noticed that I got 0.0 credit for <a href=\"http://szdg.lpds.sztaki.hu/szdg/workunit.php?wuid=114334\">WU 114334</a>. This concerns me because I know this WU had a clean run . . . It ran through from start to finish with no interruption . . . and the stderr_txt output (<a href=\"http://szdg.lpds.sztaki.hu/szdg/result.php?resultid=647457\">Result ID 647457</a>) shows no problems.

        I\'ve had problems getting credit at Leiden Classical because my operating system is Windows XP Home Edition, but this doesn\'t seem to be the problem here as I get credit for other SZTAKI WUs. Also, for this WU, one of the other results that got credit was generated on an XP Home machine.

        Can anyone enlighten me on the cause of this glitch?



        ____________

        Profile [B^S] Gamma^Ray
        Send message
        Joined: Jan 3 07
        Posts: 18
        Credit: 5,977
        RAC: 0
        Message 5947 - Posted 19 Mar 2007 22:03:00 UTC - in response to Message 5945.

          Last modified: 19 Mar 2007 22:04:47 UTC



          EDIT: I disagree with Gamma^Ray. These do not appear to be related to the \"Output is Empty\" problem.


          Nowhere did I say that the issue is relatable to the Output is Empty problem, I was merely giving comment on his question of \"Can anyone enlighten me\" based on the link I provided, And really several others with problems that go unresponded. I do believe though there is another issue with the validater that is related as both his problem, And the link I posted both ended with zero to little credit after lengthy run times. Why ? I have no clue. But he can happily wait for an answer from the Adm. It would be nice to get one concerning many of these errors, But it might be a long wait.




          ____________

          Stick
          Send message
          Joined: Jun 12 06
          Posts: 193
          Credit: 66,271
          RAC: 0
          Message 5948 - Posted 19 Mar 2007 22:24:46 UTC - in response to Message 5947.

            Nowhere did I say that the issue is relatable to the Output is Empty problem . . .


            Sorry! I misread your reply.

            But he can happily wait for an answer from the Adm. It would be nice to get one concerning many of these errors, But it might be a long wait.


            Nobody can disagree with that.

            ____________

            [B^S] DAMedek
            Send message
            Joined: Nov 23 06
            Posts: 3
            Credit: 23,716
            RAC: 0
            Message 5950 - Posted 20 Mar 2007 13:34:24 UTC

              Thanks, Gamma^Ray and Stick. I\'ve been reading the message boards ever since I joined SZTAKI, and yes, I\'m quite familiar with the \"empty output\" problem . . . have experienced that myself with restarted WUs. That\'s one reason I run SZTAKI WUs from start to finish with no interruptions. I\'ve also experienced other minor problems, but never a clean SZTAKI run with no credit.

              I\'ll chalk this one up to \"It\'s A Mystery\" and keep on crunching . . . but of course I\'ll be keeping an eye on validations and won\'t hold my breath waiting for the Admins to fix something I don\'t yet understand.

              Again, my thanks.

              ____________

              Profile [B^S] Gamma^Ray
              Send message
              Joined: Jan 3 07
              Posts: 18
              Credit: 5,977
              RAC: 0
              Message 5956 - Posted 21 Mar 2007 18:29:47 UTC

                \"Life is but a Mystery\" LoL .. Best of luck to ya ! :)

                Regards,
                G^R
                ____________

                Stick
                Send message
                Joined: Jun 12 06
                Posts: 193
                Credit: 66,271
                RAC: 0
                Message 5960 - Posted 21 Mar 2007 23:49:21 UTC - in response to Message 5945.

                  EDIT2: Two more WU\'s with similar 0.00 results: 134697 and 130460.


                  I was looking at wuid=134697 again and noticed that the two results which received credit were both restarted several times. My result (which did not receive credit) processed all the way through without interreuption (i.e. a clean run). This makes me wonder if the two units which received credit were matched by the validator because they happened to encounter the checkpointing bug on the same line (or lines). That is, did two faulty results match and validate? And, could my \"invalid\" result be the only really correct one of the group?

                  I\'ve raised the question of this possibility previously. I am raising it again because the batch of (recently validated) WU\'s (for my account anyway) seems to have an inordinately high incidence of 0.00 results and I was wondering how many of these 0.00 results might fall into this category. Has anyone else seen any WU\'s where this might have happened?
                  ____________

                  Vid Vidmar*
                  Avatar
                  Send message
                  Joined: Jul 12 05
                  Posts: 7
                  Credit: 23,827
                  RAC: 0
                  Message 5961 - Posted 22 Mar 2007 9:40:05 UTC

                    Well I just wasted 232905 seconds on this WU only to get 0.0 credits. This time I\'m really leaving here. Will check from time to time, to see if things are any better.
                    See ya\'!
                    ____________

                    robert.mouris
                    Send message
                    Joined: Nov 3 05
                    Posts: 129
                    Credit: 4,124,194
                    RAC: 0
                    Message 5965 - Posted 22 Mar 2007 19:56:21 UTC - in response to Message 5960.

                      My result (which did not receive credit) processed all the way through without interreuption (i.e. a clean run). This makes me wonder if the two units which received credit were matched by the validator because they happened to encounter the checkpointing bug on the same line (or lines). That is, did two faulty results match and validate? And, could my \"invalid\" result be the only really correct one of the group?


                      Stick, if you were right with this assumption (and I\'m afraid you are), it would mean that the mathematicians will base their research on wrong data. I don\'t know the process of their theoretical work, but in some areas one single data can lead to wrong conclusions. This project is not like empirical physics where an experiment can and must be repeated many times. Here the processing of a WU with a canonical result is definitely finished and supposed to be correct. And saying that 2 persons who agree are right, just because they outnumber that single person, can\'t work. History tells as that sometimes a scientist proves to be right, even if everyone else says he\'s wrong or crazy. So, should all these WUs be reprocessed, and if yes, how to select them if all records are deleted apart from the canonical result?

                      The project raises the opposite mistake: missing a correct result. This, however, doesn\'t seem to be an issue. I quote from here: \"Further work in this area includes two immediate generalizations. On one hand, dimensions higher than 12 could be attacked by the help of a currently unproved mathematical conjecture. This conjecture could reduce the size of the parameter space at hand so that a list of polynomials can be obtained. This list will not be 100% sure to be complete but from a practical point of view, a possibly uncomplete list is still valuable. This program is now executed on the grid for degree 12.\"


                      ____________

                      Stick
                      Send message
                      Joined: Jun 12 06
                      Posts: 193
                      Credit: 66,271
                      RAC: 0
                      Message 5968 - Posted 23 Mar 2007 0:29:21 UTC - in response to Message 5965.

                        Last modified: 23 Mar 2007 0:30:36 UTC

                        . . . it would mean that the mathematicians will base their research on wrong data . . . can lead to wrong conclusions.


                        Robert,

                        Thank you for the reply! Hopefully, my \"flawed validator\" conjecture is not a real possibility. Although, from what we know about BOINC validators in general, the idea certainly seems plausible. And (as you said), if I am right, the project\'s research and conclusions become suspect.

                        Right now, it appears there are other, more pressing, issues with the validator. (The fact that it is frequently being taken off-line seems to imply that anyway. The abundance of recent 0.00 results may be another indicator.) But, hopefully, Adam will look at this issue as well. And, until he says something to refute the idea, I guess we should assume that it is possible to validate faulty results.

                        Stick
                        ____________

                        [B^S] DAMedek
                        Send message
                        Joined: Nov 23 06
                        Posts: 3
                        Credit: 23,716
                        RAC: 0
                        Message 5969 - Posted 23 Mar 2007 1:40:24 UTC

                          I started this thread a few days ago . . . but now I\'m really miffed.

                          Another WU of mine, 114908, finished clean and completed the quorum. I claimed 37.90 credits while two other crunchers claimed 40.27 and 75.17 credits, respectively. The amount of credit awarded? . . . only 0.04 credits for each of us . . . the same amount claimed by another user with an outcome of \"Client Error\"!

                          That\'s it for me, folks, I\'m not crunching SZTAKI for awhile. . . .

                          Maybe if the admin cares enough someday to fix these types of problems, I\'ll care enough to return to SZTAKI. Otherwise, I\'ll be crunching for someone else.

                          -- DAMedek

                          Gaurav
                          Send message
                          Joined: Sep 4 06
                          Posts: 22
                          Credit: 52,275
                          RAC: 0
                          Message 5970 - Posted 23 Mar 2007 3:29:52 UTC

                            All computers that crunched the wu sucessfully claimed 200+ credit. granted - 69.79 (same as a computer whose outcome was client error).
                            This WU
                            ____________

                            miketoth1001
                            Avatar
                            Send message
                            Joined: Apr 3 06
                            Posts: 20
                            Credit: 27,967
                            RAC: 0
                            Message 5977 - Posted 24 Mar 2007 17:38:44 UTC

                              Now, why would a client error even have credits? Half of his WUs are client errors.

                              Profile Nightbird
                              Forum moderator
                              Avatar
                              Send message
                              Joined: Jul 12 05
                              Posts: 920
                              Credit: 114,924
                              RAC: 0
                              Message 5978 - Posted 24 Mar 2007 19:11:31 UTC - in response to Message 5977.

                                Last modified: 24 Mar 2007 19:12:40 UTC

                                Now, why would a client error even have credits? Half of his WUs are client errors.

                                Because credits have been granted manually (with a script), i guess.
                                ____________

                                Profile [B^S] Gamma^Ray
                                Send message
                                Joined: Jan 3 07
                                Posts: 18
                                Credit: 5,977
                                RAC: 0
                                Message 5980 - Posted 25 Mar 2007 19:48:42 UTC - in response to Message 5978.

                                  Now, why would a client error even have credits? Half of his WUs are client errors.

                                  Because credits have been granted manually (with a script), i guess.


                                  IF that was the case, Then that would mean that Adam is picking users/wu\'s to give credit to with errors based on what ? As many users have not gotten credit for wu\'s finished that have errored out (And some that ran sucessfully I do believe LoL).

                                  G^R

                                  ____________

                                  Stick
                                  Send message
                                  Joined: Jun 12 06
                                  Posts: 193
                                  Credit: 66,271
                                  RAC: 0
                                  Message 5982 - Posted 25 Mar 2007 20:36:33 UTC - in response to Message 5978.

                                    Last modified: 25 Mar 2007 20:39:05 UTC

                                    Now, why would a client error even have credits? Half of his WUs are client errors.

                                    Because credits have been granted manually (with a script), i guess.


                                    An alternate explanation is that the validator program did it. It is off-line now but was running early last week (just prior to this thread being started). Since then it has been \"on and back offline\" at least once. And, it seems the problem symptoms reported here have changed slightly. Maybe Adam has been \"tinkering\" but hasn\'t gotten the bugs out yet. (Hopefully, he is reading this thread - even if he isn\'t replying to it.)
                                    ____________

                                    NJMHoffmann
                                    Send message
                                    Joined: May 19 06
                                    Posts: 18
                                    Credit: 10,221
                                    RAC: 0
                                    Message 5983 - Posted 26 Mar 2007 9:15:54 UTC - in response to Message 5977.

                                      Last modified: 26 Mar 2007 9:18:10 UTC

                                      Now, why would a client error even have credits? Half of his WUs are client errors.

                                      Because most of the \"client errors\" are in reality \"application bugs\"?

                                      Norbert
                                      ____________

                                      Gaurav
                                      Send message
                                      Joined: Sep 4 06
                                      Posts: 22
                                      Credit: 52,275
                                      RAC: 0
                                      Message 5999 - Posted 29 Mar 2007 1:53:07 UTC

                                        Last modified: 29 Mar 2007 2:00:13 UTC

                                        Both wus claimed 116 credit but 0 & 0.47 granted.

                                        http://szdg.lpds.sztaki.hu/szdg/workunit.php?wuid=13576
                                        http://szdg.lpds.sztaki.hu/szdg/workunit.php?wuid=11035

                                        this wu claimed 267 & granted 1.78

                                        http://szdg.lpds.sztaki.hu/szdg/workunit.php?wuid=13615

                                        this one claimed 39.24 & granted 0

                                        http://szdg.lpds.sztaki.hu/szdg/workunit.php?wuid=14099

                                        claimed -643 granted -0

                                        http://szdg.lpds.sztaki.hu/szdg/workunit.php?wuid=114652
                                        ____________

                                        Profile Nightbird
                                        Forum moderator
                                        Avatar
                                        Send message
                                        Joined: Jul 12 05
                                        Posts: 920
                                        Credit: 114,924
                                        RAC: 0
                                        Message 6068 - Posted 11 Apr 2007 20:53:42 UTC

                                          Last modified: 15 Apr 2007 8:40:30 UTC

                                          I don\'t know what\'s happened but ...
                                          look here please
                                          http://szdg.lpds.sztaki.hu/szdg/workunit.php?wuid=115205

                                          result 514162 (valid) (canonical result)
                                          claimed credit : 306.98
                                          granted credit : 177.85

                                          177.85 ? what\'s ?
                                          result 514160 (Invalid -> <message>
                                          - exit code -1073741502 (0xc0000142)
                                          </message>)
                                          claimed credit : 177.85
                                          granted credit : - (?) -> got 177.85

                                          so, a valid wu got credits from an invalid wu

                                          then,
                                          result 514161 (invalid, according the validator i guess))
                                          claimed credit : 304.96
                                          granted credit : 304.96

                                          claimed credit = granted credit

                                          and the fourth :
                                          result 763143 (invalid, according the validator)
                                          claimed credit = 174.05
                                          granted credit = 0 credit

                                          What\'s this BIG mess ??

                                          ____________

                                          robert.mouris
                                          Send message
                                          Joined: Nov 3 05
                                          Posts: 129
                                          Credit: 4,124,194
                                          RAC: 0
                                          Message 6080 - Posted 15 Apr 2007 7:29:58 UTC

                                            This one is also strange. The validation is still pending, but the validator has found the canonical result! The credits are pending, but on top of the page it says that 0.88 credits are awarded. How is this to be understood?

                                            Apart from having tested (with disappointment) the new release 2 weeks ago, I have been longing since the beginning of February to come back. It seems that I must be ready to wait for a very long time more...
                                            ____________

                                            Profile Nightbird
                                            Forum moderator
                                            Avatar
                                            Send message
                                            Joined: Jul 12 05
                                            Posts: 920
                                            Credit: 114,924
                                            RAC: 0
                                            Message 6081 - Posted 15 Apr 2007 8:43:56 UTC - in response to Message 6080.

                                              Last modified: 15 Apr 2007 8:45:22 UTC

                                              This one is also strange. The validation is still pending, but the validator has found the canonical result! The credits are pending, but on top of the page it says that 0.88 credits are awarded. How is this to be understood?

                                              Apart from having tested (with disappointment) the new release 2 weeks ago, I have been longing since the beginning of February to come back. It seems that I must be ready to wait for a very long time more...

                                              People will get credits from the result 516875, it\'s invalid but it got 0.88 credits (= claimed credit)...

                                              ____________

                                              Stick
                                              Send message
                                              Joined: Jun 12 06
                                              Posts: 193
                                              Credit: 66,271
                                              RAC: 0
                                              Message 6083 - Posted 15 Apr 2007 14:20:56 UTC - in response to Message 6080.

                                                Apart from having tested (with disappointment) the new release 2 weeks ago, I have been longing since the beginning of February to come back. It seems that I must be ready to wait for a very long time more...


                                                Me, too! In fact, I have been running the project some - but very carefully (by allowing new work, one unit at a time, then going back to \"no new work\"). I have run several units with the newest version (v2.06) but I have not tried to test it for the checkpoint bug yet. (Has anyone?) The validator anomalies being reported are certainly strange and humorous (Nightbird\'s being the most outstanding one). And, it seems, there has been very little progress on that front. So, unfortunately, I agree that we \"must be ready to wait for a very long time more...\"
                                                ____________

                                                Profile jamin
                                                Avatar
                                                Send message
                                                Joined: Nov 20 06
                                                Posts: 324
                                                Credit: 100,110
                                                RAC: 0
                                                Message 6084 - Posted 15 Apr 2007 20:03:30 UTC

                                                  I\'ve been trying to test the checkpointing on my 2 last long WUs - seems the application is working alright - but I\'m not quite sure as the results from others looks different :-/
                                                  ____________
                                                  Bang! Zoom! Right in the kisser!

                                                  Stick
                                                  Send message
                                                  Joined: Jun 12 06
                                                  Posts: 193
                                                  Credit: 66,271
                                                  RAC: 0
                                                  Message 6097 - Posted 17 Apr 2007 20:36:29 UTC - in response to Message 6084.

                                                    Last modified: 17 Apr 2007 20:48:33 UTC

                                                    I\'ve been trying to test the checkpointing on my 2 last long WUs - seems the application is working alright - but I\'m not quite sure as the results from others looks different :-/


                                                    Well, I wasn\'t really trying to test for the checkpointing bug - but I had an unrelated problem with my computer and had to do a restart. This WU/Result had been running for about 2 hours but was still showing 0.00% progress. When it resumed after the restart it immediately jumped to 20%. So it appears the checkpoint bug is still there. However, since that initial restart, I have restarted BOINC 4 or 5 more times and the WU only jumped ahead once (to 40%). (BTW: The WU has now processed over 9 hours.) Therefore, I am inclined to think that v2.06 is somehow improved over previous versions. Although, I am virtually certain that my result will be invalid (or, at least, it should be), I will continue processing until it finishes (including more forced restarts) in the hopes that its stderr.txt file might be helpful.
                                                    ____________

                                                    Profile jamin
                                                    Avatar
                                                    Send message
                                                    Joined: Nov 20 06
                                                    Posts: 324
                                                    Credit: 100,110
                                                    RAC: 0
                                                    Message 6103 - Posted 18 Apr 2007 7:46:22 UTC

                                                      Last modified: 18 Apr 2007 7:47:21 UTC

                                                      As I thought - I was given 0 credit for the WU (don\'t really think it\'s connected with checkpointing as two others restarted more times as my result :-/)

                                                      edit - 32h of crunching down the drain that is ;)
                                                      ____________
                                                      Bang! Zoom! Right in the kisser!

                                                      Stick
                                                      Send message
                                                      Joined: Jun 12 06
                                                      Posts: 193
                                                      Credit: 66,271
                                                      RAC: 0
                                                      Message 6107 - Posted 18 Apr 2007 12:10:10 UTC - in response to Message 6097.

                                                        Last modified: 18 Apr 2007 12:42:23 UTC

                                                        Well, I wasn\'t really trying to test for the checkpointing bug - but I had an unrelated problem with my computer and had to do a restart. This WU/Result had been running for about 2 hours but was still showing 0.00% progress. When it resumed after the restart it immediately jumped to 20%. So it appears the checkpoint bug is still there. However, since that initial restart, I have restarted BOINC 5 or 6 more times and the WU only jumped ahead once (to 40%). (BTW: The WU has now processed over 9 hours.) Therefore, I am inclined to think that v2.06 is somehow improved over previous versions. Although, I am virtually certain that my result will be invalid (or, at least, it should be), I will continue processing until it finishes (including more forced restarts) in the hopes that its stderr.txt file might be helpful.


                                                        Result 830385 is now finished. When I looked at its stderr info and I saw something that may be worth noting. That is, I noticed the checkpoint problem after the first restart (when progress \"jumped\" from 0.00% to 20.00%). But, the stderr data says:

                                                        APP: Restarting from checkpoint: Lines processed so far 1
                                                        APP: Restarting line 1

                                                        At that point, I don\'t think the program was really restarting line 1. I think it was starting line 2.

                                                        Following that restart, there are 5 more BOINC restarts, which I induced (trying to force it to \"jump\" ahead again). These restarts seemed to work properly (progress did not \"jump\" ahead). All 5 reported the exact same restart message shown above.

                                                        However, the 6th restart caused the second \"jump\" (to 40%). It may be a coincidence, but both times the program \"jumped\" ahead, I had restarted my computer. The other times, I had only exited and restarted BOINC.

                                                        As I indicated in my earlier post, I think v2.06 may be improved over previous versions. It looks like it still has checkpointing problems but, from my limited observations, they may occur less frequently.
                                                        ____________

                                                        Stick
                                                        Send message
                                                        Joined: Jun 12 06
                                                        Posts: 193
                                                        Credit: 66,271
                                                        RAC: 0
                                                        Message 6129 - Posted 20 Apr 2007 17:53:17 UTC - in response to Message 6107.

                                                          Last modified: 20 Apr 2007 17:55:24 UTC

                                                          I am not sure what to think. WU 204629 reached a quorom and was validated. And, surprisingly, my result 830385 was valid and received credit. Maybe v2.06 has somehow overcome the checkpoint/empty output problem. On the other hand, maybe the validator still has bugs in it. It sure would be nice to hear from Adam on this issue.

                                                          Well, I wasn\'t really trying to test for the checkpointing bug - but I had an unrelated problem with my computer and had to do a restart. This WU/Result had been running for about 2 hours but was still showing 0.00% progress. When it resumed after the restart it immediately jumped to 20%. So it appears the checkpoint bug is still there. However, since that initial restart, I have restarted BOINC 5 or 6 more times and the WU only jumped ahead once (to 40%). (BTW: The WU has now processed over 9 hours.) Therefore, I am inclined to think that v2.06 is somehow improved over previous versions. Although, I am virtually certain that my result will be invalid (or, at least, it should be), I will continue processing until it finishes (including more forced restarts) in the hopes that its stderr.txt file might be helpful.


                                                          Result 830385 is now finished. When I looked at its stderr info and I saw something that may be worth noting. That is, I noticed the checkpoint problem after the first restart (when progress \"jumped\" from 0.00% to 20.00%). But, the stderr data says:

                                                          APP: Restarting from checkpoint: Lines processed so far 1
                                                          APP: Restarting line 1

                                                          At that point, I don\'t think the program was really restarting line 1. I think it was starting line 2.

                                                          Following that restart, there are 5 more BOINC restarts, which I induced (trying to force it to \"jump\" ahead again). These restarts seemed to work properly (progress did not \"jump\" ahead). All 5 reported the exact same restart message shown above.

                                                          However, the 6th restart caused the second \"jump\" (to 40%). It may be a coincidence, but both times the program \"jumped\" ahead, I had restarted my computer. The other times, I had only exited and restarted BOINC.

                                                          As I indicated in my earlier post, I think v2.06 may be improved over previous versions. It looks like it still has checkpointing problems but, from my limited observations, they may occur less frequently.


                                                          ____________

                                                          Profile Ananas
                                                          Send message
                                                          Joined: Jul 12 05
                                                          Posts: 222
                                                          Credit: 665,833
                                                          RAC: 0
                                                          Message 6174 - Posted 3 May 2007 17:39:33 UTC

                                                            Last modified: 3 May 2007 18:26:51 UTC

                                                            Same here, non-interrupted result (stay in memory) :

                                                            http://szdg.lpds.sztaki.hu/szdg/result.php?resultid=953340

                                                            The [s]randomizer[/s] validator seems to be still crappy,
                                                            second attempt to return after several weeks, I aborted all
                                                            remaining WUs and will try again in a few months.

                                                            The project still seems to be pre-alpha quality.


                                                            p.s.: the other results in this WU have a bunch of restarts,
                                                            repeating some line over and over again, strange that those
                                                            are valid, while an continous run is invalid.

                                                            If the value at the end of stderr.txt means anything, all
                                                            three results must be different.


                                                            p.p.s.: If I was the only one with 0 credits, I would look
                                                            for the problem on my side - but there are way too many
                                                            non-validating results on lots of hosts to be a specific
                                                            host problem

                                                            Stick
                                                            Send message
                                                            Joined: Jun 12 06
                                                            Posts: 193
                                                            Credit: 66,271
                                                            RAC: 0
                                                            Message 6232 - Posted 9 May 2007 14:30:27 UTC

                                                              This WU was vaidated today. I noticed that my result, which processed all the way through without interruption, received zero credit. However, both of the results which did receive credit were restarted on the same line. This is another example of a possible validator flaw which I have pointed out several times - most recently in this message. That is, did the two results which received credit match because they encountered the checkpointing bug on the same line? Were these results truly valid? Or, just matching? And, was my result truly invalid?
                                                              ____________

                                                              Stick
                                                              Send message
                                                              Joined: Jun 12 06
                                                              Posts: 193
                                                              Credit: 66,271
                                                              RAC: 0
                                                              Message 6237 - Posted 10 May 2007 18:32:16 UTC - in response to Message 6232.

                                                                Last modified: 10 May 2007 18:37:31 UTC

                                                                Here is another one. My, uninterrupted, result got zero. The other two, which received credit, were both restarted on line 4. (This one restarted line 4 only. This one also restarted line 1 several times.)


                                                                This WU was vaidated today. I noticed that my result, which processed all the way through without interruption, received zero credit. However, both of the results which did receive credit were restarted on the same line. This is another example of a possible validator flaw which I have pointed out several times - most recently in this message. That is, did the two results which received credit match because they encountered the checkpointing bug on the same line? Were these results truly valid? Or, just matching? And, was my result truly invalid?


                                                                ____________

                                                                robert.mouris
                                                                Send message
                                                                Joined: Nov 3 05
                                                                Posts: 129
                                                                Credit: 4,124,194
                                                                RAC: 0
                                                                Message 6239 - Posted 10 May 2007 20:43:31 UTC - in response to Message 6237.

                                                                  Last modified: 10 May 2007 20:44:28 UTC

                                                                  I hope to be able to come back soon! Attila and Ádám are aware of the problem.

                                                                  Robert
                                                                  ____________

                                                                  Stick
                                                                  Send message
                                                                  Joined: Jun 12 06
                                                                  Posts: 193
                                                                  Credit: 66,271
                                                                  RAC: 0
                                                                  Message 6240 - Posted 11 May 2007 0:01:27 UTC - in response to Message 6239.

                                                                    I hope to be able to come back soon! Attila and Ádám are aware of the problem.

                                                                    Robert


                                                                    Robert,

                                                                    I hope you are right about this! But, my interpretation is that they have only acknowledged the checkpoint problem - and not the (possible) validation problem. I very much look forward to your return!

                                                                    Stick
                                                                    ____________

                                                                    Profile jamin
                                                                    Avatar
                                                                    Send message
                                                                    Joined: Nov 20 06
                                                                    Posts: 324
                                                                    Credit: 100,110
                                                                    RAC: 0
                                                                    Message 6296 - Posted 21 May 2007 9:09:39 UTC

                                                                      I got another one. What\'s strange for me is that the lines processed by the 2 results are different - they were granted credits and mine not :(

                                                                      It\'s a pity - 87h down the drain :(
                                                                      ____________
                                                                      Bang! Zoom! Right in the kisser!

                                                                      ionu2002
                                                                      Send message
                                                                      Joined: Feb 14 06
                                                                      Posts: 1
                                                                      Credit: 8,145
                                                                      RAC: 0
                                                                      Message 6315 - Posted 28 May 2007 7:27:07 UTC - in response to Message 6237.

                                                                        Here is another one. My, uninterrupted, result got zero. The other two, which received credit, were both restarted on line 4. (This one restarted line 4 only. This one also restarted line 1 several times.)


                                                                        This WU was vaidated today. I noticed that my result, which processed all the way through without interruption, received zero credit. However, both of the results which did receive credit were restarted on the same line. This is another example of a possible validator flaw which I have pointed out several times - most recently in this message. That is, did the two results which received credit match because they encountered the checkpointing bug on the same line? Were these results truly valid? Or, just matching? And, was my result truly invalid?




                                                                        Me too, with this wu.
                                                                        Not interested so much in credits, more concerned about wasted time and resources.

                                                                        Ionut
                                                                        ____________

                                                                        Profile Ananas
                                                                        Send message
                                                                        Joined: Jul 12 05
                                                                        Posts: 222
                                                                        Credit: 665,833
                                                                        RAC: 0
                                                                        Message 6316 - Posted 29 May 2007 6:53:48 UTC

                                                                          Last modified: 29 May 2007 7:00:06 UTC

                                                                          Not only the wasted time is an issue, the result database might hold
                                                                          a bunch of wrong results now.

                                                                          Assuming that the uninterrupted results are correct and the interrupted
                                                                          ones are incorrect, each of those 0-credits uninterrupted ones most
                                                                          likely stands for one invalid entry (from the validated interrupted ones)
                                                                          that made it into the science database.

                                                                          p.s.: ... plus those WUs where 3 interrupted ones validated fine without
                                                                          beeing correct - and no one will ever notice it.

                                                                          Stick
                                                                          Send message
                                                                          Joined: Jun 12 06
                                                                          Posts: 193
                                                                          Credit: 66,271
                                                                          RAC: 0
                                                                          Message 6326 - Posted 3 Jun 2007 16:42:51 UTC

                                                                            Last modified: 3 Jun 2007 16:45:36 UTC

                                                                            Here is another WU that was (probably) validated incorrectly. My uninterrupted result got zero credit. The two results (this one and this one) which received credit were both restarted on line 5.
                                                                            ____________

                                                                            Stick
                                                                            Send message
                                                                            Joined: Jun 12 06
                                                                            Posts: 193
                                                                            Credit: 66,271
                                                                            RAC: 0
                                                                            Message 6327 - Posted 3 Jun 2007 17:07:40 UTC

                                                                              This WU is even stranger. My uninterrupted result was invalid and got zero credit. However, this result was also invalid (Output is empty) but received credit. And, this invalid result also received credit.
                                                                              ____________

                                                                              6dj72cn8
                                                                              Send message
                                                                              Joined: May 27 06
                                                                              Posts: 36
                                                                              Credit: 8,504
                                                                              RAC: 0
                                                                              Message 6331 - Posted 7 Jun 2007 4:40:45 UTC - in response to Message 6327.

                                                                                And another. Usual thing: invalid and 0 credit for an uninterrupted crunch but the two restarters are valid. Supposedly.
                                                                                ____________

                                                                                robert.mouris
                                                                                Send message
                                                                                Joined: Nov 3 05
                                                                                Posts: 129
                                                                                Credit: 4,124,194
                                                                                RAC: 0
                                                                                Message 6410 - Posted 14 Jul 2007 19:47:26 UTC - in response to Message 6316.

                                                                                  Not only the wasted time is an issue, the result database might hold
                                                                                  a bunch of wrong results now.

                                                                                  Assuming that the uninterrupted results are correct and the interrupted
                                                                                  ones are incorrect, each of those 0-credits uninterrupted ones most
                                                                                  likely stands for one invalid entry (from the validated interrupted ones)
                                                                                  that made it into the science database.

                                                                                  p.s.: ... plus those WUs where 3 interrupted ones validated fine without
                                                                                  beeing correct - and no one will ever notice it.

                                                                                  As it seems that nobody is bothered by supposedly wrong results in the database, neither the project administrators nor the scientists who base their theoretical research thereon, I suggest that the quorum be set to 1. The project will advance much faster than 3 times as fast, as it will also bring back those who don\'t want to wait many weeks or months before getting their credits.
                                                                                  ____________

                                                                                  Post to thread

                                                                                  Message boards : SZTAKI Desktop Grid : No Credit For A Clean Run?


                                                                                  Home | My Account | Message Boards


                                                                                  Copyright © 2017 SZTAKI Desktop Grid