Message boards :
Number crunching :
Unknown hosts?
Message board moderation
Author | Message |
---|---|
Send message Joined: 19 Aug 04 Posts: 39 Credit: 18,866 RAC: 0 |
Recently checked the \"Results\" section of my account and discovered two strange things. First - I have 7(!?!) WUs in my list. Well, some crashes and experiments with BOINC led to this situation, I confess, but how to remove them from the account list while the client shows only one and works only one? Second - those WUs from the account list: when I ckick on WU ID ALL of them are showing more than one host, two actually. One is mine and the other is uknown for me. Is it normal to asign experiments to two or more accounts for testing? 10x in advance <img src="http://boinc.mundayweb.com/cpdn/stats.php?userID=70"> |
Send message Joined: 17 Aug 04 Posts: 56 Credit: 63,814 RAC: 0 |
> First - I have 7(!?!) WUs in my list. Well, some crashes and experiments with > BOINC led to this situation, I confess, but how to remove them from the > account list while the client shows only one and works only one? You can't remove them. They were assigned to your computer. Whether or not you finished them makes no difference. They were assigned so they are associated with your computer. > Second - those WUs from the account list: when I ckick on WU ID ALL of them > are showing more than one host, two actually. One is mine and the other is > uknown for me. Is it normal to asign experiments to two or more accounts for > testing? Not just testing. seti@home and predictor@home actually send all work units out to 3 people. Then they verify the results and credit is only granted if all three results agree with each other. CPDN had to take a little different approach due to the LONG duration of their work units. However this long duration also increases the chances that there will either be an error during computation or that the user will become bored and quit before it is done. Hence sending the work units out to 2 people from the start greatly reduces the chance of a work unit just being dropped and then having to be re-sent out at a later date. Also I'm sure under ideal circumstances they wouldn't mind having 2 results per model for error checking. <br> ---------------------------- A member of The Knights Who Say Ni! Yet another stats page: <a href="http://boinc-kwsn.no-ip.info">http://boinc-kwsn.no-ip.info</a> |
Send message Joined: 10 Aug 04 Posts: 94 Credit: 309,849 RAC: 0 |
One, you need to merge your computer (host) from your users account page? Two, since your computers, i.e CPU and OS (computers) are hidden there there is no where to start helping you. Do you have hyperthreading as an example? Are you using Windows? You may need to go into your projects folder and remove the offending WU's? Example C:/Project Files/Boinc/Projects/hadsmXXXX This should only be done as a last resort. Had you reset or detached? Giving the necessary info is important to help fix a problem. Using the "Question & Problems" thread would be a better way to go. (forward slashes used instead of back slashes for html reasons) <img src="http://boinc.mundayweb.com/cpdn/stats.php?userID=35&trans=off"><a href="http://mysite.wanadoo-members.co.uk/thefinalfrontear/index.html"> Team Site Link</a> |
Send message Joined: 5 Aug 04 Posts: 1283 Credit: 15,824,334 RAC: 0 |
> Not just testing. seti@home and predictor@home actually send all work units > out to 3 people. Then they verify the results and credit is only granted if > all three results agree with each other. CPDN had to take a little different > approach due to the LONG duration of their work units. However this long > duration also increases the chances that there will either be an error during > computation or that the user will become bored and quit before it is done. > Hence sending the work units out to 2 people from the start greatly reduces > the chance of a work unit just being dropped and then having to be re-sent out > at a later date. Also I'm sure under ideal circumstances they wouldn't mind > having 2 results per model for error checking. All 7 of the units that Fram has been allocated had been recirculated (10 or 11 days) after the first host completed the job with a client error. <a href="http://climateapps2.oucs.ox.ac.uk/cpdnboinc/team_display.php?teamid=3"><img src="http://www.teampicard.net/templates/fisubice/images/phpbb2_logo.jpg"></a> |
Send message Joined: 17 Aug 04 Posts: 56 Credit: 63,814 RAC: 0 |
> All 7 of the units that Fram has been allocated had been recirculated (10 or > 11 days) after the first host completed the job with a client error. Oops. Guess I was wrong. I have had several work units error out myself and all of them had already been assigned to another host so I assumed all work units were just sent out to 2 hosts immediately. But looking at them again and comparing the sent/recieved dates, it is clear that you are right, Thyme. That'll teach me to assume things! 'cause you know what happens when you assume things, right? :) Ah... sense of humor is decaying. Time for bed. <br> ---------------------------- A member of The Knights Who Say Ni! Yet another stats page: <a href="http://boinc-kwsn.no-ip.info">http://boinc-kwsn.no-ip.info</a> |
Send message Joined: 5 Aug 04 Posts: 1283 Credit: 15,824,334 RAC: 0 |
> Two, since your computers, i.e CPU and OS (computers) are hidden there there > is no where to start helping you, Do you have hyperthreading as an example? > Are you using Windows? You may need to go into your projects folder and > remove the offending WU's. Example C:/Project Files/Boinc/Projects/hadsmXXXX You can find out, but its not obvious, and I'm not sure if it's really intended ;-) If you click on a trickle on the account data page and then on the host ID on the results page you can see the computer details (with owner listed as anonymous). You can see all units sent to that computer by clicking on the results link. <a href="http://climateapps2.oucs.ox.ac.uk/cpdnboinc/team_display.php?teamid=3"><img src="http://www.teampicard.net/templates/fisubice/images/phpbb2_logo.jpg"></a> |
Send message Joined: 19 Aug 04 Posts: 39 Credit: 18,866 RAC: 0 |
Toby, thank you for the explanation. :) Dogbytes, Im sorry that my computers are hidden - its because of the LAN I belong. Im using WInXP SP1, ADM Barton. In my projects folder the only one WU is that actually calculated. The others dont exist on my hdd - only in the list in my account. In fact that isnt a real problem for me. I meant that it might be a problem for CPDN :) <img src="http://boinc.mundayweb.com/cpdn/stats.php?userID=70"> |
Send message Joined: 19 Aug 04 Posts: 39 Credit: 18,866 RAC: 0 |
Thyme Lawn, thank you too :) <img src="http://boinc.mundayweb.com/cpdn/stats.php?userID=70"> |
Send message Joined: 10 Aug 04 Posts: 94 Credit: 309,849 RAC: 0 |
Problem solved. Dobra! |
Send message Joined: 19 Aug 04 Posts: 39 Credit: 18,866 RAC: 0 |
> All 7 of the units that Fram has been allocated had been recirculated (10 or > 11 days) after the first host completed the job with a client error. > Does it mean that the WU Im doing right now is obsolete - the most recent 7th one? <img src="http://boinc.mundayweb.com/cpdn/stats.php?userID=70"> |
Send message Joined: 19 Aug 04 Posts: 39 Credit: 18,866 RAC: 0 |
> Problem solved. Dobra! > LOL Sounds almost like bulgarian "Dobre" :):):) Thanks, buddy :) :) <img src="http://boinc.mundayweb.com/cpdn/stats.php?userID=70"> |
Send message Joined: 5 Aug 04 Posts: 1283 Credit: 15,824,334 RAC: 0 |
> > All 7 of the units that Fram has been allocated had been recirculated (10 > or > > 11 days) after the first host completed the job with a client error. > > > Does it mean that the WU Im doing right now is obsolete - the most recent 7th > one? No. One of your WUs is over with a client error, the others are still active. The one that's trickling will be fine, the others might eventually be tagged as over and get recirculated again. <a href="http://climateapps2.oucs.ox.ac.uk/cpdnboinc/team_display.php?teamid=3"><img src="http://www.teampicard.net/templates/fisubice/images/phpbb2_logo.jpg"></a> |
Send message Joined: 19 Aug 04 Posts: 39 Credit: 18,866 RAC: 0 |
> No. One of your WUs is over with a client error, the others are still active. > The one that's trickling will be fine, the others might eventually be tagged > as over and get recirculated again. > Thank you once again, Thyme Lawn :) <img src="http://boinc.mundayweb.com/cpdn/stats.php?userID=70"> |
Send message Joined: 5 Aug 04 Posts: 1283 Credit: 15,824,334 RAC: 0 |
> Thank you once again, Thyme Lawn :) My pleasure, Fram. And if you're looking to join a team you could always click on the image below ;-) <a href="http://climateapps2.oucs.ox.ac.uk/cpdnboinc/team_display.php?teamid=3"><img src="http://www.teampicard.net/templates/fisubice/images/phpbb2_logo.jpg"></a> |
Send message Joined: 19 Aug 04 Posts: 39 Credit: 18,866 RAC: 0 |
> > Thank you once again, Thyme Lawn :) > > My pleasure, Fram. And if you're looking to join a team you could always > click on the image below ;-) > Thanks for the invitation :) <img src="http://boinc.mundayweb.com/cpdn/stats.php?userID=70"> |
©2024 cpdn.org