1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

enigma@home - Boinc stats team challenge

Discussion in 'Team OcUK Distributed Computing Projects' started by Carlos Pinho, Feb 3, 2016.

  1. PDW

    Wise Guy

    Joined: Feb 21, 2014

    Posts: 1,763

    Only about 25% is actually being used (mostly by windows) of the 16Gb.

    Have found that it is also getting:

    02/01/2017 06:44:24 | Enigma@Home | Reporting 3 completed tasks
    02/01/2017 06:44:24 | Enigma@Home | Requesting new tasks for CPU
    02/01/2017 06:44:27 | Enigma@Home | Scheduler request completed: got 0 new tasks
    02/01/2017 06:44:27 | Enigma@Home | No tasks sent
    02/01/2017 06:44:27 | Enigma@Home | Message from server: Enigma 0.76b needs 13482166625837707296508455294996900044251027422596401029782012726271763384636952730618783393505932024335246350758056564557460147738613726978611546151458668996605602112795106310885474304.00 MB RAM but only 16268.52 MB is available for use.
    02/01/2017 06:44:27 | Enigma@Home | Message from server: Enigma 0.76b needs 3633355906759967822163392472118519509897936333762380538799349237146650150447872889552536326477638749902077901780791787612123406287017170888293484786718292115578013593460801536.00 MB RAM but only 16268.52 MB is available for use.
    02/01/2017 06:48:30 | Enigma@Home | Sending scheduler request: To fetch work.
    02/01/2017 06:48:30 | Enigma@Home | Requesting new tasks for CPU
    02/01/2017 06:48:34 | Enigma@Home | [error] Can't parse file info in scheduler reply: unexpected XML tag or syntax
    02/01/2017 06:48:34 | Enigma@Home | [error] No close tag in scheduler reply


    The XML error occurred after at least 5 hours of the memory message and getting 0 new tasks (and no problems with tags).
    Now if I update it gives the error and backups off even more, suspect others will be getting it, they are getting the memory error.
    Hopefully it will eventually sort itself out (WCG had a similar problem last year), but that box will be going back to Gene as well when it finishes the last few in the cache to be done :rolleyes:
     
  2. PDW

    Wise Guy

    Joined: Feb 21, 2014

    Posts: 1,763

    After reading 7 pages of the WCG thread found that setting the cache to 0 days, set no new tasks, then update would report/remove the completed work.

    Then allowed new work and am getting 'Resent lost task' messages for each new task that comes down when more work is needed. Setting cache to anything but 0 results in the return of the XML tag error.

    Am hoping once all the 'lost' tasks (about 112) have been downloaded, completed and returned things will get back to normal and I can increase the cache again.
     
  3. PDW

    Wise Guy

    Joined: Feb 21, 2014

    Posts: 1,763

    Was going to post that we had just hit 10th place but hitting refresh again and we have gone up to 9th :)
     
  4. PDW

    Wise Guy

    Joined: Feb 21, 2014

    Posts: 1,763

    That fix didn't last long :(

    Came back to check progress, it had got down to 95 left to do but nothing being downloaded so was idling.
    Hadn't changed anything, tried a few things and gave up on it. That one is back on Gene now !

    PS. Up to 7th now :)
     
  5. hanluc

    Mobster

    Joined: Oct 22, 2010

    Posts: 2,958

    Location: Ratae Corieltauvorum

    I can't help you, other than to suggest asking in their forums.
    I have the big rig on Numberfields atm, I'll finish off what I'm doing and switch to enigma later on.
     
  6. scole250

    Associate

    Joined: Jun 27, 2014

    Posts: 28

    Post the contents of your enigma app_config.xml and app_info.xml if you have one. Sometimes the broken XML structure in the global_prefs.xml in the WCG section causes a problem. You can either fix it or delete the file. The fix is to look for the <venue> tag without the closing </venue> tag at the end.

    As for the memory message, try restarting the the boinc client. I've had that error a couple times but restarted the client and no more problem.
     
  7. PDW

    Wise Guy

    Joined: Feb 21, 2014

    Posts: 1,763

    The app_config just restricted the number of threads so the PPS Sieve on the GPU and a couple of WOO WUs could keep running.
    The app_info was the one from the Beta3 link Carlos gave that I copied onto other Window boxes. They are all happily deciphering the code without a problem.

    Some of the other boxes (I haven't looked at all) are also getting the memory error message but still chugging away.

    Thanks for the reminder of the <venue> tag, I have seen that cause problems over at Gerasim and had to fix it. Just seems strange that having got rid of the error by setting the cache to 0 it thens makes a come back :confused:
     
  8. PDW

    Wise Guy

    Joined: Feb 21, 2014

    Posts: 1,763

    Up to 4th place now having overtaken P3D :)
    Will they wake up and claim a position back like they did last Feb ?

    Can't see us getting to 3rd.
    The fun is now watching Seti.USA and TSBT who are fighting amongst themselves for the top spot :D
     
  9. Coleslaw

    Associate

    Joined: Apr 16, 2014

    Posts: 78

    Did you try a project reset? And does any other projects have this issue currently? (I'm guessing you just have the one pulling work but gotta ask.)

    Edit: Oh..and I had the same issue with Gerasim. Sad part was, it was only on a random PC here and there. Setting global preferences at another project fixed my issue there at that time, but like I said in their forums, I find it odd that Gerasim is the only project I have ever seen have the issue with WCG's global preferences. I wasn't even using the global preferences is the odd part.
     
    Last edited: Jan 3, 2017
  10. PDW

    Wise Guy

    Joined: Feb 21, 2014

    Posts: 1,763

    Are you talking about this XML issue ?

    02/01/2017 06:48:34 | Enigma@Home | [error] Can't parse file info in scheduler reply: unexpected XML tag or syntax
    02/01/2017 06:48:34 | Enigma@Home | [error] No close tag in scheduler reply


    After I had been able to return all the completed work and get nothing new to do then yes, I tried a project reset. It made no difference.
    The 120+ tasks on the server (in my account) remained.

    Some time after that I found the 'setting cache to 0' suggestion which made the lost tasks download to the client to be worked on. 8 where on the PC, plus 112 left to download and process. When it finally gave up there were 95 still be 'processed' and 0 on the PC.

    Other projects happily carried on pulling work down without a problem.

    I tend to think it is probably related to the Enigma server being overloaded at the start of the challenge with all those who had bunkered returning their work, some communications didn't go according to plan and it messed up. No idea what version of server code they are running and whether that might affect things.


    With regards to Gerasim, there is a bit of a language problem but when I found a bug in their implementation I gave them enough evidence (screen dumps, the process I followed) that they identified the problem and fixed it very quickly. I also tried to pin the XML tag problem down to them but came to the conclusion it wasn't their project at fault from the evidence I could gather. If you want to find a project that really messes things up then try Albert !

    I never use the Global Preferences, not knowingly anyway, I always want to set things locally how I want them.
    I really want to re-write the BOINC client how I would like to use it but I know I'm most unlikely to get round to doing that :(
     
  11. Coleslaw

    Associate

    Joined: Apr 16, 2014

    Posts: 78

    Yeah... I ran Albert to the million point goal and then will every now and again attach it to a borg or low end gpu system.

    Gerasim has a huge language barrier and it doesn't have to be that bad. Sucks to have to bumble through the settings just to get to English.

    Server overload could very well be the enigma issue. Hope all is well going forward. The project has pretty well been on auto pilot for years and nothing really gets updated. So, old server software is most likely used there.
     
  12. PDW

    Wise Guy

    Joined: Feb 21, 2014

    Posts: 1,763

    Someone else reported the XML tag error over at this Primegrid thread yesterday...

    03/01/2017 19:57:56 | PrimeGrid | [error] Can't parse workunit in scheduler reply: unexpected XML tag or syntax
    03/01/2017 19:57:56 | PrimeGrid | [error] No close tag in scheduler reply
     
  13. hanluc

    Mobster

    Joined: Oct 22, 2010

    Posts: 2,958

    Location: Ratae Corieltauvorum

    Nice tussle for 7th in the personal stats. I've lost count how many times myself and PDW have switched places.
     
  14. Bowmore

    Associate

    Joined: Mar 6, 2016

    Posts: 20

    When you've been running WCG and switch over to other BOINC project and run into that "tag" error then there is usually a very simple fix. Go to WuProp and change something in the global preferences, save them, and then update the offending computer on WuProp. I usually just toggle the Run on Battery field since that is irrelevant on my machines.
     
  15. Bowmore

    Associate

    Joined: Mar 6, 2016

    Posts: 20

    If you guys continue to pound Enigma I foresee you taking the number 3 position from GC :) Of course my crystal ball isn't always accurate!
     
    Last edited: Jan 5, 2017
  16. hanluc

    Mobster

    Joined: Oct 22, 2010

    Posts: 2,958

    Location: Ratae Corieltauvorum

    Welcome new Team OcUK members Bowmore & Scole! lol
    Thanks guys.
     
  17. Bowmore

    Associate

    Joined: Mar 6, 2016

    Posts: 20

    You're welcome :D

    Happy New Year OcUK ... now let's go forth and kick some GC backside ;) Hopefully we have enough to pull it off.
     
  18. PDW

    Wise Guy

    Joined: Feb 21, 2014

    Posts: 1,763

    We'd been swapping places for about 10 hours since lunchtime then I assume you looked down the back of the sofa and found some more cores ?

    My phone defected to the dark side and decided it was better to run Rosetta Alpha work :(

    The last two WOO Wus completed late yesterday afternoon so that gave me 2 more cores and the opportunity to restart Boinc and run the optimised app again, I thought the extra 7 cores would be enough to overtake and stay there :rolleyes:

    The SoBs had also completed Wednesday night so that was another 2 cores, but I didn't suspend the Cullen and 321 work in case the cache of PPS Sieve for the GPUs ran out. I've been swapping 1st place in Primegrid on FB with PNT for a few days so didn't want to spoil that fun either :)

    Congratulations, I'll let you have my lunch money later :D
     
  19. PDW

    Wise Guy

    Joined: Feb 21, 2014

    Posts: 1,763

    Happy New Year to [insert Team_Name here] :D

    Was going to suggest that you should go to the Drs and get your eyesight and crystal checked !


    PS. Hanluc, good work on the new team recruitment drive, just need to keep pushing on the team retention side of things :D
     
  20. Bowmore

    Associate

    Joined: Mar 6, 2016

    Posts: 20

    The Doc said my eyesight was okay and he polished the crystal ball for me lol

    The Borg have been disassemilated :D All that remains now is to hold them for 9 more hours. It appears Enigma is having some trouble with their validator :(