EMIII gone all haywire in 2 puters.

dragonV8dragonV8 not here much New
edited October 2004 in Folding@Home
Sally updated 2 computers running 4 instances of folding to the latest version of EMIII and also changed from V4 to V5.02.

Each computer has a console and a GUI. The problems are with the consoles only.

EMIII shows "Time per Frame" and "Time per WU" as 00:00:00. When she brings up the log, it shows as still folding ok. :rolleyes:

Just curious if anyone has come across this problem before. It's a pain as we keep checking to ensure it is still working.

Though we have lost a few WU's in the last 24-36 hours, i don't think it's related. Sally did try to explain it to me, but i must be a bit thick as i could not grasp it real well, lol. :scratch: (I'm ok with heavy weights though). :)

Comments

  • GHoosdumGHoosdum Icrontian
    edited October 2004
    How long has EMIII been running since the update? Sometimes it takes a while to update the timing info, once it's worked its way into a few frames and been running for a bit.
  • dragonV8dragonV8 not here much New
    edited October 2004
    Wish it was that problem GHoosdum. This one is different. It's like that for the whole time the WU is folding from start to finish. :(

    Even now checking the WU it still shows 00:00:00 in EMIII, yet shows quite a few steps done in the actual log. :rolleyes:

    At least it is still working (folding). :)
  • mmonninmmonnin Centreville, VA
    edited October 2004
    How about getting the latest version of EM3. V3.00 just came out.
  • dragonV8dragonV8 not here much New
    edited October 2004
    Sally has been doing the new EMIII V3 since she had notification this morning. Don't know if it will/has made a difference yet. Don't want to ask her as she is concentrating on the job at hand and we are expecting the electrician any minute.

    She is trying to do as much as possible before he gets there. Then it's off with all the puters.....temporary loss of points, :( but in the long run should stop the farm from intermittant tripping. :)
  • CryptoCrypto W.Sussex UK Member
    edited October 2004
    Jon,

    sometimes I have a problem with EMIII pointing to the right file location for the Log, have you checked that?

    When you check to see if the WU is running, do you look via EMIII or do you right click on the F@H icon in the system tray and look at configure/view log file?

    Good luck with the farm re-wiring :)

    Cheers

    Crypto
    :smokin:
  • Straight_ManStraight_Man Geeky, in my own way Naples, FL Icrontian
    edited October 2004
    Um, do the computers that have this issue have clients with the -advmethod switch on??? If so, you might have something interesting going on.... That switch, lately, has been getting me both a new core, Core_82 which is a core that processes WUs that are type coded Amber, and Amber is a whole new WU type. AFAIK, EMIII does not know where Amber saves its data partials and HAS NO STEP AND FRAME DATA for Amber WUs yet.

    Second, EMIII uses WU tables to figure out how many frames there are in aWU, and my guess is that it has only established WUs in it. You are probably getting WUs it does not know.

    Some data to tell from logs if you are getting Amber work:

    First, Core shown will be Core_82.

    Second, it will show as a PMD core, version 1.0 as core starts.

    Third, you will get a check point (Completion Point) file written every percentage after the first percentage completes for every percentage point completed AFTER the first percentage completes when the folding is started. This will happen irregardless of percentage calc time, as the amber core figures out for itself when to save CP files. I had one WU here saving a Completion point file every 9 minutes on my XP SP2 box. The save painted in the log during every percentage except one-- the time between the client start and the completion of percentage 1 (frame 1 complete), shown as (1%) in log. It writes NO data to .nfo files, adn I noted byut forgot the other type that Gromacs and Tinkers use but which were ZERO byte size after each of four Amber Wus were worked on herr. Completion Point (.CP file type) is used a lot.

    AFAIK, EMIII has not been coded for Amber WUs yet, nor the files it uses to store things which differ some from how Gromacs and Double Gromacs and Tinker partials are stored. I will let this double as a new WU TYPE rollout notify that has just begun in last two-four days. I got my first Amber WU and a Core_82 about 3.7 days ago, and I do not accept beta WUs here and use same clients as you folks who do not beta fold do. They showed up in the Project Summary page October 11 or 12 US calendar days. So, I think it is safe to say that Amber is now officially being assigned as something other than early Beta, some Ambers are being publicly assigned to normal cores that use the -advmethods switch, and I can talk about it now in summary form.

    John.
  • csimoncsimon Acadiana Icrontian
    edited October 2004
    Hey Jon ...does your emIII look similar to this?
    This machine seems to not be running emIII correctly but the one at work is ...let me know if you get it figured out and how you did it!

    thanks
    csimon :fold:

    maybe this has something to do with it "The old protein data file will no longer be supported or updated. The old file has been removed from the site and the new one has been put up in its place. If this new method works out, I may be able to make EM update automatically and the days of posting new protein data files may be over... I hope. Only time will tell."
  • Straight_ManStraight_Man Geeky, in my own way Naples, FL Icrontian
    edited October 2004
    Not running EMIII right now, but folks on Folding Community were complaining about EMIII and some had Amber WUs. Amber Core is enough different that EMIII cannot understand it right yet. In fact, tried it on one Amber, it did similar things to that except it had no idea of the WU number either. Yes, the changeover of update downloading you quoted could play into this also, the data file is a WU ID table or Table set in a file.
  • dragonV8dragonV8 not here much New
    edited October 2004
    Chris, yours is more condensed than ours, though the similarities are there.

    Time per frame kept resetting itself. It appears EMIII was getting info from the latest WU and also picking/keeping info from the previous one. In the top left corner it would show: P****(pick a number) and the expanded one next to "Time per WU" would be totally different. :(

    Sally just got up to check the puters and said the problem ones now (finally) sorted themselves, with no help from us. Been keeping an eye on any info available @ Stanford too. I've got a case of "Data Overload", lol. :rolleyes:

    Here's to hoping they stay that way.

    ps...the "Amber" core shows up in "Pink" in EMIII, at least with us it does. :scratch:

    Thanks for all the input guys. :thumbsup:
  • csimoncsimon Acadiana Icrontian
    edited October 2004
    dragonV8 wrote:
    Chris, yours is more condensed than ours, though the similarities are there.

    Yes that is how I have mine set ...to read 4 boxes although I only have one. I expand whenever I wanna see the info but mostly I use what you see there.

    Mine still hasn't sorted itself out it just shows 0000000's
  • dragonV8dragonV8 not here much New
    edited October 2004
    For the record, from memory ours started showing 0000000's 5 days ago and only fixed it self in the early hours of this morning. Just wish i knew how. :rolleyes:
  • csimoncsimon Acadiana Icrontian
    edited October 2004
    mine fixed itself in the wee hours too! :fold:
    maybe it was a new protien that did it cause the protien data hasn't been updated since v3 arrived.
  • dragonV8dragonV8 not here much New
    edited October 2004
    Sally noticed about 15 minutes ago that Larry has another update to V3. It is the latest protien data file.

    This may also have a bearing on the problems we were having, Chris.
    Glad to see yours came good as well. :thumbsup:
Sign In or Register to comment.