Results 1 to 5 of 5

Thread: Overloading RAM on new work units

  1. #1
    Join Date
    Nov 2005
    Location
    Central Pennsylvania
    Posts
    4,333

    Overloading RAM on new work units

    I downloaded some work units only to have the system fall back to smaller number of cores due to exceeding the 50% realm of usable RAM from my machine. Never have I had a problem except in Atlas@home and Lattice@home use more then 50% which in my case is my system has 32 Gig Ram available. So 16 Gig of RAM is pretty rough and tough system use. I adjusted the usability of % to 90% and I still got a shut down of usability to only 4 work units, those are serious work units. I also wanted to let anyone know that each work unit has nearly 100 meg per work unit being downloaded for each work unit. So if your paying for bandwidth and data used, watch out.! I reset my Lattice @Home project and left it lie there whimpering as it won't be getting me to help anytime soon.

    Name 222437370.04178208515455539.517_6
    Workunit 5162896
    Created 25 Mar 2017, 21:03:43 UTC
    Sent 26 Mar 2017, 15:58:38 UTC
    Report deadline 6 Apr 2017, 15:58:38 UTC
    Received 26 Mar 2017, 17:22:14 UTC
    Server state Over
    Outcome Computation error
    Client state Compute error
    Exit status 1 (0x1) Unknown error number
    Computer ID 111085
    Run time 23 sec
    CPU time 19 sec
    Validate state Invalid
    Credit 0.00
    Device peak FLOPS 3.26 GFLOPS
    Application version GARLI v5.06
    Peak working set size 745.65 MB
    Peak swap size 743.34 MB
    Peak disk usage 97.16 MB

    Beware





    Challenge me, or correct me, but don't ask me to die quietly.

    …Pursuit is always hard, capturing is really not the focus, it’s the hunt ...

  2. #2
    Join Date
    Sep 2010
    Location
    Leiden, the Netherlands
    Posts
    4,372
    Quote Originally Posted by Nflight View Post
    I reset my Lattice @Home project and left it lie there whimpering as it won't be getting me to help anytime soon.
    Just Lettuce@Home then (Man, was I glad to reach 100k with this project!)


  3. #3
    NeoGen's Avatar
    NeoGen is offline AMD Users Alchemist Moderator
    Site Admin
    Join Date
    Oct 2003
    Location
    North Little Rock, AR (USA)
    Posts
    8,451
    I've struggled with those too for a bit, it's just too much memory. I'm going to try and see if an app_config would work to limit how many it can run for Lattice only, and leave the rest of the cores for other projects.

    <app_config>
    <app>
    <name>garli</name>
    <max_concurrent>2</max_concurrent>
    </app>
    </app_config>
    EDIT: After about 15 mins monitoring, it seems to be working, only 2 Lattice WUs are running and the other cores are being taken by Leiden Classical. Task Manager showing 12.6GB of RAM usage and slowly climbing.
    Last edited by NeoGen; 03-28-2017 at 05:10 AM.

  4. #4
    Join Date
    Sep 2010
    Location
    Leiden, the Netherlands
    Posts
    4,372
    My advice: do not run more than one at a time.


  5. #5
    NeoGen's Avatar
    NeoGen is offline AMD Users Alchemist Moderator
    Site Admin
    Join Date
    Oct 2003
    Location
    North Little Rock, AR (USA)
    Posts
    8,451
    After almost 24 hours seems to have stabilized at 20Gb of RAM usage in Windows Task Manager for 2 Lattice workunits. I'm guessing between 8 and 9Gb per WU plus a couple of Gb for the OS and applications.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •