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

NFS@home challenge

Discussion in 'Team OcUK Distributed Computing Projects' started by hanluc, Dec 10, 2012.

  1. Carlos Pinho

    Associate

    Joined: Dec 12, 2012

    Posts: 73

    Location: Milton Keynes , UK

    Don't worry, there's plenty of work available. As I said before more than 1M wu's.
     
  2. PDW

    Wise Guy

    Joined: Feb 21, 2014

    Posts: 1,460

    I've put a few cores onto this but can't do too many due to the memory requirements. Can I ask why you are asking for help, I couldn't see anything on your forum about it ?
     
  3. Carlos Pinho

    Associate

    Joined: Dec 12, 2012

    Posts: 73

    Location: Milton Keynes , UK

    Asking for help to finish these two numbers as quick as possible so we can start working on harder numbers to justify next year grant of the cluster used for the post-processing stage.
    When I can I do like to go to each individual forum and request help. It is a way to approach people to NFS@Home project.

    Carlos
     
  4. andy_taximan

    Wise Guy

    Joined: Apr 20, 2014

    Posts: 1,675

    Location: Herts/uk

    Just got 16e now on main pc , had to turn down cpu usage to 60% to stop the waiting for memory messages
     
  5. colinmcr

    Wise Guy

    Joined: Mar 17, 2009

    Posts: 2,465

    Location: Scotland

    Well if I ever get my 2P rig finished I'll try do some ;)
     
  6. PDW

    Wise Guy

    Joined: Feb 21, 2014

    Posts: 1,460

    Okay, thanks for the info.

    Does it run better under Window or Linux ?
     
  7. Carlos Pinho

    Associate

    Joined: Dec 12, 2012

    Posts: 73

    Location: Milton Keynes , UK

    "Runs better", do you want me to comment about speed?

    Last info I had was identical speed when they were compiling the windows 64-bit version of lasieve5f and new linux version.
     
  8. Carlos Pinho

    Associate

    Joined: Dec 12, 2012

    Posts: 73

    Location: Milton Keynes , UK

    Have the same issue on my work laptop, running 3 threads out of 8. Only 8GB of memory.

    In the meantime:

    2,1285- sieve has started. We will go up to q=2200M or 2400M, meaning we will have 1.09M wus to 1.19M wus to process. Can we sieve this integer in time record?!

    For 6,490+ we stopped at q=1800M. Still awaiting to complete the left overs, ETA= 1 week.

    Best Regards,

    Carlos
     
  9. Carlos Pinho

    Associate

    Joined: Dec 12, 2012

    Posts: 73

    Location: Milton Keynes , UK

    Right now we have 21505 outstanding wus to complete the sieve of 6,490+ to q=1800M.
     
  10. Carlos Pinho

    Associate

    Joined: Dec 12, 2012

    Posts: 73

    Location: Milton Keynes , UK

    andy_taximan,

    Not sure if you notice but at this stage G2m1285_ wus only consume 430Mb therefore you can turn up cpu usage to 100%.

    Carlos
     
  11. andy_taximan

    Wise Guy

    Joined: Apr 20, 2014

    Posts: 1,675

    Location: Herts/uk

    tried 100% and got "waiting for memory" so backed it down to 80% , so got 12 threads running ok

    Also got some G6p490_ wus what do they need?
     
    Last edited: Jul 20, 2015
  12. hanluc

    Mobster

    Joined: Oct 22, 2010

    Posts: 2,744

    Location: Ratae Corieltauvorum

    Andy have you changed the default boinc client memory usage? I think the default is set at only 50% when in use. It's under tools/computer preferences/disk and memory usage.
     
  13. andy_taximan

    Wise Guy

    Joined: Apr 20, 2014

    Posts: 1,675

    Location: Herts/uk

    Ah well spotted lol yer was at 50%
     
  14. Carlos Pinho

    Associate

    Joined: Dec 12, 2012

    Posts: 73

    Location: Milton Keynes , UK

    The G6p490 ones are the left overs that people didn't complete or aborted. They all end at _1 or _2, etc...
    I have like 10 in queue.

    Carlos
     
  15. colinmcr

    Wise Guy

    Joined: Mar 17, 2009

    Posts: 2,465

    Location: Scotland

    Ok I'm gonna try run some on my laptop here in India just for fun. Managed to get 6/8 running then read about default memory and managed to get all 8 going buy upping mem to 80%
     
    Last edited: Jul 20, 2015
  16. Carlos Pinho

    Associate

    Joined: Dec 12, 2012

    Posts: 73

    Location: Milton Keynes , UK

    At current pace 2,1285- will be completely sieved within 100 days.
     
  17. PDW

    Wise Guy

    Joined: Feb 21, 2014

    Posts: 1,460

    Thanks, that was one of the questions I had been meaning to ask :)

    The other question I have is when is your deadline for getting them done by so you can start on 'harder' numbers ?
     
  18. andy_taximan

    Wise Guy

    Joined: Apr 20, 2014

    Posts: 1,675

    Location: Herts/uk

    Woo , just taken 1st place in Top Hosts !! Greg (former number 1) is running 4 Tesla gpu 64gig ram pc! lol

    Seems Greg is the Project administrator lol

    We are now 40th place overall
     
    Last edited: Jul 28, 2015
  19. Carlos Pinho

    Associate

    Joined: Dec 12, 2012

    Posts: 73

    Location: Milton Keynes , UK

    Sorry for the late reply. Some comments from Greg the Admin.

    1) Server issues: "I think I tracked down the database slowdown. Turns out that the size of the database, specifically the host table, has grown substantially thanks to Syracuse’s 1.8 million host entries. It became larger than the cache assigned to it, so records were being read from disk rather than memory. That’s bad. I increased the max cache size to much larger than the table, so things should be back to normal now."

    2) Records: "As far as setting a new record, the current lasieve16e binaries aren’t up to the task. They came close to running out of steam on our current records."
    Also " I need to take a look at the full data set I have now and set priorities for the next grant proposal"

    So we are trying to check from the Cunningham Tables and on other projects where we can be of assistant inside the limits of the sievers, closer to record pinch.

    Thank you all for your support.
    I'll let you know what will be the next integer to sieve.

    Keep crunching, we are still far away to finish 2,1285-, more or less 850k wus.

    Carlos
     
  20. PDW

    Wise Guy

    Joined: Feb 21, 2014

    Posts: 1,460

    Thanks for the update Carlos.

    Rough guess at less than 10 weeks to complete 2,1285- at current rates then !