![]() |
![]() |
#1 |
Nov 2015
916 Posts |
![]()
Howdy!
I work in HPC and I would really love to donate some extra cycles from our systems, but the setup for it seems a bit... involved. I'd rather not manually run the setup for each and every one of our ~4000 hosts to make sure it all reports correctly. Is there anyone who has successfully set up GIMPS as a PBS job who might be able to help me out? I have a couple of ideas on how to do it automatically (largely through abuse of expect commands), but I wanted to know if there was a more elegant option. Thanks for your time! |
![]() |
![]() |
![]() |
#2 |
Aug 2002
100001011011012 Posts |
![]()
Please make sure you have permission first! (The electrical and cooling cost increase will not be trivial!)
![]() |
![]() |
![]() |
![]() |
#3 |
Nov 2015
32 Posts |
![]()
I do indeed! I already do something similar for folding@home.
In fact, it's generally appreciated when we submit a bunch of super-low-priority jobs, because it bumps up our cluster utilization percentages and gets our CIO more interested in putting down money for another cluster. |
![]() |
![]() |
![]() |
#4 | |
"Kieren"
Jul 2011
In My Own Galaxy!
100111101011102 Posts |
![]() Quote:
![]() |
|
![]() |
![]() |
![]() |
#5 |
Nov 2015
32 Posts |
![]()
I know it's dumb, but boy, if the difference between say 83% utilization and 92% utilization is another 200 nodes to next year's cluster and the extra utilization is actually going to a good cause, I don't lose any sleep
![]() But I take it that there's nobody who has done it in this format? |
![]() |
![]() |
![]() |
#6 |
If I May
"Chris Halsall"
Sep 2002
Barbados
101011010011112 Posts |
![]() |
![]() |
![]() |
![]() |
#7 | |
Nov 2015
32 Posts |
![]() Quote:
Are you saying that I'm making myself a target, are you saying that others wouldn't talk because they don't want to become a target, are you saying that anyone who would have done this would have talked about it before, or did you mean something entirely else? |
|
![]() |
![]() |
![]() |
#8 | |
Jan 2014
2·73 Posts |
![]() Quote:
|
|
![]() |
![]() |
![]() |
#9 | |
Nov 2015
916 Posts |
![]() Quote:
The majority of our systems are RHEL, but we do have a wide variety of different flavors of linux for special use cases. |
|
![]() |
![]() |
![]() |
#10 |
Jan 2014
2228 Posts |
![]()
See this thread for init file and logging on linux flavours. I assume you have something like cfengine or puppet to distribute software and configs. prime.txt and local.txt are host-specific, but I do not know what would a minimum set-up here, or how to create a batch of IDs beforehand. Also, you would need to insert some system info (CPU freq, memory) there as well. Mprime generates some of this info on first start.
Last fiddled with by blip on 2015-12-01 at 22:54 |
![]() |
![]() |
![]() |
#11 | |
"Kieren"
Jul 2011
In My Own Galaxy!
2×3×1,693 Posts |
![]() Quote:
![]() |
|
![]() |
![]() |
![]() |
Thread Tools | |
![]() |
||||
Thread | Thread Starter | Forum | Replies | Last Post |
Auto-XYYXF tool | fivemack | NFS@Home | 3 | 2016-07-06 04:01 |
auto off/on machines | wildrabbitt | Hardware | 3 | 2016-05-14 16:31 |
Auto scroll and auto maximize worker windows ? | wyattwong | Software | 3 | 2011-02-25 20:20 |
Spammer auto-ban system? | 10metreh | Forum Feedback | 6 | 2009-08-24 17:35 |
Auto updates from mprime | rgiltrap | Linux | 1 | 2007-01-08 22:33 |