View previous topic :: View next topic |
Select GPU tie-breaker |
Collatz |
|
36% |
[ 4 ] |
Poem |
|
63% |
[ 7 ] |
|
Total Votes : 11 |
|
Author |
Message |
LanDroid Prince


Joined: 11 Jun 2002 Posts: 4476 Location: Cincinnati, OH U.S.
|
Posted: Fri Aug 30, 2013 11:23 am Post subject: GPU Tie-Breaker VOTE! / Poem Vs. Collatz |
|
|
VOTE! |
|
Back to top |
|
 |
stooper101 Prince


Joined: 16 May 2009 Posts: 1248 Location: Cheese Country
|
Posted: Fri Aug 30, 2013 12:03 pm Post subject: |
|
|
How has POEM been with work supply? _________________
>stats< |
|
Back to top |
|
 |
LanDroid Prince


Joined: 11 Jun 2002 Posts: 4476 Location: Cincinnati, OH U.S.
|
Posted: Fri Aug 30, 2013 1:54 pm Post subject: |
|
|
Poem GPU WU's are not very steady* in my experience, but if that wins we can use Collatz as the back up project (similar to what is being done with Correlizer & Asteroids in the CPU arena).
*I think there's a trick to it, something along the lines of getting up at 4:00 a.m. EST to grab what's available? |
|
Back to top |
|
 |
Gemjunkie Prince


Joined: 03 Jul 2010 Posts: 3519 Location: Earth, lately
|
Posted: Fri Aug 30, 2013 2:23 pm Post subject: |
|
|
I'm shrubbing Poem now. I had to update a few times to get about a dozen WU's, went to bed and woke up to a full cache. The problem is Boinc Manager's habit of not requesting tasks for up to 24hrs once it gets a 'project has no work' message. Just update manually to get some work and tell Boinc Manager to report results immediately so it updates every time you finish a WU.
 _________________
(older, before split CPID)
 |
|
Back to top |
|
 |
Plomos Prince

Joined: 26 Apr 2011 Posts: 859
|
Posted: Fri Aug 30, 2013 3:36 pm Post subject: |
|
|
How do i tell it to report immediately when i am not sitting here and it is still allowed to get new tasks? _________________
 |
|
Back to top |
|
 |
Pooh Bear 27 Prince


Joined: 28 Jan 2005 Posts: 1358 Location: Fond du Lac, WI
|
Posted: Fri Aug 30, 2013 4:45 pm Post subject: |
|
|
The timer is somewhat controlled by the project, unless the project is down then a back off timer is used. Many projects only set that wait period for minutes, or even an hour or two. Poem sets it at 24 hours. _________________ Watch the movie about me The 4-1-4s: The Original Teenage Hackers (It is only about 12 minutes long.)
My lucky numbers are 121*2^4553899-1 and 3756801695685*2^666669±1
 |
|
Back to top |
|
 |
Gemjunkie Prince


Joined: 03 Jul 2010 Posts: 3519 Location: Earth, lately
|
Posted: Fri Aug 30, 2013 9:29 pm Post subject: |
|
|
Do you have a config.xml file in place? This is mine on this PhenomIIx6:
Code: | <cc_config>
<log_flags>
<coproc_debug>0</coproc_debug>
<sched_op_debug>0</sched_op_debug>
<work_fetch_debug>0</work_fetch_debug>
</log_flags>
<options>
<max_file_xfers>10</max_file_xfers>
<max_file_xfers_per_project>2</max_file_xfers_per_project>
<report_results_immediately>1</report_results_immediately>
<save_stats_days>360</save_stats_days>
<ncpus>6</ncpus>
</options>
</cc_config> |
The filename should be cc_config.xml
It should be placed in:
C:\ProgramData\BOINC\cc_config.xml
(Ex. under Win7)
Naturally, don't put ncpus>6</ncpus if you don't have 6 cores.
There's more info here:
boinc.berkeley.edu/wiki/Client_configuration _________________
(older, before split CPID)
 |
|
Back to top |
|
 |
LanDroid Prince


Joined: 11 Jun 2002 Posts: 4476 Location: Cincinnati, OH U.S.
|
Posted: Sat Aug 31, 2013 7:51 am Post subject: |
|
|
So if we just want a tweak to "report immediately", it would look as follows? I don't know what those other options do, seems to work fine without those tweaks?
<cc_config>
<options>
<report_results_immediately>1</report_results_immediately>
</options>
</cc_config>
Here's another FAQ on configgerin' this file. Noticed one in particular.
<ncpus>
Act as if there were N CPUs: run N tasks at once. This is for debugging, i.e. to simulate 2 CPUs on a machine that has only 1. Don't use it to limit the number of CPUs used by BOINC; use general preferences instead.
For use with CUDA: When you want to do CUDA only, not use the CPUs at all, set <ncpus>0</ncpus>
If you want to use all CPUs, either omit using this option, or set <ncpus>-1</ncpus>
http://boincfaq.mundayweb.com/index.php?language=1&view=91 _________________ [img]http://stats.free-dc.org/cpidtagb.php?cpid=6533a276b3a1dde393be350eb3cfda70&theme=16&cols=5/.png[/img]
[img]http://stats.free-dc.org/badgesbanner.php?cpid=6533a276b3a1dde393be350eb3cfda70/.png[/img] |
|
Back to top |
|
 |
Plomos Prince

Joined: 26 Apr 2011 Posts: 859
|
Posted: Sat Aug 31, 2013 7:59 am Post subject: |
|
|
Ok got it all set, thanks Sir Gem.
Made sure to change the 6 to a 4 _________________
 |
|
Back to top |
|
 |
Gemjunkie Prince


Joined: 03 Jul 2010 Posts: 3519 Location: Earth, lately
|
Posted: Sat Aug 31, 2013 12:10 pm Post subject: |
|
|
You don't need to include ncpus at all if you're not going to manipulate it. I leave it in as I sometimes want to increase it for projects that limit tasks in progress on a host according to the number of cores. _________________
(older, before split CPID)
 |
|
Back to top |
|
 |
Plomos Prince

Joined: 26 Apr 2011 Posts: 859
|
Posted: Sat Aug 31, 2013 10:17 pm Post subject: |
|
|
i went ahead and removed that line. Working thx _________________
 |
|
Back to top |
|
 |
|