climateprediction.net home page
Posts by Belfry

Posts by Belfry

1) Message boards : climateprediction.net Science : Misconfiguration e-mail (Message 52020)
Posted 5 Jun 2015 by Belfry
Post:
Yikes! 1352118.
2) Message boards : Number crunching : Weird Clouds on hadam3p_anz (Message 50372)
Posted 3 Oct 2014 by Belfry
Post:
I've seen this before, the Selectric effect--simultaneously annoying and fantastic IMHO. I think this is related to a permissions issue that crops up on Linux machines--see this post. Check the permissions on the graphic jpegs and support files. Run if necessary:

chmod 644 /var/lib/boinc-client/projects/climateprediction.net/*.jpg
chmod 755 /var/lib/boinc-client/projects/climateprediction.net/t*f
chmod 644 /var/lib/boinc-client/projects/climateprediction.net/t*f/*
3) Message boards : Number crunching : UK Met Office HADAM3P (global only) with MOSES II landsurface scheme v7.03 (Message 49293)
Posted 3 Jun 2014 by Belfry
Post:
So I restarted two from scratch and ran them in lock-step along with one newly started.

16603602: restart--finished!
16608984: restart--failed.
16611240: new--failed.

The failures show this message repeated 5 to 6 times near the end in stderr: "Model crashed: æM".

Tempermental things, for sure.

(I learned to restart when working with Iain's slab model anaylsis--just involves some careful file deletion and xml editing.)
4) Message boards : Number crunching : Credit updates? (Message 49184)
Posted 21 May 2014 by Belfry
Post:
1) I know.
2) It undoubtedly has some effect. When CPDN credits freeze for an extended period (like last August/September) I have to suspend CPDN tasks to get any WCG tasks done. Maybe disabling BOINC network activity for the entire credit freeze period would work, but it's impractical for several reasons (10-day work limit among them).

It has happened many times over the past five years or so. I also know of it happening to another member of my team. If you don't believe me, join WCG with a CPDN machine and assign it 25%. Then wait for the next CPDN credit freeze.

Before all the pure science people jump all over me, let me make clear: I'd prefer BOINC abandon credits altogether if it would mean running without user intervention.
5) Message boards : Number crunching : Credit updates? (Message 49182)
Posted 21 May 2014 by Belfry
Post:
And don't forget people who run multiple projects on a single machine: credit freezes can lead to inequitable CPU time distribution (often leading to a user having to suspend/activate on a task-by-task basis). Fortunately CPDN is not having that problem right now.
6) Message boards : Number crunching : UK Met Office HADAM3P (global only) with MOSES II landsurface scheme v7.03 (Message 49129)
Posted 16 May 2014 by Belfry
Post:
Thanks. I see now from your earlier post that I made you repeat yourself, so I apologize. I already rebooted because of a kernel update, so I'll see if I can restart these from the beginning.
7) Message boards : Number crunching : UK Met Office HADAM3P (global only) with MOSES II landsurface scheme v7.03 (Message 49124)
Posted 16 May 2014 by Belfry
Post:
Can these reboot?
8) Message boards : Number crunching : UK Met Office HADAM3P (global only) with MOSES II landsurface scheme v7.03 (Message 49116)
Posted 15 May 2014 by Belfry
Post:
...and maybe the too-heavily-weighted points situation could be fixed with the next batch? It's just with my predilection for AMD machines there's just no way I merit a spot in the top 30 hosts. (And pause with the recognition that AMD is synonymous with slow).
9) Message boards : Number crunching : UK Met Office HADAM3P (global only) with MOSES II landsurface scheme v7.03 (Message 49115)
Posted 15 May 2014 by Belfry
Post:
TWO! (16546960 and 16525908)

Regarding my remaining MOSES II tasks, is there a way I can reboot my machine without these erroring?
10) Message boards : Number crunching : UK Met Office HADAM3P (global only) with MOSES II landsurface scheme v7.03 (Message 48831)
Posted 17 Apr 2014 by Belfry
Post:
Klepel, you might be running into an old issue that keeps cropping up with RHEL derived distributions like CentOS. Unfortunately CPDN's application developers target a distribution with newer libraries than those in these distributions. See this sticky.

If you:
strings /usr/lib/i386-linux-gnu/libstdc++.so.6 | grep GLIBCXX

(modify the path for your libstdc++.so.6 location)

...the most recent version supported should be 3.4.10 or greater.
11) Message boards : Number crunching : "No resubmission" tasks (Message 48827)
Posted 17 Apr 2014 by Belfry
Post:
Task aborted--and now it's gone here. Well, at least it's on Linux.

Probably too late for this batch, but perhaps in the future the distribution script could have a few lines added that grep for the '7xxx' and nix them (in Python this can be done with a subprocess call.) The filename check would be more robust than querying the database for 'no resubmission' (which apparently is not working).

Apology in advance if this strategy has already been considered.
12) Message boards : Number crunching : "No resubmission" tasks (Message 48822)
Posted 17 Apr 2014 by Belfry
Post:
No, I saw them. I wondered if it might pique the interest of a scientist or two to know some are completing. On the other hand, if I'm just wasting electricity I'd like to hear a moderator or admin say so.
13) Message boards : Number crunching : "No resubmission" tasks (Message 48819)
Posted 17 Apr 2014 by Belfry
Post:
So the silence means moderators want me to continue with the one have?
14) Message boards : Number crunching : "No resubmission" tasks (Message 48808)
Posted 15 Apr 2014 by Belfry
Post:
Just got one of these and noticed I've completed two in recent months (16275890 and 16291996).

Maybe Linux just rules, eh?
15) Questions and Answers : Unix/Linux : hnddler vs handler (Message 47293)
Posted 12 Oct 2013 by Belfry
Post:
Eirik Redd, does this happen when running other projects?

Ed: I know WCG also uses the "file_upload_handler" syntax.
16) Message boards : Cafe CPDN : Letting BOINC run 24/7 (Message 47211)
Posted 30 Sep 2013 by Belfry
Post:
In my experience it's best to shutdown BOINC before machine shutdown.

Ed: it might be okay to suspend BOINC without the "leave applications in memory while suspended" setting checked, but to the tasks it's essentially the same as shutting down BOINC--me thinks. This setting is probably what you want to save energy and keep the tasks loaded through S3 suspend or hibernation. Check that your suspend/hibernation is good and stable though.
17) Message boards : Cafe CPDN : Letting BOINC run 24/7 (Message 47201)
Posted 29 Sep 2013 by Belfry
Post:
CPDN, as with all distributed computer projects, will increase electricity consumption. Some things you can do:

-Use a laptop.*

-Don't run CPDN or any CPU-based BOINC project on a machine with a dedicated graphics card.

-Run BOINC only certain times of the day and then suspend or hibernate.

-Undervolt and/or underclock.

* Ed: I'd recommend an extra one you don't need to carry around a lot--CPDN tasks are prone to crashing with a lot of shutdown cycles.
18) Message boards : Cafe CPDN : 50:1 Project (Message 47198)
Posted 29 Sep 2013 by Belfry
Post:
Their simplistic formula implies they know exactly what climate change will cost. What if impacts on food prodcution are much worse (e.g. tropical pests and fungi that establish themselves midlatitude, collapses in marine ecosystems)? I'm guessing the 50-to-1 project doesn't price in the mortality required to adapt to a world with 10% fewer calories.
19) Message boards : Number crunching : Virtualisation (running CPDN inside virtual PCs) (Message 47019)
Posted 12 Sep 2013 by Belfry
Post:
In my experience running BOINC on both host (Linux with KVM/Virtualbox) and guest (any OS, all CPU's allocated), the hypervisor will give priority to the guest. Can others verify my config and/or test Windows hosts with all CPUs allocated in guests?

On my laptop running 50/50 CPDN/WCG, I was able to download one hadcm3n by pausing WCG and setting processors to 50%. Then after getting the task, setting CPDN to no new tasks. My CPDN scheduling priority swings wildly in the course of a few minutes after updates (i.e. -15.59 to -.90), so BOINC must need some kind of credit confirmation to balance things out. It's strange though, you'd think it would be based strictly on crunching time.
20) Questions and Answers : Windows : Optimise PC build for CPDN (Message 46970)
Posted 5 Sep 2013 by Belfry
Post:
...,but digging around it seems to only work when configured as RAID.

No, all it requires is an SSD and and HDD. It sounds like it will be active by default on Intel 68 and up even without the configuration program installed (the driver is part of Win7/8).

Ed: actually you're right. For it to work a single HDD must be put in RAID mode--pretty weird.


Next 20

©2020 climateprediction.net