climateprediction.net home page
Posts by Nigel Garvey

Posts by Nigel Garvey

1) Message boards : Number crunching : Work Units for MacIntosh (Message 58910)
Posted 24 Oct 2018 by Nigel Garvey
Post:
Art Masson wrote:
I'm considering dropping the CPDN project on my Mac, since there has been no work unit developed or available for quite some time. Any insight appreciated.

I've considered detaching from this project too, since no-one there seems capable of developing a Mac app that works. But I've been put off by CPDN's purge of no-long-active users a year or two back, which essentially disassociated them from their CPDN credit (eg. https://stats.free-dc.org/userdate/bcpdn/14734). At least having my Macs ask for tasks every now and then seems to be keeping them "active" in my account, which hopefully will protect it from similar purges in the future. The inconvenience to me and my computers is minimal.

This shouldn't be an issue if you're also running other platforms on your account (as I see you are).
2) Message boards : Number crunching : Credit Status (Message 57326)
Posted 6 Nov 2017 by Nigel Garvey
Post:
Since CPDN is from UK, they have to comply with UK law, and now, apparently the law mandates that they have to delete inactive accounts. It's a bit stupid, we aren't giving CPDN any crucial info.


Given that this is the UK and that this is CPDN, it's more likely to be either the law being applied stupidly or the law being used as an excuse to cover something else such as freeing up server space or reducing the time spent updating credit.

Two other projects for which I've crunched in the past have shut down altogether, but their participants' association with their credit hasn't been expunged from the BOINC record, as has apparently happened here: http://stats.free-dc.org/stats.php?page=userdate&proj=bcpdn&startdate=14734. (An additional, CPDN-specific cross-project identifier has been generated for every user deleted from CPDN, thus shifting the strain onto the stats sites' servers.)

This was in the news thread last month:
From Andy at CPDN staff:



Project credits from these accounts will be retained but will no longer be linked to an email address or username.


What a useless statement! :D

Current, active, volunteers, or those who have been active in the last three years, will not be affected by this policy.


What does "current, active" mean, exactly? Will all the remaining Mac and Linux users' accounts be deleted too in three years' time if viable apps aren't produced for them by then? Or will we be able to hang on to our current credit simply by requesting tasks in vain for the rest of our lives?
3) Message boards : Number crunching : Credit Status (Message 57219)
Posted 27 Oct 2017 by Nigel Garvey
Post:
So, do I care about my 'loss'? Not a whit! Even if I did care, I know from experience, that the folks at Oxford will sort it out when the important stuff is sorted. You lose nothing by hanging-in with what has always been true.


I've never understood the obsession with credits.


It's not just an "obsession with credits". No-one seriously expects to crunch for CPDN and go sailing up to the top of the BOINC charts. The reason many people get sensitive about credit is that CPDN is supposedly a BOINC project — and the ethos of BOINC is that people allow projects' tasks to run on their computers and in return receive credit. It's a two-way thing. The usual understanding is that the credit will be granted immediately the work's validated, not just when project administrators don't have anything "more important" to do. Credits aren't worth anything in material terms terms and don't enhance anyone's status in the real world, but they're a nice recognition of the bother and expense people have gone to on your project's behalf and a helpful indicator of the usefulness of their results.

CPDN's tasks are long and prone to crashing and have to be nurtured through by leaving the machines switched on permanently and without restarts for several days or weeks at a time. Anyone who continues to crunch for the project on this basis after a few weeks of crashing models, models which only run on certain types of computer, bugs in the apps for certain types of computer which no-one knows how to fix, server problems which can only be addressed when it's not a weekend or a bank holiday, and credit scripts which are only run once every two or three months is dedicated and deserves to be treated with respect. With all CPDN's existing peculiarites, it's perfectly natural for people to feel nervous when yet another glitch appears to wipe out their credit. Haughty replies about what's important do little to alleviate their fears about the competence and attitudes at that end or to reassure them that crunching for the project isn't actually contributing more to global warming than to a solution for it.
4) Questions and Answers : Macintosh : wah2 v8.24 (Message 56252)
Posted 18 May 2017 by Nigel Garvey
Post:
I wrote:
Just to note that my first task with the new app bombed out during its first upload.


I received another task later that day — very much shorter at just over 17 hours to process instead of several days — which has survived and been recorded on the site as a success. Only three trickle files, plus a "_restart" file at the same time as the first and an "_out" file two minutes after the last.
5) Questions and Answers : Macintosh : wah2 v8.24 (Message 56226)
Posted 16 May 2017 by Nigel Garvey
Post:
Just to note that my first task with the new app bombed out during its first upload. Rather disappointing after receiving no credit for 1,260,505.84 seconds spent on the previous task.

Mon 15 May 16:04:06 2017 | climateprediction.net | Starting task wah2_eas50_n0ju_201212_12_567_011011021_0

Tue 16 May 09:22:01 2017 | climateprediction.net | Started upload of wah2_eas50_n0ju_201212_12_567_011011021_0_r1299423490_out.zip
Tue 16 May 09:22:02 2017 | climateprediction.net | Computation for task wah2_eas50_n0ju_201212_12_567_011011021_0 finished
Tue 16 May 09:22:02 2017 | climateprediction.net | Output file wah2_eas50_n0ju_201212_12_567_011011021_0_r1299423490_1.zip for task wah2_eas50_n0ju_201212_12_567_011011021_0 absent
Tue 16 May 09:22:02 2017 | climateprediction.net | Output file wah2_eas50_n0ju_201212_12_567_011011021_0_r1299423490_2.zip for task wah2_eas50_n0ju_201212_12_567_011011021_0 absent
Tue 16 May 09:22:02 2017 | climateprediction.net | Output file wah2_eas50_n0ju_201212_12_567_011011021_0_r1299423490_3.zip for task wah2_eas50_n0ju_201212_12_567_011011021_0 absent
Tue 16 May 09:22:02 2017 | climateprediction.net | Output file wah2_eas50_n0ju_201212_12_567_011011021_0_r1299423490_4.zip for task wah2_eas50_n0ju_201212_12_567_011011021_0 absent
Tue 16 May 09:22:02 2017 | climateprediction.net | Output file wah2_eas50_n0ju_201212_12_567_011011021_0_r1299423490_5.zip for task wah2_eas50_n0ju_201212_12_567_011011021_0 absent
Tue 16 May 09:22:02 2017 | climateprediction.net | Output file wah2_eas50_n0ju_201212_12_567_011011021_0_r1299423490_6.zip for task wah2_eas50_n0ju_201212_12_567_011011021_0 absent
Tue 16 May 09:22:02 2017 | climateprediction.net | Output file wah2_eas50_n0ju_201212_12_567_011011021_0_r1299423490_7.zip for task wah2_eas50_n0ju_201212_12_567_011011021_0 absent
Tue 16 May 09:22:02 2017 | climateprediction.net | Output file wah2_eas50_n0ju_201212_12_567_011011021_0_r1299423490_8.zip for task wah2_eas50_n0ju_201212_12_567_011011021_0 absent
Tue 16 May 09:22:02 2017 | climateprediction.net | Output file wah2_eas50_n0ju_201212_12_567_011011021_0_r1299423490_9.zip for task wah2_eas50_n0ju_201212_12_567_011011021_0 absent
Tue 16 May 09:22:02 2017 | climateprediction.net | Output file wah2_eas50_n0ju_201212_12_567_011011021_0_r1299423490_10.zip for task wah2_eas50_n0ju_201212_12_567_011011021_0 absent
Tue 16 May 09:22:02 2017 | climateprediction.net | Output file wah2_eas50_n0ju_201212_12_567_011011021_0_r1299423490_11.zip for task wah2_eas50_n0ju_201212_12_567_011011021_0 absent
Tue 16 May 09:22:02 2017 | climateprediction.net | Output file wah2_eas50_n0ju_201212_12_567_011011021_0_r1299423490_12.zip for task wah2_eas50_n0ju_201212_12_567_011011021_0 absent
Tue 16 May 09:22:02 2017 | climateprediction.net | Output file wah2_eas50_n0ju_201212_12_567_011011021_0_r1299423490_restart.zip for task wah2_eas50_n0ju_201212_12_567_011011021_0 absent
Tue 16 May 09:22:07 2017 | climateprediction.net | Finished upload of wah2_eas50_n0ju_201212_12_567_011011021_0_r1299423490_out.zip
6) Questions and Answers : Macintosh : No Tasks Available for Mac? (Message 54298)
Posted 13 Jun 2016 by Nigel Garvey
Post:
I wrote:
OK. The upload took two and a half hours, but the server was up to it. ;) The task was successfully reported an hour later. All smiles on the task page, except (as I write) for the absence of trickles.


Many apologies. The link in my earlier post is to an hadcm3n task I had in March/April. I've no idea how that happened. The correct link is http://climateapps2.oerc.ox.ac.uk/cpdnboinc/result.php?resultid=19601252.
7) Questions and Answers : Macintosh : No Tasks Available for Mac? (Message 54294)
Posted 13 Jun 2016 by Nigel Garvey
Post:
OK. The upload took two and a half hours, but the server was up to it. ;) The task was successfully reported an hour later. All smiles on the task page, except (as I write) for the absence of trickles.
8) Questions and Answers : Macintosh : No Tasks Available for Mac? (Message 54293)
Posted 13 Jun 2016 by Nigel Garvey
Post:
My own task too has finished overnight and the zips have all appeared in BOINC Manager's Transfers tab, ready for when BOINC reconnects to the network in a few minutes' time. The file sizes add up to 1004.04 MB in the Manager, but to 1052.7 MB in the Mac OS folder. Presumably this is something to do with Mac OS now using decimal kilo sizes. There's no mention of impending trickle messages in BOINC Manager's Projects tab.
9) Questions and Answers : Macintosh : No Tasks Available for Mac? (Message 54247)
Posted 3 Jun 2016 by Nigel Garvey
Post:
Thanks, Peter. My own current task is still running and has an estimated week of processor time to go before it finishes. Uploading fourteen 73MB files may take a while too. ;) I'll report what happens.
10) Questions and Answers : Macintosh : No Tasks Available for Mac? (Message 54244)
Posted 2 Jun 2016 by Nigel Garvey
Post:
Thanks again. I'm not bothered about the graphics. I just mentioned their absence here in case it should turn out to be a clue. :)
11) Questions and Answers : Macintosh : No Tasks Available for Mac? (Message 54242)
Posted 2 Jun 2016 by Nigel Garvey
Post:
Ah right. Thanks. How does one tell what kind one has — apart from the frequency with which the trickles and uploads don't happen? ;)

Also, I have no graphics with this app (wah2). Is that right?
12) Questions and Answers : Macintosh : No Tasks Available for Mac? (Message 54240)
Posted 2 Jun 2016 by Nigel Garvey
Post:
I have three what look like upload zips in the project folder (wah2_eu25_120h_203212_13_395_010518631_0_1.zip to wah2_eu25_120h_203212_13_395_010518631_0_3.zip) and nothing that looks like a trickle. Current task progress is 27.76%. Presumably the zip interval's 8.33% (ie. 12 or 13 zips per task) as with HadAM3Ps, although this doesn't chime with Iain having nine zips ready to go after only 39%.
13) Questions and Answers : Macintosh : No Tasks Available for Mac? (Message 54234)
Posted 2 Jun 2016 by Nigel Garvey
Post:
Iain Inglis wrote:
That theory didn't last long: no trickle, no Zip by this afternoon. Checked the model's XML files, which expect 25 Zips and show the model at 29% (81,409 steps out of an expected 276,779). So the progress indicator is right but no trickles or Zips are being produced. There's a problem somewhere.


Just to confirm the absence of trickles. The one wah2 task I have has just passed 26% in BOINC Manager's Tasks tab and is dominating the pie chart in the Disk tab, but hasn't attempted any trickles or uploads so far. No mention of it in the Event Log since it was downloaded and started on Sunday.

http://climateapps2.oerc.ox.ac.uk/cpdnboinc/result.php?resultid=19601252
14) Questions and Answers : Macintosh : UK Met Office HadAM3P-HadRM3P Australia New Zealand v6.10 (Message 53358)
Posted 28 Jan 2016 by Nigel Garvey
Post:
15) Questions and Answers : Macintosh : UK Met Office HadAM3P-HadRM3P Australia New Zealand v6.10 (Message 53355)
Posted 28 Jan 2016 by Nigel Garvey
Post:
This fault hasn't been fixed three days later but another batch of the tasks has just been released. Seems a trifle clueless to me.
16) Message boards : Number crunching : Computing Preferences Percentage? (Message 52715)
Posted 13 Oct 2015 by Nigel Garvey
Post:
17) Questions and Answers : Macintosh : BOINC on El Capitan OS? (Message 52711)
Posted 11 Oct 2015 by Nigel Garvey
Post:
I've had no problems with BOINC itself since upgrading to El Capitan, although I had to reinstall it after the upgrade to re-establish its permissions. I've not had any CPDN tasks since before the upgrade, so I can't say anything about these specifically.

BOINC itself recently had a "recommended" update to 7.6.9, shortly after which the Mac recommended version went up to 7.6.11 and I see while writing this that it's now 7.6.12. I've no idea what issues have been addressed, but presumably its worth having the latest Mac version.
18) Questions and Answers : Macintosh : OS X Yosemite (Message 50675)
Posted 29 Oct 2014 by Nigel Garvey
Post:
19) Message boards : Number crunching : HadCM3 short - errors galore (Message 50545)
Posted 17 Oct 2014 by Nigel Garvey
Post:
You also had:
process exited with code 9
Might be a Mac issue, or it could be related to you using a test version of BOINC.
The 2 trickle_up files got returned, so it may just be a post-processing error.


Thanks for the interesting info.

But most of my failures are the "code 9", so...


Similarly here.

My Mac's had three "shorts" so far. The first two ran to completion, but their final uploads took surprisingly little time and their outcomes are shown on the site as "Client error", the Stderr reports being "process exited with code 9 (0x9, -247)". No errors were reported in my BOINC Manager (7.2.42).

The third's currently at 48% completion and has just successfully completed its first upload.

All three tasks have also been attempted by Windows machines, which have errored after only a few seconds. The second task was later successfully completed by a Linux machine, which got the same amount of credit for it as my Mac did.
20) Message boards : Number crunching : Possible problem with new EU work units (Message 49557)
Posted 13 Jul 2014 by Nigel Garvey
Post:
Iain Inglis wrote:
The one I've just finished returned 13 trickles! I thought the zip file numbers were out with the percentage completion in the later stages too, but the correct number (13) appear to have been uploaded, with no gaps or additions (according to my BOINC "stdoutdae.txt" file).

That is odd. The alien trickle is this one:

08 Jul 2014 10:08:52 1073433 16706651 hadam3p_eu_j9dj_2013_1_008791627_0 1 69,222 225,267 3.2543


Hmm. It's timed shortly after trickles 5 and 6, which, because my BOINC client's set not to use the network overnight, were both sent at the same time one morning after I'd manually enabled network access earlier than the set time. The corresponding zip files, however, kept failing to upload. The reason given was "transient HTTP error", but I knew from previous experience that this probaby meant my Mac was overdue for a reboot. So, at an opportune moment, I carefully suspended all BOINC computation, quit BOINC Manager, and shut down the machine. Then I restarted it and resumed BOINC computation once everything was up and running again. The two zip files immediately uploaded with no problem at all and the additional trickle was apparently sent an hour after the uploads completed, being recorded on the task page five minutes after that.

This is probably all connected, but I can't think why a spurious trickle should have been sent and with an incorrect timestep.


Next 20

©2020 climateprediction.net