Hi Dave

I put a sticky in Preferences about getting rid of stop/start BOINC options.
It may make it easier to answer.

markmoss

If you want to get rid of the errors, run those models on Linux.

Thanks Les, if I had been more awake I would just have linked to the sticky :)

... and if it's the pop-up messages that are annoying then run BOINC in service mode. The messages are sometimes caused by a local problem and sometimes by a model that is crashing: in service mode a crashing model will still crash, but silently.

I would be happy to continue my many years of running climateprediction.net if it would never cause these Visual Fortran errors. Is there a notification that I could receive if this is fixed?

---

Hello Christopher

The errors seem to mostly be about a computer doing something that
interrupts the main program unexpectedly, causing it to fail. It then
issues a message about where it has failed.

One of the problems is that the "short" models don't clean up after they finish. And NO model type cleans up its many miscellaneous files if it fails.
As a consequence, people have to do this manually.

If your area allocated for BOINC and/or cpdn has filled up with the
remnants of failed models plus all of the "short" models, then your
problem may be "failure to write", which could be due to effectively being out of disk space.
These errors occur on some computers some of the time.
They start appearing mysteriously, and just as mysteriously stop.

Thank you for the informative reply. Luckily, I can still run climateprediction on my Mac, where there are none of these errors.

---

Luckily, I can still run climateprediction on my Mac, where there are none of these errors.

As they are likely to be different model types and not having this error yes, you should be ok once there are some tasks for Mac's again.
May be only of academic interest now but in the days when tasks would go out to multiple operating systems, the ones that failed with the fortran errors on Windows machines would also fail on nix and macs.

Was getting Visual Fortran run-time errors for a hadcm3s task.

System: Win 10 x64 system with an i7-3770K @3GHz & two GTX970's. Stable system.

Was only running one climate model. Other tasks include WCG, TN_Grid, PG, GCC (NCI) & GPUGrid. CPU not saturated.

Intel(r) Visual Fortran run-time error

forrtl: severe (24): end-of-file during read, unit 5, file C:\BOINC\projects\climateprediction.net\hadcm3s_a036_203412_120_599_011122128\jobs\climate.cpdc, line 873, position 0
Image ...
hadcm3s_um_8.34_w ...

Stack trace terminated abnormally.

The Intel(r) Visual Fortran run-time error kept popping up when I closed it (after 3sec). However, the task kept running & progressing (until it failed with Computation error). Tried restarting - same error.

I expected the progress wasn't genuine & I was considering ending the task when it failed. I will probably just stop asking for work, but I'll check back to see if there is any better option. Have read some of the threads & posts on this & think I'm probably just not going to be able to run the app on said system. Thanks,

Task 20562523

Name hadcm3s_a036_203412_120_599_011122128_1
Workunit 11122128
Created 22 Jul 2017, 3:21:22 UTC
Sent 22 Jul 2017, 3:21:33 UTC
Report deadline 4 Jul 2018, 8:41:33 UTC
Received 22 Jul 2017, 8:43:40 UTC
Server state Over
Outcome Computation error
Client state Compute error
Exit status 22 (0x16) Unknown error number
Computer ID 1269301
Run time 37 min 35 sec
CPU time
Validate state Invalid
Credit 0.00
Device peak FLOPS 3.38 GFLOPS
Application version UK Met Office HadCM3 short v8.34
windows_intelx86
Peak working set size 49.77 MB
Peak swap size 156.57 MB
Peak disk usage 0.01 MB
stderr out

<core_client_version>7.6.33</core_client_version>
<![CDATA[
<message>
The device does not recognize the command.
(0x16) - exit code 22 (0x16)
</message>
<stderr_txt>
Suspended CPDN Monitor - Suspend request from BOINC...
Suspended CPDN Monitor - Quit request from BOINC...
Controller:: CPDN process is not running, exiting, bRetVal = 1, checkPID=0, selfPID=7420, iMonCtr=1
Model crash detected, will try to restart...
Controller:: CPDN process is not running, exiting, bRetVal = 1, checkPID=0, selfPID=7420, iMonCtr=1
Model crash detected, will try to restart...
Controller:: CPDN process is not running, exiting, bRetVal = 1, checkPID=0, selfPID=7420, iMonCtr=1
Model crash detected, will try to restart...
Controller:: CPDN process is not running, exiting, bRetVal = 1, checkPID=0, selfPID=7420, iMonCtr=1
Model crash detected, will try to restart...
Controller:: CPDN process is not running, exiting, bRetVal = 1, checkPID=0, selfPID=7420, iMonCtr=1
Model crash detected, will try to restart...
Controller:: CPDN process is not running, exiting, bRetVal = 1, checkPID=0, selfPID=7420, iMonCtr=1
Model crash detected, will try to restart...
Sorry, too many model crashes! :-(
09:35:31 (7420): called boinc_finish(22)

</stderr_txt>
]]>

---

see this thread: HadCM3 short 8.34 workunits errors in number crunching.