Atom topic feed | site map | contact | login | Protection des données personnelles | Powered by FluxBB | réalisation artaban
You are not logged in.
Pages: 1
Hello,
I ran into an error (Ubuntu 11.04 64bit on QuadCore Intel CPU) compiling mpi version of aster. The error from the mpi version is shown below:
--------------------------------------------------------------------------------
Prepare cata_ele.pickled
build cata_ele.pickled with make_capy_offi.py [ OK ]
<F>_PICKLED_ERROR error during making cata_ele.pickled
--------------------------------------------------------------------------------
<F> errors previously raised :
<F>_PICKLED_ERROR error during making cata_ele.pickled
<E>_UPDATE_FAILED Exception raised by AutoUpdate:
------------------------------------------------------------
Traceback:
Traceback (most recent call last):
File "/opt/aster/lib/python2.7/site-packages/asrun/maintenance.py", line 1309, in AutoUpdate
Update(run, num_update=num_update, *lpkg)
File "/opt/aster/lib/python2.7/site-packages/asrun/maintenance.py", line 1189, in Update
_build_aster(run, conf, True, REPREF, None, reptrav_built, lardv=lardv)
File "/opt/aster/lib/python2.7/site-packages/asrun/maintenance.py", line 962, in _build_aster
kret = build.MakePickled(REPREF, reptrav, repdest=destdir, **kargs)
File "/opt/aster/lib/python2.7/site-packages/asrun/build.py", line 1170, in MakePickled
'<F>_PICKLED_ERROR')
File "/opt/aster/lib/python2.7/site-packages/asrun/build.py", line 156, in _mess
self.run.Mess(msg, cod, store)
File "/opt/aster/lib/python2.7/site-packages/asrun/run.py", line 640, in Mess
self.Sortie(4)
File "/opt/aster/lib/python2.7/site-packages/asrun/run.py", line 679, in Sortie
raise RunAsterError, exit_code
RunAsterError: 4
------------------------------------------------------------
Exit code : 4
--------------------------------------------------------------------------------
<F> errors previously raised :
<F>_PICKLED_ERROR error during making cata_ele.pickled
---------------------------------------------------------------------------------
cpu system cpu+sys elapsed
---------------------------------------------------------------------------------
Download archives 0.40 0.65 1.05 461.57
Extraction of archives 2.44 0.52 2.96 3.42
Copy of updated files 0.15 0.59 0.74 5.10
Deletion of old files 0.40 0.70 1.10 2.96
Compilation in debug mode 6.26 2.49 8.75 12.16
Add object files to library 0.61 0.92 1.53 9.06
Build executables 5.75 2.03 7.78 9.56
Test executables 0.69 0.22 0.91 5.08
Compilation in nodebug mode 13.10 1.82 14.92 7.70
Compilation of commands catalogue 0.39 0.26 0.65 3.72
Make pickled of elements 0.38 0.07 0.45 2.52
---------------------------------------------------------------------------------
Total 30.60 10.28 40.88 522.89
---------------------------------------------------------------------------------
(*) cpu and system times are not correctly counted using mpirun.
as_run 1.10.2
------------------------------------------------------------
--- DIAGNOSTIC JOB : <F>_PICKLED_ERROR
------------------------------------------------------------
EXIT_CODE=4
Same everything (except SingleCore AMD Athlon PC "ubuntu31") updated with no problems! Why / what is happening? How can I diagnose or help diagnose the problem?
uname -a
Linux ubuntu31 2.6.38-11-generic #50-Ubuntu SMP Mon Sep 12 21:17:25 UTC 2011 x86_64 x86_64 x86_64 GNU/Linux -OK
Linux ubuntu32 2.6.38-11-generic #50-Ubuntu SMP Mon Sep 12 21:17:25 UTC 2011 x86_64 x86_64 x86_64 GNU/Linux -NOK
Linux ubuntu33 2.6.38-11-generic #50-Ubuntu SMP Mon Sep 12 21:17:25 UTC 2011 x86_64 x86_64 x86_64 GNU/Linux - NOK
Regards, JMB
Last edited by JMB365 (2011-10-27 02:47:24)
SalomeMeca 2021
Ubuntu 20.04, 22.04
Offline
Hello,
As an update, when I ran:
cd /opt/aster/NEW11.0
/opt/aster/bin/as_run --make > /tmp/output 2> /tmp/error
cd /opt/aster/NEW11.0_mpi
/opt/aster/bin/as_run --vers=NEW11.0_mpi --make > /tmp/output 2> /tmp/error
The /tmp/output shows that everything compiled fine on the sequential version! But the error remains when compiling the mpi version. So there is an elusive BUG somewhere...
Maybe this topic gives us the clue: http://www.code-aster.org/forum2/viewtopic.php?id=12856 Hopefully a clean solution to the problem can be found...
Regards, JMB
Last edited by JMB365 (2011-10-27 03:26:02)
SalomeMeca 2021
Ubuntu 20.04, 22.04
Offline
Please attach the output and error files.
Code_Aster release : last unstable on Ubuntu 16.04 64 bits - GNU Compilers
Please do not forget to tag your first post as *SOLVED* when it is!
Offline
Hello courtois,
Thanks for looking into it. The files are attached.
Regards, JMB
SalomeMeca 2021
Ubuntu 20.04, 22.04
Offline
Hello,
It does not seem normal that ALL the 30 test (list /opt/aster/NEW11.0_mpi/astest/liste_short) would fail with INCORRECT_PARA, ERROR or NO_RESU_FILE; as I am getting. Can somebody comment on whether the changes being made to the test cases in the latest versions (11.0.30) are causing this or is my compiled asteru_mpi at fault? Also a key indicator is that the full suite of tests has no [OK] tests. This makes me suspect my asteru_mpi Thanks!
Regards, JMB
Last edited by JMB365 (2011-10-27 17:06:57)
SalomeMeca 2021
Ubuntu 20.04, 22.04
Offline
Hi,
I had a similar error and when I wanted to checked the files, there were none. All the files that gave ERROR_PARAM weren't in the astout folder. Did you checked those files?
Cheers
S.
Offline
Hi,
It seems that some files (at least in the astest directory) have not been correctly updated.
Perhaps because the forma* testcases have been heavily changed (removed and re-added later)...
...a bug in as_run ? I can't believe it
MC
Code_Aster release : last unstable on Ubuntu 16.04 64 bits - GNU Compilers
Please do not forget to tag your first post as *SOLVED* when it is!
Offline
It seems that some files (at least in the astest directory) have not been correctly updated.
Perhaps because the forma* testcases have been heavily changed (removed and re-added later)...
...a bug in as_run ? I can't believe it ;)MC
I hear what you are saying, but I have a situation were every one of the 30 (liste_short) tests fail and every one of the 2400+ tests are failing for one reason or another wherein after 500 errors the testing just stops. Now that is abnormal... So I continue to suspect my .../NEW11.0_mpi/asteru_mpi
Regards, JMB
SalomeMeca 2021
Ubuntu 20.04, 22.04
Offline
Did you checked those files?.
Thanks for the suggestion, but those files do exist. The mpi version of CodeAster is failing on every one of them, while the sequential version passes OK with most of them.
Regards, JMB
SalomeMeca 2021
Ubuntu 20.04, 22.04
Offline
The mpi version of CodeAster is failing on every one of them
No, your mpi version fails and that's what we have to understand !
Code_Aster release : last unstable on Ubuntu 16.04 64 bits - GNU Compilers
Please do not forget to tag your first post as *SOLVED* when it is!
Offline
No, your mpi version fails and that's what we have to understand !
Before you read too much into what you think I may be implying; please note that two postings ago I stated "So I continue to suspect my .../NEW11.0_mpi/asteru_mpi
-JMB
SalomeMeca 2021
Ubuntu 20.04, 22.04
Offline
We may have more infos by adding the option '-g' :
as_run -g --make --vers=NEW11_mpi > /tmp/output
I checked it, it will print the Code_Aster output when compiling the file of elements.
Hope this will show us the problem !
MC
Code_Aster release : last unstable on Ubuntu 16.04 64 bits - GNU Compilers
Please do not forget to tag your first post as *SOLVED* when it is!
Offline
Hello courtois,
Thanks for looking into the problem. The files are attached below.
Regards, JMB
SalomeMeca 2021
Ubuntu 20.04, 22.04
Offline
There is yet the problem you reported earlier :
testing executable /opt/aster/NEW11.0_mpi/asteru_mpi... [ OK ]
mpiexec: unable to start all procs; may have invalid machine names
remaining specified hosts:
192.168.1.231 (ubuntu31.mshome)
But it exits with exit code = 0 (OK) so I don't know if it starts or not and it prints no output ?!
This thread is probably not solved :-) : http://www.code-aster.org/forum2/viewtopic.php?id=15853
Try running your Pi executable with the same command lines as used by as_run (see the error file).
Do you have several MPI installation on the machines ? Check that /usr/local/bin/mpirun comes with mpdboot ?
(Probably not relevant here but) mpiexec and mpirun point to the same target but nevertheless they may have a different behavior (show and get point together to as_run...).
MC
Code_Aster release : last unstable on Ubuntu 16.04 64 bits - GNU Compilers
Please do not forget to tag your first post as *SOLVED* when it is!
Offline
Hello courtois,
Thank for all the suggestions. Let me look into them more carefully. I am sure that the calculation of Pi worked correctly on the 3-noded cluster, as reported by me in Msg#8 of http://www.code-aster.org/forum2/viewtopic.php?id=15853
I may have zipped up the wrong output/error files in my previous message. Currently I am trying to install CAELinux2011 on two other PCs so that I can compare configurations, in an effort to better understand what I need to do for the problems I am experiencing. After I investigate as thoroughly as I can, I will respond back. Have a good weekend!
Regards,
JMB
SalomeMeca 2021
Ubuntu 20.04, 22.04
Offline
I just suggested to
Try running your Pi executable with the same command lines as used by as_run (see the error file).
to see what does not work on your installation : comparing the command line (yours) which works and the one (as_run) it does not !
MC
Code_Aster release : last unstable on Ubuntu 16.04 64 bits - GNU Compilers
Please do not forget to tag your first post as *SOLVED* when it is!
Offline
Pages: 1