|
[Sponsors] |
Fluent 17.0 with PBS Torque job scheduler on LINUX |
|
LinkBack | Thread Tools | Search this Thread | Display Modes |
December 29, 2019, 11:27 |
Fluent 17.0 with PBS Torque job scheduler on LINUX
|
#1 |
New Member
mardar
Join Date: Dec 2019
Posts: 17
Rep Power: 6 |
Hi,
For a project with Academic Research License HPC I'm trying to use Fluent 17.0 which is installed on my universites HPC (CentOS 6.10) with PBS Torque Job Scheduler. I haven't succeeded yet. I read in Fluent Documentation that LSF, SGE, and PBS Pro are examples of job scheduler on Linux. I want to make sure that PBS Torque can be used for Fluent 17.0 on HPC with CentOS 6.10. Could you confirm it? Also, if you can confirm this, I'd like to ask you why I can't run it with detailed information about my problem. Best Regards |
|
December 29, 2019, 17:05 |
|
#2 |
Senior Member
Alain Islas
Join Date: Nov 2019
Location: Mexico
Posts: 142
Rep Power: 7 |
Yes, you can run Fluent with PBS torque scheduler. You will need to use a script file, containing the #PBS instructions to request queues, nodes, processors, emailing, etc...
Let me know if you need help with this. |
|
December 29, 2019, 17:52 |
|
#3 |
New Member
mardar
Join Date: Dec 2019
Posts: 17
Rep Power: 6 |
Hi alainislas,
Actually, I need help and that would be great When I try to schedule job with "qsub script_filename" from the terminal of HPC. 2 files are created in my home directory in these formats: JobName.oJobID (this file is empty) JobName.eJobID (this file has a message and says that: /var/spool/torque/mom_priv/jobs/23454.poyraz.ae.metu.edu.tr.SC: line 34: ./opt/ansys_inc/v170/fluent/bin/fluent*: No such file or directory script_filename is in the following format: # Give your job a name so it can be found readily with qstat. #PBS -N Your_Job # Specify time limit of your job. (HH:MM:SS) #PBS -l walltime=02:00:00 # Specify the number of threads/cores you need (MAX:64) # number of nodes can only be 1 for shared memory openmp apps #PBS -l nodes=1pn=64 export OMP_NUM_THREADS=64 # mail alert at the (e)nd of execution #PBS -m e #PBS -M your_email_address # Standard error output merge to the standard output file. #PBS -j oe # Specify the file name for the standard output/error #PBS -o output # Enter your work directory under /cfd cd /users/project/aozuzun # execute your shared memory application ./opt/ansys_inc/v170/fluent/bin/fluent* 3ddp -t np -ssh -pib -g -i /users/project/aozuzun/aa.jou |
|
December 29, 2019, 18:32 |
|
#4 |
Senior Member
Alain Islas
Join Date: Nov 2019
Location: Mexico
Posts: 142
Rep Power: 7 |
The output files you are getting, contain the terminal display before and after your job is executed. This is done by #PBS -j oe. If they are empty or showing an error message, this means fluent is not starting correctly. Please refer to the following link http://docs.adaptivecomputing.com/to...%23mailoptions
to setup your script file. Another thing, you may need to change your working directory to the PBS directory, this is done by: cd $PBS_O_WORKDIR Is fluent installed as a module in the HPC? you can verify this by typing on the terminal: module avail if this is the case, load the module and then use the command fluent -g 3ddp -ssh -tx -i journal.jou -g (runs in background mode) 3ddp (3D simulation using double precision) -ssh (Secure Shell turned on) -tx (specify the number of processors, where x is the number) -i (run the fluent with a journal file) Need to specify the console commands to run this. Please check the fluent getting started guide and console commands. Let me know if this is helpful |
|
December 30, 2019, 17:21 |
|
#5 |
New Member
mardar
Join Date: Dec 2019
Posts: 17
Rep Power: 6 |
Firstly, I check the fluent is installed as a module and it gives the message below. Is the module necessary?
[aozuzun@poyraz:~]$ module avail ------------------------ /usr/share/Modules/modulefiles ------------------------ dot module-cvs module-info modules null use.own ------------------------------- /etc/modulefiles ------------------------------- compat-openmpi-i386 openmpi-1.4-i386 openmpi-1.5.3-i386 compat-openmpi-x86_64 openmpi-1.4-x86_64 openmpi-1.5.3-x86_64 ------------------------ /opt/share/Modules/modulefiles ------------------------ Secondly, I found a mistake in my script file and correct it by deleting the dot in last line of my script.The last version is: /opt/ansys_inc/v170/fluent/bin/fluent* -g 3ddp -ssh -t64 -i journal.jou Now, the files created in my home directories says different things. But as I understand at least pbs can start fluent anymore. May I ask what is these errors? The file named "Your_Job.e23465" Error: eval: unbound variable Error Object: /opt/ansys_inc/v170/fluent/bin/fluent_arch Error: eval: unbound variable Error Object: /opt/ansys_inc/v170/fluent/bin/fluent_arch1.0 Error: eval: unbound variable Error Object: /opt/ansys_inc/v170/fluent/bin/fluentbench.pl Error: eval: unbound variable Error Object: /opt/ansys_inc/v170/fluent/bin/fluent-cleanup.pl Error: eval: unbound variable Error Object: /opt/ansys_inc/v170/fluent/bin/fluent_sysinfo Cleanup script file is /users/project/aozuzun/cleanup-fluent-atmaca65-5648.sh The file named "Your_Job.o23469" /opt/ansys_inc/v170/fluent/fluent17.0.0/bin/fluent -r17.0.0 /opt/ansys_inc/v170/fluent/bin/fluent_arch /opt/ansys_inc/v170/fluent/bin/fluent_arch1.0 /opt/ansys_inc/v170/fluent/bin/fluentbench.pl /opt/ansys_inc/v170/fluent/bin/fluent-cleanup.pl /opt/ansys_inc/v170/fluent/bin/fluent_sysinfo -g 3ddp -ssh -t32 -i /users/project/aozuzun/journal.jou /opt/ansys_inc/v170/fluent/fluent17.0.0/cortex/lnamd64/cortex.17.0.0 -f fluent /opt/ansys_inc/v170/fluent/bin/fluent_arch /opt/ansys_inc/v170/fluent/bin/fluent_arch1.0 /opt/ansys_inc/v170/fluent/bin/fluentbench.pl /opt/ansys_inc/v170/fluent/bin/fluent-cleanup.pl /opt/ansys_inc/v170/fluent/bin/fluent_sysinfo -g -i /users/project/aozuzun/journal.jou (fluent "3ddp -pshmem -host -r17.0.0 -t32 -mpi=pcmpi -path/opt/ansys_inc/v170/fluent -ssh") /opt/ansys_inc/v170/fluent/fluent17.0.0/bin/fluent -r17.0.0 3ddp -pshmem -host -t32 -mpi=pcmpi -path/opt/ansys_inc/v170/fluent -ssh -cx atmaca65:46733:34485 Starting /opt/ansys_inc/v170/fluent/fluent17.0.0/lnamd64/3ddp_host/fluent.17.0.0 host -cx atmaca65:46733:34485 "(list (rpsetvar (QUOTE parallel/function) "fluent 3ddp -flux -node -r17.0.0 -t32 -pshmem -mpi=pcmpi -ssh") (rpsetvar (QUOTE parallel/rhost) "") (rpsetvar (QUOTE parallel/ruser) "") (rpsetvar (QUOTE parallel/nprocs_string) "32") (rpsetvar (QUOTE parallel/auto-spawn?) #t) (rpsetvar (QUOTE parallel/trace-level) 0) (rpsetvar (QUOTE parallel/remote-shell) 1) (rpsetvar (QUOTE parallel/path) "/opt/ansys_inc/v170/fluent") (rpsetvar (QUOTE parallel/hostsfile) "") )" Welcome to ANSYS Fluent Release 17.0 Copyright 2015 ANSYS, Inc.. All Rights Reserved. Unauthorized use, distribution or duplication is prohibited. This product is subject to U.S. laws governing export and re-export. For full Legal Notice, see documentation. Build Time: Dec 02 2015 13:21:15 EST Build Id: 10359 Revision: 867598 -------------------------------------------------------------- This is an academic version of ANSYS FLUENT. Usage of this product license is limited to the terms and conditions specified in your ANSYS license form, additional terms section. -------------------------------------------------------------- Host spawning Node 0 on machine "atmaca65" (unix). /opt/ansys_inc/v170/fluent/fluent17.0.0/bin/fluent -r17.0.0 3ddp -flux -node -t32 -pshmem -mpi=pcmpi -ssh -mport 10.10.10.65:10.10.10.65:42858:0 Starting /opt/ansys_inc/v170/fluent/fluent17.0.0/multiport/mpi/lnamd64/pcmpi/bin/mpirun -e MPI_USE_MALLOPT_MMAP_MAX=0 -np 32 /opt/ansys_inc/v170/fluent/fluent17.0.0/lnamd64/3ddp_node/fluent_mpi.17.0.0 node -mpiw pcmpi -pic shmem -mport 10.10.10.65:10.10.10.65:42858:0 ------------------------------------------------------------------------------- ID Hostname Core O.S. PID Vendor ------------------------------------------------------------------------------- n0-31 atmaca65 32/64 Linux-64 5819-5850 AMD Opteron(TM) 6276 host atmaca65 Linux-64 5648 AMD Opteron(TM) 6276 MPI Option Selected: pcmpi Selected system interconnect: shared-memory ------------------------------------------------------------------------------- Reading journal file /users/project/aozuzun/journal.jou... > /file/read-case "/users/project/aozuzun/FFF.msh" Multicore SMT processors detected. Processor affinity set! Reading "/users/project/aozuzun/FFF.msh"... Buffering for file scan... 9870 nodes, binary. 2047 nodes, binary. 127893 triangular interior faces, zone 1, binary. 1052 triangular velocity-inlet faces, zone 5, binary. 219 triangular pressure-outlet faces, zone 6, binary. 1694 triangular symmetry faces, zone 7, binary. 1092 triangular wall faces, zone 8, binary. 33 triangular wall faces, zone 9, binary. 64969 tetrahedral cells, zone 2, binary. Building... mesh auto partitioning mesh by Metis (fast), distributing mesh parts................................, faces................................, nodes................................, cells................................, materials, interface, domains, zones, wing-tip wing-surface symmetry outlet inlet fluid interior-fluid parallel, Done. > /define/boundary-conditions/velocity-inlet (inlet) inlet no yes yes no 0 yes no 50 no 0 no 0 > /solve/initialize/hyb-initialization Initialize using the hybrid initialization method. Checking case topology... -This case has both inlets & outlets -Pressure information is not available at the boundaries. Case will be initialized with constant pressure iter scalar-0 1 1.000000e+00 2 3.204116e-03 3 6.282962e-04 4 2.183606e-04 5 6.469976e-05 6 2.289022e-05 7 7.921233e-06 8 2.911090e-06 9 1.077119e-06 10 4.187437e-07 hybrid initialization is done. > /solve/iterate 10 iter continuity x-velocity y-velocity z-velocity time/iter 1 1.0000e+00 9.3978e-03 8.4207e-03 8.6167e-03 0:00:01 9 2 4.5170e-01 3.2794e-03 3.2977e-03 3.2374e-03 0:00:00 8 3 2.8198e-01 1.4109e-03 1.5119e-03 1.5059e-03 0:00:00 7 4 2.1502e-01 7.9329e-04 8.4129e-04 8.4017e-04 0:00:00 6 5 1.7309e-01 5.6978e-04 5.5468e-04 5.6431e-04 0:00:00 5 6 1.3927e-01 4.5844e-04 3.9632e-04 4.0327e-04 0:00:00 4 7 1.0968e-01 3.7456e-04 2.9209e-04 2.9621e-04 0:00:00 3 8 8.6301e-02 3.0567e-04 2.2128e-04 2.2177e-04 0:00:00 2 9 6.8001e-02 2.4458e-04 1.7117e-04 1.7187e-04 0:00:00 1 10 5.3917e-02 1.9585e-04 1.3448e-04 1.3641e-04 0:00:00 0 > /report/forces/wall-forces yes 1 0 0 yes "/users/project/aozuzun/force" > Halting due to end of file on input. BEST REGARDS |
|
December 30, 2019, 18:33 |
|
#6 |
Senior Member
Alain Islas
Join Date: Nov 2019
Location: Mexico
Posts: 142
Rep Power: 7 |
By module I meant, to avoid typing the whole ansys installation directory and run it by typing a single word. This doesn't look to be done by your HPC staff. Anyway, Now fluent runs correctly.
This Unbound Variable error may be due to an incomplete setup of your case. Make sure to check all the boundary conditions are well defined. Please note that once Fluent runs in Batch mode, it will depend directly on the journal file set of instructions you are using. This is very sensitive to errors, so it will exit and terminate the job once any error is presented. I suggest you to do the case setup with the GUI activated, so you can edit your case more comfortably. If you are exporting reports, monitors, or calculation activities, make sure you are naming and placing them in the proper formats and directories. Otherwise, fluent will detect error and lead you to a job termination. Greetings |
|
Tags |
fluent, hpc, linux, pbs torque, ssh |
Thread Tools | Search this Thread |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
Stopping a Fluent batch job AND saving the data! Possible? | Volker Pawlik | FLUENT | 13 | December 28, 2020 05:16 |
fluent 12.1, batch mode in linux | sakalido | FLUENT | 13 | September 27, 2017 09:55 |
parallel fluent under pbs not working | kharnabnew | FLUENT | 0 | January 6, 2011 04:28 |
Hints for Job Interview with Fluent? | Andy | Main CFD Forum | 4 | August 15, 2005 10:46 |
Linux: when the job suddenly stops working. | Stevie Wonder | CFX | 3 | August 15, 2004 19:36 |