Issues with running tests (Peridigm 1.5)

[ricomorasata]

@djlittl @johnfoster
Dear Developers,
I am trying to run Peridigm version 1.5 on my machine (running on Ubuntu 16.04).

Running "make -j 4" after the configuration does not result in any error. There seems to be no issue with "make install" as well, though there is one detail I find quite curious:


Is this normal? Otherwise, what could be the problem?

However, upon running "make test", it seems that all multi-processor tests fail.
LastTestsFailed.log
Please find attached the log file in one of these failed tests' directory:
NOX_QS_Newton_NoPrec.log
The requested .e file is not in the directory:

I would be most grateful if you could give me some insight into these issues and potential solutions.
Thank you for your time and consideration.
Looking forward to your feedback.

[johntfoster]

Looks like a problem with your MPI distribution. Did you build the MPI libraries from source?

On Wed, Jul 3, 2019 at 2:35 PM AeroEngProgrammer24 ***@***.***> wrote: @djlittl <https://github.com/djlittl> @johnfoster <https://github.com/johnfoster> Dear Developers, I am trying to run Peridigm version 1.5 on my machine (running on Ubuntu 16.04). Running "make -j 4" after the configuration does not result in any error. There seems to be no issue with "make install" as well, though there is one detail I find quite curious: [image: make_install_1] <https://user-images.githubusercontent.com/37848373/60618963-3b061400-9de0-11e9-9acd-51f76e99def8.png> [image: make_install_2] <https://user-images.githubusercontent.com/37848373/60618964-3b9eaa80-9de0-11e9-9234-5680814f57b9.png> Is this normal? Otherwise, what could be the problem? However, upon running "make test", it seems that all multi-processor tests fail. LastTestsFailed.log <https://github.com/peridigm/peridigm/files/3356561/LastTestsFailed.log> I check one of these failed tests' directory and please find attached the log file: NOX_QS_Newton_NoPrec.log <https://github.com/peridigm/peridigm/files/3356570/NOX_QS_Newton_NoPrec.log> The requested .e file is not in the directory: [image: NOX_QS_Newton_NoPrec_np4_dir] <https://user-images.githubusercontent.com/37848373/60619680-e5cb0200-9de1-11e9-97a5-61a380e6d419.png> I would be most grateful if you could give me some insight into these issues and potential solutions. Thank you for your time and consideration. Looking forward to your feedback. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#100?email_source=notifications&email_token=AAEQR3NYXUABMBXT2TUWIKTP5T5O7A5CNFSM4H5I65U2YY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4G5G3N7Q>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAEQR3MMHMAYG7BUILLB3SDP5T5O7ANCNFSM4H5I65UQ> . This message is from an external sender. Learn more about why this matters. <https://ut.service-now.com/sp?id=kb_article&number=KB0011401>

-- Sent from iPhone

[ricomorasata]

@johntfoster Dear Dr. Foster,
Thank you for your response.
Yes, I installed OpenMPI from source, and all the other dependencies as well.
How should I proceed?
Do I need to reinstall everything?
Thank you in advance

[Hostudent]

did you set environmental variables properly?
sth like this:

export CC=gcc
export CXX=g++
export FC=mpif90
export F77=mpif77
export PATH="/usr/local/Mybulds/openmpi/bin:$PATH"
export LIBRARY_PATH="/usr/local/Mybuilds/openmpi/lib/:${LIBRARY_PATH}"
export LD_LIBRARY_PATH="/usr/local/Mybuilds/openmpi/lib/:${LD_LIBRARY_PATH}"
export DYLD_LIBRARY_PATH="/usr/local/Mybuilds/openmpi/lib/:${DYLD_LIBRARY_PATH}"

[ricomorasata]

Yes,
My environment variables are set like that.

[ricomorasata]

Dear all,

I think I have managed to install Peridigm. The versions of the dependencies are as follows:
Boost 1.70.0, HDF5 1.8.20, NetCDF 4.4.0, Trilinos 12.12.1 (configured with yaml-cpp), Peridigm 1.5.0.
So I tried to run one of the examples, "tensile_test".
The input mesh was successfully partitioned; however, upon running the solver, I think there might be an issue so I would like to ask for some clarifications.
I ran the program with the command: "mpirun -np 4 /path/to/Peridigm ./tensile_test.yaml"
Please find attached a snippet of the iteration process.

Is this normal? If not, what could be the issue and how could I solve it?

Once the iterations are complete, there is only a single output file "tensile_test.e" generated.

In the User's Manual, it is indicated that the MergeFiles.py script is to be used to merge the output files.
Is this a feature of the current version? Otherwise, what could cause the problem and what are potential workarounds?

Thank you for your time and consideration.

[Coooooco]

Dear author,
There is a same mistake happened in my Peridigm 1.5 while I was running "make test". I installed the all dependencies from source, and set environmental variables properly.BUT all multi-processor tests fail.
Could you please tell me how you solve this problem? I would appreciate it very much.
Looking forward to your reply.

[github-actions]

This issue is stale because it has been open 30 days with no activity. Remove stale label or comment or this will be closed in 5 days.

[github-actions]

This issue was closed because it has been stalled for 5 days with no activity.