some questions about the log file

[Eliot Bush]

Dear Benoit,

Thanks for developing this tool. It seems promising for my applications.

I've been running it, and have a few questions / things in the log that are bothering me.

I'm running it using mpiexec with 60 cores. The data set has 4 species about about 2500 families. I always get the following error at the end of my runs:

===================================================================================
=   BAD TERMINATION OF ONE OF YOUR APPLICATION PROCESSES
=   PID 17020 RUNNING AT siva
=   EXIT CODE: 134
=   CLEANING UP REMAINING PROCESSES
=   YOU CAN IGNORE THE BELOW CLEANUP MESSAGES
===================================================================================
YOUR APPLICATION TERMINATED WITH THE EXIT STRING: Aborted (signal 6)
This typically refers to a problem with your application.
Please see the FAQ page for debugging suggestions
mpiexec -np 60 /usr/local/GeneRax/build/bin/generax -f geneRaxInputList.txt


As far as I've been able to determine so far, it seems all my families end up with gene trees. So not sure what this error is about.

Also, here's what the top of the log looks like:

GeneRax was called as follow:
/usr/local/GeneRax/build/bin/generax -f geneRaxInputList.txt -s ../example.tre -r UndatedDTL --max-spr-radius 3 -p tempOut.6.11

General information:
- Output prefix: tempOut.6.11
- Families information: geneRaxInputList.txt
- Species tree: ../example.tre
- You are running GeneRax without MPI (no parallelization)

I'm not sure why it's telling me generax was run without MPI. It did run with MPI, I observed it using 60 cores.

I also got a number of messages like this:

Error in family family_1102: Cannot read alignment file (file exists but is invalid)

which again seem not to be true. family_1102 exists in the results and has a gene tree in it at the end.

Anyway, any light you can shed on this would be helpful. Things seem to be working, but given the log is telling me otherwise I thought I'd check.

thanks!
Eliot

[Benoit Morel]

Dear Eliot,

I am on vacation until Monday and can only try to give you a quick answer. The only explanation I can see for the "no parallelization" message is that GeneRax was compiled without MPI (because cmake didn't not find the MPI libraries when GeneRax was installed). Running GeneRax with mpiexec would then be equivalent to running it sequentially but 60 times simultaneously, and could cause a lot of weird behaviors.

Please try to reinstall GeneRax, and check the cmake messages (whether it finds MPI or not).

Let me know if this helps,

Benoit

[Eliot Bush]

Dear Benoit,

You were right. It must not have found the MPI libraries the first time I compiled it. I've redone it, and now it does say it's using MPI, and it goes much much faster. (and no error messages).

Thanks for the quick reply!

Eliot