BAD TERMINATION OF ONE OF YOUR APPLICATION PROCESSES

[JoseLozano]

Hello dear community members,

I would like to ask if you can help me out to solve an issue that I've been facing for the last two weeksl. I'm not well versed in the world of Intel Fortran Compilers. I'm trying to run a test from a cosmological code called CosmoMC (https://cosmologist.info/cosmomc/readme.html).

I want to use the Planck Likelihood Code (PLC) where all the instructions are explained in https://cosmologist.info/cosmomc/readme_planck.html . After installing CosmoMC and PLC succesfully without any errors and warning since I followed all the instructions, I tried to run a test. However, as it is written at the end  of the section "Using the Planck likelihood with CosmoMC" (https://cosmologist.info/cosmomc/readme_planck.html) I typed 

mpirun -np 1 ./cosmomc test_planck.ini

to check if everything works fine. However, I got the following output:

Number of MPI processes: 1
file_root:test
Random seeds: 22226, 21161 rand_inst: 1
Using clik with likelihood file
./data/clik_14.0/hi_l/plik/plik_rd12_HM_v22b_TTTEEE.clik
corrupted size vs. prev_size

===================================================================================
= BAD TERMINATION OF ONE OF YOUR APPLICATION PROCESSES
= RANK 0 PID 3760 RUNNING AT Heisenberg
= KILLED BY SIGNAL: 6 (Aborted)
===================================================================================

Although the installation process was very smooth, it fails when I try to run the test_planck.ini

Could be that the Compiler is not working properly or something is off with Planck Likelihood Code ? 

I hope you can give me a hand with this issue.

I'm looking forward to your kind reply

[mecej4]

I am afraid that only someone who uses and is familiar with the large code in question (CosmoMC) may be able to attempt to answer your question. I doubt that there are any such people who are also members of this forum. Please seek help in a CosmoMC forum.

[TobiasK]

@JoseLozano 
you may want to try to compile without MPI and see if the error stills persists. If it persists I would advice you to add -check all,nouninit to the Fortran flags and see if that catches some error.

If the error disappears with MPI disabled at compile time, try to check if the MPI installation works. 

[JohnNichols]

 From the COSMOS notes

Some people have reported problems using Intel MPI 2019, earlier versions should be OK.

[jimdempseyatthecove]

In the same terminal window where

mpirun -np 1 ./cosmomc test_planck.ini

produces the error, what happens when you issue:

./cosmomc test_planck.ini

If this works, then is it possible that you are running a non-intel mpirun? iow forgot to soruce the setvars.sh from the Intel installation folder.

Jim Dempsey