ABINIT Discussion Forums

Hi everyone,

Now that the old tests script has been removed, I would like to use the new python script.

I build abinit in a directory:
[ABINIT]/build

Then, I compile and at the end I go to the directory tests, as usual:
cd [ABINIT]/build/tests

Then I launch a test using the new python script (following Matteo's presentation in the ABINIT workshop):
[ABINIT]/tests/runtests.py v1[1] -v

This is the output I get:

Running on u-adidas -- system Linux -- ncpus 4 -- Python 2.7.2 -- runtests.py-0.4
mpi_nprocs 1, omp_nthreads 0, py_nthreads 1
('Initalizing JobRunner for sequential runs.',)
Loading database from file: /media/DISK/7.5.1-private/tests/test_suite.cpkl
Test_suite directory already exists! Old files will be removed
Running ntests = 1, MPI_nprocs = 1, py_nthreads = 1...
Command
returned exit_code: 1

[v1][t01] fldiff.pl fatal error:
Unable to open input file /media/DISK/7.5.1-private/build2/tests/Test_suite/v1_t01/t01.out



Indeed, t01.out does not exist.
However, if I go to /Test_suite/v1_t01 and there I launch the command :
/media/DISK/7.5.1-private/build2/src/98_main/abinit < /media/DISK/7.5.1-private/build2/tests/Test_suite/v1_t01/t01.stdin > /media/DISK/7.5.1-private/build2/tests/Test_suite/v1_t01/t01.stdout 2> /media/DISK/7.5.1-private/build2/tests/Test_suite/v1_t01/t01.stderr

It works fine and it produces t01.out

What am I doing wrong? How should I run the tests?

Best wishes,
Tonatiuh Rangel

Hi Tonatiuh,

[v1][t01] fldiff.pl fatal error:
Unable to open input file /media/DISK/7.5.1-private/build2/tests/Test_suite/v1_t01/t01.out

There should be an error message in /media/DISK/7.5.1-private/build2/tests/Test_suite/v1_t01/t01.stderr that
should explain why the subprocess failed.

Matteo

Hi Matteo,

I see in stdout:
Commande timeout inaccessible. Essayer cea_timeout

so there is something missing about timing
Do you know how to fix this?

I already tried to give a value for timeout:
../..//tests/runtests.py v1[1] -v --timeout=1000
But it did not work.



Best
Tonatiuh

By default, the script executes the tests under the control of the linux tool timeout (if found)
so that we kill the job if the run gets stuck. It seems that the version of timeout installed in your machine
is not working properly. To bypass this problem you can use the option -t [--timeout] that allows the user to define
the timeout value in seconds.
Setting --timeout to zero is equivalent to disabling the use of the timeout tool.
Try with:

runtests.py -t 0 [other_options]

and let me know if it works.

Cheers,
Matteo

Thanks, now it works!