Finding pMatlab Output
When you run a pMatlab job, the standard output of each process is captured in out
files. The .out
filenames include a number indicating the process id that created the file: <pMatlab_script_name>.<pid>.out
where:
<Matlab_script_name>
is the filename of your pMatlab script<pid>
is the id of the process
Depending on how you submit your job the location of the files will be in either of these 2 locations:
- The
./MatMPI
directory (which is created in your working directory) - In subdirectories within the
MatMPI
directory:./MatMPI/p<start-pid>-p<end-pid>_<compute node name>
where:p<start-pid>
is the id of the first process running on the compute node whose name is<compute node name>
p<end-pid>
is the id of the last process running on the compute node whose name is<compute node name>
- If your job ran on multiple compute nodes, the log files will be spread across multiple subdirectories, each one containing the compute node name in the subdirectory's name.
The name of the output files will be <pMatlab_script_name>.<pid>.out
.
Where will the .out files reside?
Default location
In most cases, the .out
files reside in the subdirectories that
include the process id and compute node name within
the MatMPI
directory:
./MatMPI/p<start-pid>-p<end-pid>_<compute node name>/<pMatlab_script_name>.<pid>.out
For example, this is the path and file name for a job named param_sweep_parallel_v2 that ran on 4 processes on compute node b-11-16-3:
./MatMPI/p0-p3_b-11-16-3/param_sweep_parallel_v2.1.out
If you set GRIDMATLAB_MANYCORE to 'NO'
If you set the GRIDMATLAB_MANYCORE
environment variable to 'no' (the
default setting is 'yes'), your output files will be located in
the MatMPI
directory:
./MatMPI/<pMatlab_script_name>.<pid>.out
For example:
./MatMPI/param_sweep_parallel_v2.1.out
The other files in ./MatMPI
pRUN_Parallel_Wrapper.err
The file pRUN_Parallel_Wrapper.err will be created in the MatMPI
directory. This file contains all standard error output from your
pMatlab job. Each line in the file is prefixed with the Pid where the
error occurred which you can use to identify the .out
file and
investigate further. You'll find the same error message in the
associated .out
file.
Unix_Commands*.sh
The Unix_Commands*.sh files are the shell scripts that were used to launch your pMatlab job. This may be useful to the SuperCloud team to help diagnose any problems you may run into when you launch your job or during runtime.
If you set the GRIDMATLAB_MANYCORE
environment variable to 'no' (the
default setting is 'yes'), gridMatlab generates one
Unix_Commands.GRIDMATLAB_KEEP_TEMP_FILES
environment variable as shown below: