Wayne State University

AIM HIGHER

Wayne State University

High Performance Computing Services

Running a Tensorflow Job

HOW TO RUN A JOB FOR TENSORFLOW

Follow these steps to run a Tensorflow job.

NOTE: Make sure you have access to nodes with GPUs

Log on to the Grid.

Copy the required contents using the following command

cp -R /wsu/apps/scripts/tutorial/addition.py .

cp -R /wsu/apps/scripts/tutorial/tensorflow_job .

The file that has jobscript is tensorflow_job. It contains the following script :

#!/bin/bash
#PBS -N Tensorflow
#PBS -l select=1:ngpus=1 -q egacq
#PBS -m ea
#PBS -V

module load gnu-4.8.2/tensorflow-20161005
python addition.py

Make sure that addition.py in in your home directory.

Press the Escape key and type :wq to save and quit.

To submit the job type:

qsub tensorflow_job

The job is submitted. You can check to see what node it is running on with the following command.

qme

Input: qme of Tensorflow

The red boxes show the job ID, the pink one shows the name of the script and the blue one shows the node name on which the job is running.

The output can be found in the Tensorflow.oJOB_ID .

Computer Center Services * 5925 Woodward Avenue Detroit, Michigan 48202