SLURM

1. [c]

   On cumulus, create a subdirectory ~/PDC/hw4, and copy pub/cs300/hw/trap.cpp in that homework subdirectory. Also create a file trapjob.sh with the following contents.

   #!/bin/sh
   
   mpirun --map-by node /home/rab/PDC/F16/trap
   

   Compile trap.cpp with mpiCC, naming the executabe trap . Then launch an MPI job on SLURM using

   cumulus$  sbatch -n 9 -o trapjob.out trapjob.sh
   

   You can use the squeue command and the job number printed by sbatch to determine when your job has finished (should be quick). Then examine the results in trapjob.out.

   Make a commit of your code:

   cumulus$  git add trap.cpp trapjob.sh trapjob.out 
   cumulus$  git commit -m "hw4: SLURM submission of an MPI job"
   

2. [c]

   Experiment with your SLURM job in the previous exercise, for example, using options identified in the cluster job management reading. (Note: avoid launching large jobs in class, since that will slow everyone's work! use scancel if necessary).

   Use commit to record the source code for your results.

Hadoop introduction

1. [c] Carry out steps 1-3 of the Hadoop intro lab on cumulus. Locate your work in the subdirectory of ~/PDC directed in the lab.

Submitting this homework

All of your code for this assignment should already be contained in commits. Modify the most recent commit message to indicate that you are submitting the completed assignment.

cumulus$  git commit --amend

Add the following to the latest commit message.

    hw4: complete

If your assignment is not complete, indicate your progress instead, e.g.,

    hw4: items 1-5 complete, item 6 partial

You can make later commits to submit updates.

Finally, pull/push your commits in the usual way.

cumulus$  git pull origin master
cumulus$  git push origin master

---

Files: trap.cpp trapjob.sh trapjob.out

Also, files from A2 and the Hadoop lab