Report Open Access
Khadidja Hadj Henni
The HPC service at CERN provides linux batch infrastructure to run high performance computing
applications that require MPI clusters.The HPC cluster is therefore dedicated to run MPI programs.
with numerous applications running on the cluster and big number of users demanding resources to
run their jobs,a monitoring tool is needed to measure performances and so this project aims to
integrate a framework that includes intel performance tools in the hpc cluster in order to gather data
for both users and service managers about job execution efficiency and give feedback to users about
how to improve the performance.