Francesca Schiavello

Francesca Schiavello

Hey there, I’m Francesca and I am one of the lucky participants of the PRACE summer of HPC, edition 2020. I did my bachelor studies in Applied Mathematics and after working for a few years I decided I wanted to go into Computational studies. I am currently doing a master in such a topic in the University of Amsterdam and University of Vrije, and early on in the master I found out about PRACE. The research projects offered and the opportunity to work and learn in one of the leading supercomputer facilities in Europe was an opportunity I could not pass up.

Several months later here I am, starting my summer adventure in Machine learning for the rescheduling of SLURM jobs in the Hartree Centre in the UK. Despite COVID19, the program moves on and I will be based in Rome whilst working on this project online. I am excited to learn as much as possible and after the first training week in CUDA, openMP and MPI and I am even more eager to do some hands-on work.

Besides my academic interests I am also quite an athletic person, and love to try all sorts of activities. Right now my go to sports are karate and soccer, but I really will try any sport. Fun fact, I am currently trying to learn how to do a handstand. If you have any tips for me drop them in the comments!

Stay tuned for the development of this project. Ömer Faruk and I will be starting off with a review of the literature and studying some of the background theory, before delving into the machine learning applications themselves. The ultimate objective is to properly estimate execution times of user’s submitted jobs to Hartree’s clusters. Given that most users are not submitting optimal or correct estimations of the resources they need, it is our job to see if we can predict these run times more efficiently…for both a happier user and a happier, more efficient running cluster.

Tagged with: , , , ,

Leave a Reply

Your email address will not be published. Required fields are marked *


This site uses Akismet to reduce spam. Learn how your comment data is processed.