This workshop is dedicated to students with some programming experience to learn the three parallel programming models MPI, CUDA and OpenMP. It starts on beginners level but also includes some advanced features like the parallelization of a Krylov type solver. Hands-on sessions (in C and Fortran) allow students to immediately test and understand the main building blocks of these parallel programming models.
NTUA School of Chemical Engineering (PC-Lab)
On-Site
Hands-on
July 8 - 11, 2024
June 23th, 2024
The main topics are:
Laboratories and Units from NTUA
Georgios Goumas is an Associate Professor at the School of ECE of the National Technical University of Athens. He graduated from the Dept. of Electrical and Computer Engineering of the National Technical University of Athens (NTUA) (1999). He received a PhD Degree from the School of Electrical and Computer Engineering of NTUA in January 2004. He is currently a senior researcher at the Computing Systems Laboratory (CSLab) in the School of Electrical and Computer Engineering of NTUA. His research interests include high-performance computing and architectures, cloud computing, resource allocation policies, resource-demanding applications, sparse algebra, automatic parallelizing compilers, parallel programming models, etc. He has published more than 80 research papers in journals, international conferences and peer-reviewed workshops. He has worked in several European and National R&D programs in the field of High Performance Computing, Cloud Computing, Networking and Storage for IT systems.
Dr. Antony Spyropoulos is a Laboratory Teaching Stuff at the School of Chemical Engineering of the National Technical University of Athens (NTUA). He received a degree in Chemical Engineering from the NTUA (1996) and a PhD (2003) from the same institution. His research interests are: parallel algorithms for computational mechanics, non-linear analysis (bifurcation, stability) and iterative methods for large scale linear equation systems and eigenvalue problems. He has taught courses in programming languages and numerical methods.
Dr. Xenofon Trompoukis is currently a research officer in the Parallel CFD & Optimization Unit of the Lab of Thermal Turbomachines of the School of Mechanical Engineering of NTUA. He received a degree in Mechanical Engineering from the NTUA (2007) and a PhD (2012) from the same institution. He has participated in several research projects with emphasis on aerodynamic and recently multi-disciplinary gradient-based optimization methods assisted by the adjoint technique while developing and using a GPU-enabled CFD s/w running on HPC systems. His research interests are: parallel programming, GPU programming, CFD, adjoint-based optimization.
| Monday, July 8 | |
|
10:00 - 10:15 |
Welcome |
|
10:15 - 12:00 |
Introduction to parallel computer architectures |
|
12:00 - 12:15 |
Break |
|
12:15 - 14:00 |
Introduction to MPI |
|
14:00 - 14:30 |
Break |
|
14:30 - 16:00 |
Programming with MPI (Hands-on) |
| Tuesday, July 9 | |
|
10:00 - 12:00 |
Programming with MPI (Hands-on) |
|
12:00 - 12:15 |
Break |
|
12:15 - 14:00 |
The SPMD (Single Program Multiple Data) parallel programming model - Application to the finite element method |
|
14:00 - 14:30 |
Break |
|
14:30 - 16:00 |
Introduction to GPUs and their programming through the CUDA environment |
Wednesday, July 10 |
|
10:00 - 12:00 |
Programming with CUDA (Hands-on) |
|
12:00 - 12:15 |
Break |
|
12:15 - 14:00 |
Programming with CUDA (Hands-on) |
|
14:00 - 14:30 |
Break |
|
14:30 - 16:00 |
Programming with OpenMP (Hands-on) |
| Thursday, July 11 | |
|
10:00 - 16:00 |
Final exercise (MPI, CUDA, OpenMP) |
The HPC Workshop: Large Scale Scientific Computations was held at the School of Chemical Engineering of the National Technical University of Athens from July 5 to July 8, 2022.
A total of 20 participants from academic institutions (National Technical University of Athens, University of Ioannina), research centers (National Hellenic Research Foundation) and private companies (Naval DME - Engineering Company) attended this 4-day workshop.
At the last day, 13 participants successfully completed at least one parallel implementation of the final exercise (iterative solution of a linear system). Namely:
Moreover, 1 participant completed both the MPI and OpenMP implementations and 1 participant completed all implementations (MPI, CUDA and OpenMP).
