The registration to this course will open in March. Please, bring your own laptop. All the PATC courses at BSC are free of charge.
Course convener: Filippo Mantovani
Objectives:
- Get a comprehensive view of the architecture of the heterogeneous ARM-based platforms deployed in the framework of the Mont-Blanc project, including ARM multicore clusters based on mobile and server technology.
- Understand the challenges associated with large-scale HPC simulations.
- Understand the basic concepts of architectural simulations and introduce the MUlti-scale Simulation Architecture (MUSA), developed within Mont-Blanc 3 for the simulation of next-generation HPC architectures.
- Hands-on and study of a use case using MUSA. Tools for architectural simulation that include a complete software stack for performing architectural simulations at scale of thousands of cores will be made available.
Learning Outcomes: Understand how ARM multicore and ARM+GPU clusters work. Work with some example codes, compiling, debugging and executing trough system queues.
Get a global view of the machines with their specific configuration (Operating system, batch system, debuggers, compilers, how to access to the machines, how to use the resources, etc)
Target Group: Standard HPC users, some knowledge of multicore and GPU-accelerated systems is desirable, but not required.
Level: INTERMEDIATE: for participants with some theoretical and practical knowledge (All courses are designed for specialists with at least 1st cycle degree)
Requirements: Standard HPC users, with some knowledge in ARM CPU/GPU architecture (recommended, but not required).
Agenda:
Day 1 / 9:30am - 5:30pm: (3h lectures with 5 min breaks on the hour)
Session 1 (2h):
- The Mont-Blanc Project
- ARM based machines: Architecture Overview
- Discussion
Session 2 (1h + 1h hands-on):
- Simulating ARM based machines (TaskSim and Gem5) (1:30h)
- Lunch Break (13:00 to 14:00)
Session 3 (1h + 1h hands-on ):
Simulation at scale: the MUltiscale Simulation Architecture (MUSA), performance and power evaluation of next generation large ARM based systems for scientific computing
END of COURSE
