HPC Summer Workshop: MATLAB Parallel Computing Hands-On Using Wulver¶
Overview¶
Join us for an interactive webinar hosted at NJIT's HPC facility, where MATLAB expert (Evan Cosgrov) will guide participants through practical techniques for accelerating code and workflows using MATLAB’s parallel computing tools. Through live demonstrations and guided examples, you’ll gain a solid understanding of how to parallelize MATLAB code, overcome common challenges, and optimize performance across distributed computing environments.
Each participant will have access to the OnDemand Matlab server running on Wulver.
Guide to access Matlab via OnDemand:¶
Users need use Jupyter Matlab Proxy
to use MATLAB on OnDemand. Check the tutorial here.
Date and Location:¶
- Date: June 12th 2025
- Location: Campus Center 235
- Time: 1:00 PM - 4.00 PM
Topics Covered¶
- How to identify bottlenecks in serial MATLAB code and convert them to run in parallel.
- Practical differences between
parfor
andparfeval
, and how to choose the right one. - Creating asynchronous tasks and managing outputs with DataQueue.
- Running Simulink models in parallel and accelerating simulation tasks.
- Leveraging GPU resources to boost performance in compute-intensive operations.
- Working with large datasets using tall arrays and distributed arrays.
- Best practices for launching and managing MATLAB parallel pools.