This workshop will focus on approaches to porting Matlab applications to a cluster environment such as that of ASU's Agave cluster. This is not an intro to Matlab course. The intended audience member will have developed Matlab code that runs on a desktop machine but now would like to run this code in a parallel environment. This may be implemented through either:
1) Batch submission of multiple single-threaded instances (e.g. parameter sweep)
2) Multithreading m file using "parfor" command
3) Confronting large datasets using distributed arrays or tall arrays
4) Exploiting Matlab functions ported to GPU
5) Multithreading C-code using OpenMP or writing cuda kernels and compiling with mex compiler to be called by Matlab
6) Parallel C code implemented with MPI invoking the Matlab engine on multiple nodes.
All of these approaches will be discussed. In preparation for the workshop all attendees are encouraged to obtain an account on Agave if they do not already have one: https://cores.research.asu.edu/research-computing/get-started/create-an-account
Date: March 23rd, 2:00 - 4:00 PM
Location: Zoom https://asu.zoom.us/j/566944398
For more information about Research Computing, please visit researchcomputing.asu.edu
Questions: Gil Speyer, Lead Scientific Software Engineer, Research Computing, speyer@asu.edu