Dr. Bharathwaj Muthuswamy

Visiting Assitant Professor, CS department
The College of New Jersey

February 20, 2018 (Tues)
12:30 – 1:30pm
Ed 115

CUDA and OpenCL:  Programming GPUs

Abstract:

In this talk, Dr. Muthuswamy will go through a high level overview of parallel programming, specifically using (NVIDIA’s) Compute Unified Device Architecture (CUDA) and Khronos Group’s OpenCL (Open Computing Language), for programming Graphics Processing Units (GPUs).  We will show examples of parallel programming in practice, starting from the classic “hello, world” to nonlinear dynamics simulations.  We will start with CUDA but primarily focus on OpenCL, since the specification is hardware independent.

Speaker Bio:

Dr. Bharathwaj “Bharath” Muthuswamy is currently a visiting professor of computer science at The College of New Jersey (TCNJ).  Prior (2015 – 2017) to TCNJ, he was a software engineer for Tarana Wireless, working on implementing real-time network layer code for non-line of sight wireless backhaul devices.  Prior (2000 – 2015) to his Tarana Wireless position, Dr. Muthuswamy has worked in both academia (University of California, Berkeley; Milwaukee School of Engineering) and industry (Los Alamos National Laboratory, SUN Microsystems, National Semiconductor and National Instruments). His primary research interests are in nonlinear dynamical (chaotic) systems and embedded (FPGA systems).  He is also very passionate about undergraduate, high school and middle school education.  He holds a BS (2002), MS (2005) and PhD (2009) all in EECS from the University of California, Berkeley, advised by Dr. Leon O. Chua.  His primary PhD contribution was in the design and implementation of a “Muthuswamy-Chua” circuit (system) – the simplest possible chaotic circuit involving only the fundamental circuit elements – an inductor, capacitor and memristor in series.