Jared Brzenski

Jared Brzenski

Institution/Organization: San Diego State University

Department: Computational Science

Academic Status: Doctoral Student

Biography:

My desire for more education has never sated, and I found a program at SDSU/UCI that was full of courses that were geared toward what I already did in my free time. Self-taught in machine learning, app programming, database systems, and a credentialed math teacher, it was like the computational science program was made specifically for me.

I am a Ph.D. student working with Prof. Castillo at San Diego State University on mimetic ocean models. I also work with Prof. Kristen Davis at UC Irvine numerically modeling flexible vegetation. I also collaborate with assorted other faculty on projects from HPC to fluid-structure interactions.

Learning is my favorite activity. My dream job is to learn. My family has been very encouraging of me in following my dream job. I enjoy continuing it through the UCI/SDSU joint doctoral program.

Motivation:

High-performance computing (HPC) has become one of my favorite areas of research. In terms of computer programs and programming, running code in an HPC environment is like racing a Formula 1 car. Everything is sleeker and quicker. Expecting code written on your laptop to work on a supercomputer as well as other HPC optimized code is what most people assume happens. I was one of those. That is not how it works. That is like taking your Honda Accord onto the Formula 1 track and expecting to go faster. I have come to enjoy taking some program or bit of code and optimizing it as much as possible to take full advantage of the HPC system it is running on. I guess it is like tuning regular cars to be race cars. Take parts of code out, and put in better more optimized code. These skills can be learned in a classroom but I feel the best way to learn them is by working with specific codes on specific HPC systems.

The SRP-HPC fellowship allows for this hands-on learning and experience with super-computing systems. Last year I was given code to optimize and ran my code on the CORI system, which at the time was the number 30 supercomputer in the world. There is no way I could have been able to get time on a machine like that, with expert help from my mentors at Lawrence Berkeley National Lab, without the SRP-HPC fellowship. I learned as much in those ten weeks as I did in any semester-long class. I also was given the opportunities to interact with the developers of HPC codes, and conversations with them and their insights cannot be captured by articles or white papers.