Faces of HPC: Sunita Chandrasekaran
Sunita Chandrasekaran works at the University of Delaware, Newark, Delaware, USA as an Assistant Professor in the Department of Computer Science and Information Sciences. Her work focuses on high-level programming models and looking at these models can help migrate legacy code to current and future HPC platforms.
Sunita is originally from Chennai, India, where she studied for both a Bachelors in Electrical and Electronics Engineering. She moved to Singapore to complete a PhD using Field Programmable Gate Arrays (FPGAs) and uses the knowledge from both her undergraduate and postgraduate in her current work in High Performance computing. Sunita is also an active member of the Women in HPC network.
Tell us a bit about yourself.
I work at the University of Delaware as an Assistant Professor in the Department of Computer and Information Sciences. I am from Chennai, India. I did my Bachelors in Electrical and Electronics Engineering from SRM Easwari, Anna University, Chennai, India. My Ph.D. is from Nanyang Technological University, Singapore from the Department of Computer Science Engineering where I used Field Programmable Gate Arrays (FPGAs) and my research was exploring feasible ways to program this device. I then joined University of Houston, Texas, USA to work as a Postdoctoral researcher.
Outside of work, I love to paint, travel and cook. Painting helps me streamline my thoughts, cooking helps me with creative thinking and traveling helps me refresh my mind!
What is your current job?
I work at the University of Delaware as an Assistant Professor in the Department of Computer and Information Sciences. My current research interests include exploring high-level programming models and their suitability for legacy code migration to current and future platforms. I explore applications and their characteristics for different kinds of platforms. This became my main interest after spending quite a number of years working on hardware platforms. My undergraduate degree was in Electrical and Electronics, which involved learning about AC/DC motors, electronic circuits etc.. My PhD was on embedded devices and I spent quite an amount of time in exploring its design space. I then moved on to explore larger computing systems such as multicore/manycore, supercomputers and ways to program these devices and obtain effective performance. The initial years spent on smaller systems definitely helped to think big in terms of HPC.
How did you become interested in HPC?
While talking to application scientists belonging to various domains such as bioinformatics, climate modelling, I observed that they were spending weeks and months using desktop computers in order to successfully execute their scientific computations. Some of these computations have direct impact on everyday life. For instance with some applications, if the computations were to run faster, it would improve the chance of detecting cancer in its earlier stages and as we all know cancer at their early stages are the most treatable. In another instance, using HPC enables faster and better simulations for climate modelling and predictions leading to faster deployment of evacuation plans. Observing such real-life scenario and how HPC can make a difference naturally got me interested in it and ever since I realized the impact, I have been exploring algorithms belonging to real-world applications and expediting computations using machines with massive power and compute capability.
Is there something about you that’s given you a unique or creative approach to what you do?
I think professionally, extensive knowledge on both hardware and software has certainly helped me approach scientific problems from both the angles and be able to strike a balance when need.
Personally, I try to carry a positive attitude towards anything and everything. The word ‘no’ does not exist in my dictionary since there are workarounds to everything just like there is always a workaround to a buggy code!
Were there any challenges when you first entered the field? How have you overcome these, or do they continue to challenge you?
I wouldn’t call it a challenge per se, but I spent quite some time to understand who would benefit if I make an algorithm go 10x faster. Since I do not have a background in science in general (biology, physics, chemistry, astrophysics), it takes me a while wearing my computer science hat on, before I am able to put things into perspective. I have been able to overcome this to an extent by interacting with domain scientists, attending relevant talks and communicating with a lot of application-specific scientists. I think I still have a long way to go but having built a huge network of contacts, I at least now know which door to knock.
What’s the best thing about working in HPC?
Users or application scientists have legacy code written 20-40 years old for computers that existed back then. These codes are related to drug discovery, carbon sequestration, using alternate fuels, chemical evolution of galaxy and so on. HPC has the power and the capacity to migrate these codes to today’s and future platforms and help improve our lifestyle.
If there’s one thing about HPC you could change, what would it be?
A fundamental question to the HPC space from my end would be – do you want to build hardware to compete between other hardware vendors or do you want to build a hardware that the application scientists could really use? Similarly a suggestion to the software developers would be – before you go about creating yet another new programming language, take a step back and study the legacy code. Talk to the application scientists and find out what kind of language they are comfortable with instead of creating a new language and giving no choice to the application scientists but forcing them to use your language just because you created it.
Secondly I would “open source” all HPC publications (for that matter not just HPC but all publications) to make science more reachable to anybody.
What’s next for you in HPC – where does your career lead you?
HPC is growing, exponentially. My vision is to stay ahead of the curve, enable advancements in science and make the world a better place to live in. I am not a doctor in medicine or cancer research but if my research on computational science and bioinformatics helped cure or treat cancer, I would consider my career to have taken the right path.