Professor, Student Research Changes How We Use Computers

Publication Date: 10/30/2019

Dr. Nary Subramanian, associate professor of computer science, and Andrew Fobel, bachelor of computer science 2019 graduate, have been working on cutting-edge research the past four years, which has resulted in a patent submission related to the use of graphics processing unit (GPU) programming for fast-search algorithms and data analytics.

Typically, data is analyzed on a standard CPU (central processing unit) which is often called the "brain" of a computer. While the computer's GPU (graphics processing unit) handles processing all of the visuals. 

Dr. Subramanian and Fobel focused on developing a way to utilize the power of the GPU to process information and extract data--instead of the CPU--in hopes to increase the speed and decrease the memory consumption for data search and retrieval.

"There isn't a program that we could use to start searching through data and information on GPUs as we had discussed, so Andrew created a way for this information to be represented and captured on the GPU and also for it to run on the GPU," said Subramanian. "Once he solved those problems, we knew it would work so much faster."

They were able to test their algorithms in the department’s Networking Lab, which has high-performance machines with NVidia’s Tesla GPU card. And when they did there was an eightfold increase in extracting data on the GPU using their program compared to using the CPU, which they believe will have several opportunities for application to real-world, practical problems. 

"By generating benchmarks for when GPU-based processing will outperform CPU-based processing for similar operations, we can help practitioners and researchers know what will be most efficient and cost-effective to use for performing analytics," said Subramanian.

They began the patent process last November after meeting with an attorney in Dallas who, after doing preliminary research on the subject, felt confident their approach to the matter was unique. 

Fobel, who graduated UT Tyler summa cum laude in May 2019, now works at Advanced Micro Devices, Inc., one of the top manufacturers of GPUs.

We will keep the UT Tyler community updated on Dr. Subramanian's research and patent process.


UT Tyler