Acing GPU U: NVIDIA Awards $25,000 Graduate Fellowships to Top PhD Students

by Sylvia Chanak

We’re continuing a tradition of supporting graduate research in GPU computing with this year’s NVIDIA Graduate Fellowship Program.

The NVIDIA Graduate Fellowship Program awards $25,000 to Ph.D. students involved in GPU computing research. The aim is to support graduate students doing outstanding GPU-based work by offering a financial incentive and technical support.

“NVIDIA has a longstanding commitment to supporting outstanding academic research, because we know it fuels innovation,” said Bill Dally chief scientist and senior vice president of research. “We’ve invested millions of dollars in support of ground-breaking research in science, engineering and medicine. We’re delighted to support the work of these exceptional graduate students, whose efforts will help define the future of computing.”

Our Graduate Fellowship award winners were selected from more than a hundred applicants in 21 countries. Their projects involve a variety of technical challenges, including computer architecture, programming models, character animation, computer graphics, and computational methods for simulating chemical events.

This year’s winners:

Sandeep.AgrawalSandeep Agrawal, from Ahmedabad, Gujarat, India

Studying at Duke University

Web service workloads like document search, image search and dynamic web pages exhibit a high degree of similarity among requests, and this makes them a good fit for warp scheduling on modern GPUs. Sandeep’s research aims to design future server platforms using data parallel hardware to achieve maximal throughput per Watt while keeping latency under tolerable bounds.

Benjamin EckartBenjamin Eckart, from Clarksville, TN

Studying at Carnegie Mellon University

Ben’s research focuses on the creation of real-time algorithms for robotic perception using machine learning techniques accelerated by parallel algorithms and novel data structures. For systems using high-density 3D range sensing to generate large amounts of point cloud data (such as time-of-flight cameras and structured light sensing), he is exploring ways to use many-core architectures such as the GPU to rapidly create compact models to facilitate and unify common low-level perceptive tasks like point cloud segmentation, registration, and classification.

Andrew MaimoneAndrew Maimone, from Warsaw, NY

Studying at the University of North Carolina at Chapel Hill

Andrew’s research explores new display technologies that use simple optics and powerful GPU software control to improve performance and add new functionality. He is applying these techniques to develop compact, wide field of view augmented reality eyeglasses and glasses-free 3D displays that support multiple users and additional depth cues.

Soham Uday MehtaSoham Uday Mehta, from Nagpur, India

Studying at the University of California, Berkeley

Soham’s research focuses on using signal processing ideas like frequency analysis to understand the light fields in rendering, and develop faster algorithms for rendering photo-realistic images. Through the use of adaptive sampling and filtering on the GPU, his algorithm can produce accurate images with depth of field and area light direct and illumination at least 30x faster.

Jin WangJin Wang, from Ningbo, China

Studying at the Georgia Institute of Technology

Jin is developing a new execution model for dynamic parallelism found in unstructured applications with irregular data structures such as relational queries and graph analysis. The goal is to pursue both the compiler and micro-architectural support of this model to discover and exploit dynamically formed pockets of structured data parallelism in irregular applications that can use GPU compute and memory bandwidth.

We would also like to acknowledge our five finalists:

  • William Chan, Carnegie Mellon
  • Sundeep Jolly, Massachusetts Institute of Technology
  • Jungsuk Kim, Carnegie Mellon
  • Timothy Rogers, the University of British Colombia
  • Yangzihao Wang, University of California, Davis

The NVIDIA Graduate Fellowship Program is open to applicants worldwide. Eligibility criteria include completion of the first year of Ph.D. level studies in the areas of computer science, computer engineering, system architecture, electrical engineering or a related area. In addition, the student must be a member of an  active research team.

For more information on the NVIDIA Graduate Fellowship Program please visit our website.