The SCI Institute

Profile of Sowmya Yellapragada

Sowmya Yellapragada - Graduate Research Assistant

WEB 2750

Supervisor: Dr Manish Parashar

Background

I am a Ph.D. student in Computer Science at the University of Utah, advised by Prof. Manish Parashar at SCI. I earned my master’s degree in computer science from the University of Illinois Urbana-Champaign, where I worked with Prof. Laxmikant (Sanjay) Kale in the Parallel Programming Laboratory. My master’s thesis, Analyzing Communication Impacts of Graph-Refinement Load-Balancing, focused on evaluating communication overheads in Charm++ diffusion algorithms using large-scale Stencil3D simulations.

Before joining Utah, I interned at Lawrence Berkeley National Laboratory (NERSC), where I developed heterogeneity-aware load-balancing algorithms for AMReX-based applications, integrating CPU-GPU performance weighting and exploring space-filling curves. I also hold a B.Tech. in Computer Science from Guru Gobind Singh Indraprastha University, where my thesis High-Performance Image Convolution with Parallel Processing explored OpenMP-based acceleration techniques.

Current Responsibilities

  • Conducting research on data management and in-situ coupling frameworks for large-scale scientific simulations, focusing on scalability, resilience, and deployment automation.
  • Researching data management strategies for large-scale scientific workflows.
  • Analyzing communication costs, node locality, and performance scalability on heterogeneous HPC architectures.
  • Developing containerized and distributed runtime environments for scientific workflows, including efforts to orchestrate complex data services on Kubernetes clusters for dynamic, multi-component HPC applications.
  • Extending the prior work in load balancing and communication-aware runtime systems to explore efficient data movement and coordination across heterogeneous compute environments, and performance modeling for next-generation exascale infrastructures.

Research Interests

  • High-Performance Computing
  • Parallel Programming and Load Balancing
  • Large Scale Data Management
  • Heterogeneous and Distributed Systems
  • Performance Optimization and Modeling