Graduate Research Assistant
Habanero Extreme Scale Software Research Laboratory at Georgia Institute of Technology
Atlanta, Georgia, United States
May 2022 - PRESENT
Recently, my research is concentrated on building runtime systems and distributed applications for large-scale computations, focusing on novel communication strategies, efficient data distributions, scalability performance, and SW/HW co-design.
In more detail, I focus on the following:
- Increasing resiliency and performance of the HClib Actor-based runtime system by extending automatic communication termination protocols, distributed graph generation, and multithread execution.
- Building large-scale distributed graph algorithms, including triangle centrality, jaccard index, page rank, pattern matching, genome comparisons, internet network topology analysis, deep learning, and GNN.
- Implementing a distributed and shared-memory parallel Actor-based runtime system for cloud computing, allowing for HPC on the Cloud.
- Optimizing and fine tuning the runtime system using an architecture-aware approach, such as evaluating intra-node core, socket, NUMA, and software-level buffer bindings.
My thesis is Scalable Asynchronous Actor-based Approaches for Distributed-Memory Parallel Applications. I am mentored by Dr. Vivek Sarkar.
Research Interests
Distributed/parallel systems, High Performance Computing (HPC), programming models, deep learning
Publications
🌐 The publications can be retrieved from the publications page. I've published in conferences such as CCGrid, KDD, SC, ISC, and more.