معرفی
Sreepathi Pai is an Associate Professor of Computer Science at the Hajim School of Engineering & Applied Sciences, University of Rochester. He holds a PhD from the Indian Institute of Science. His research focuses on experimental computer systems, particularly the performance of computer programs, with emphasis on heterogeneous architectures combining CPUs and GPUs. He develops optimizing compilers for irregular/graph algorithms on GPUs and creates performance models for these systems.
Key research interests include compilers, heterogeneous architectures, GPU algorithms, and performance modeling. His work addresses challenges in large-scale automata processing, graph processing frameworks like Groute, and modular GPU emulation (Horus). He collaborates on projects such as TriForce for multiphysics simulations and explores efficient electromagnetic algorithms for plasma simulations.
Recent projects include formal methods for scientific computing (DANUBE), lightweight GPU performance simulation, and autotuning for volunteer computing. His research also spans statistical caching, compiler-assisted memory management, and parallel triangle counting algorithms.
His work bridges theory and practice, with contributions to frameworks like PLASMA for heterogeneous accelerators and tools for analyzing memory utilization in shared-memory systems. He has published extensively on topics ranging from tensor core modeling to preemptive thread block scheduling in GPUs.
