Sudip Seal

Sudip Seal is a Senior Scientist in the Computer Science and Mathematics Division, Oak Ridge National Laboratory. He is also a Joint ORNL-UT Associate Professor in the Department of Electrical Engineering and Computer Science, University of Tennessee, Knoxvile. He holds a PhD in computer engineering and a PhD in theoretical high energy physics. He specializes in the design and development of HPC solutions for large-scale science-based computational problems and, in that capacity, is actively involved in the fusion, neutron and materials sciences at present.

Sudip leads the DAPPER (Data Analysis Parallel Package Maker) project. DAPPER is the prototype of a highly-scalable and versatile application-independent parallel framework targeted toward the rapid development of scalable application-specific data analysis software. As part of the DAPPER project, he is leading the development of the world's first large-scale distributed memory framework for at-scale study and analysis of massive atom probe tomography data with hundreds of millions to billions of atoms collected by the state-of-the-art Local Electrode Atom Probe (LEAP) microscope.

Sudip is the lead architect of PARVMEC (Parallel Variational Moments Equilibrium Code), the fastest 3D plasma equilibrium solver in the world. PARVMEC enabled high-speed high-resolution plasma equilibria reconstructions that led to the first confirmation of a helical core in the DIII-D tokamak at the National Fusion Facility, San Diego, USA in 2017. Sudip also developed the distributed-memory parallel version of SIESTA (Scalable Iterative Equilibrium Solver for Toroidal Applications) which is over 30 times faster than its predecessor and is used to study plasma equilibira with magnetic islands.

Sudip is leading the development of PARADYSE (Parallel Debye Summation Engine). PARADYSE, which has already been demonstrated to deliver over three orders of magnitude in speedup, is scoped to replace the current Debye summation engine used for structure refinement by all users of the SNS (ORNL).