High-Performance Computing


Photo of Professor El-Ghazawi


High-Performance Computing Lab (HPCL)

The High-Performance Computing Lab (HPCL) focuses its research on parallel and distributed computing with emphasis on parallel computer architectures, including multicore and heterogeneous processing based on graphical processing units and reconfigurable computing; parallel programming languages using the partitioned global address space (PGAS) paradigm; hardware support for programming and system software; and efficient architectures and scheduling for cloud computing and big data with applications to remote sensing, business intelligence, and bioinformatics. The lab operates through CHREC (the Center for High-Performance Reconfigurable Computing), with joint funding from NSF, industry and government; and IMPACT (the Institute for Massively Parallel Applications and Computing Technologies), a university-wide chartered institute. Learn more about Professor El-Ghazawi.


Photo of Elias Balaras


Laboratory for Computational Physics and Fluid Mechanics

This lab aims to develop high-fidelity modeling tools applicable to multiphysics/multiscsale flow problems in physical and biological systems. Central to its work is the synergy of mathematical modeling and computational algorithms coupled to the remarkable advancements in computer hardware technology, which enable us to simulate a wide variety of natural phenomena. The current thrusts are fluid-structure interactions, multiphase turbulent flows, and multiscale modeling of the blood circulation and related biomedical devices. An example problem in the latter category is the two-way interaction between the macroscopic flow patterns and blood elements, which represents the link between fluid mechanics and clinical applications. These computations involve the interactions of fluid flow with millions of deformable particles, which are only possible on the latest petascale supercomputing platforms. Learn more about Professor Balaras.


Photo of Professor Sarkar


Computational Biomechanics Lab

This lab, led by Professor Kausik Sarkar of the Department of Mechanical and Aerospace Engineering, simulates complex, multiphase flows of emulsions and suspensions of drops, particles, biological cells and vesicles. The lab's current projects are: 1) investigation of the fundamental physics of viscoelastic multiphase flows and their emulsion rheology for polymer processing; 2) rheology of blood and blood cell adhesion to vessel walls underlying atherosclerosis and inflammatory diseases; and 3) rheology of dense microbubble foams for oxygen delivery. The lab's projects are funded by the National Science Foundation and the National Institutes of Health. Learn more about Professor Sarkar.


Photo of Professor Howie Huang


Computer and Systems Architecture

Professor Howie Huang and his students conduct advanced computer systems and architecture research in large-scale computer systems from supercomputers to data centers, and develop high-performance, scalable, and reliable computer systems for data-intensive applications. Current projects include a middleware for managing data-intensive applications in virtualized data centers, the development of energy-efficient computer systems, high-performance storage systems, GPU-accelerated scientific simulations, and providing reliability as a service in the Cloud. Learn more about Professor Huang.


Photo of Professor Guru Prasadh Venkataramni


Computer Architecture

Professor Guru Prasadh Venkataramani leads the GW CompArch (Computer Architecture research) group. His research group works on many-core computing, architectural support for debugging and programmability, emerging memory technologies, and performance/power analysis of applications. Dr. Venkataramani's research is supported by a National Science Foundation (NSF) Faculty Early CAREER Award (2012), and through grants from NSF's Division of Computer and Communication Foundations and a 2010 ORAU Ralph E. Powe Junior Faculty Enhancement Award. The GW CompArch group also won a Best Poster Award at the IEEE International Conference on Parallel Architectures and Compilation Techniques (PACT 2011). Learn more about Professor Venkataramani.


Lorena Barba


Barba group

Professor Lorena A. Barba and her research group apply high-performance and heterogeneous computing and develop algorithms and software in research settings involving fluid dynamics and biophysics. She was an early adopter of GPU hardware for scientific computing and was recognized as CUDA Fellow in 2012 by Nvidia Corp. Her group has developed high-performance implementations of the fast multipole method, which are being used in boundary element solutions of bioelectrostatics problems. They also apply immersed boundary methods in the simulation and analysis of gliding animals. Professor Barba and her team have a focus on open-source software and open science, high-quality software engineering practices for science, and scientific reproducibility and replicability. Learn more about Professor Barba.