Industry, Research and User Symposium HiPC 2014 Speaker Profiles-Talks

(Material for the following profiles was compiled from various sources. Apologies to the speakers for any errors or information not accurately listed. Please contact HiPC organizers for any corrections to this page.)

IRUS1: Computational Atmospheric and Climate Science in HPC: Methods being developed to tackle the performance issues

Prof. Ravi Nanjundiah, (Head of Department, Centre for Atmospheric Sciences, IISC, Bangalore)

Profile:

Ravi S Nanjundiah is a professor at Centre for Atmospheric and Oceanic Sciences (CAOS), Indian Institute of Science (IISc), Bengaluru. He is currently the chairman of CAOS. He did his Masters in Mechanical Engineering and PhD in Atmospheric Science from IISc. He did his post-doctoral fellowship at the Mathematics and Computer Science Division of Argonne National Laboratory. He is also an adjunct faculty of the International Centre for Theoretical Sciences, TIFR, Bengaluru and associate faculty at the Divecha Centre for Climate Change, IISc. His research interests include studying the Indian Summer Monsoons with climate models, using high performance computing for climate and weather modelling and downscaling of climate data to finer scales. He has collaborated with computational scientists on improving throughput of climate models on HPC systems and for developing visualization and analysis techniques for large atmospheric datasets. He has received the CV Raman young scientist award from Govt of Karnataka.

Title of Talk: Climate Simulations and HPC

Abstract:

Weather and Climate models are computationally intensive and it has always been a challenge to get good throughput for these models from HPC systems. Numerical Techniques used in climate and weather models have also evolved with system architecture to exploit their capabilities. With the advent of many-core and accelerated systems, the challenge to climate modelers has increased to exploit the capabilities of these systems. Some techniques to exploit the capabilities of these systems such as overlapping of computations from different components of a climate system model between CPUs and accelerators will be discussed.

Dr. Surya Rao/Ashok Kumar, (Scientist F, Indian Tropical Metrology, Pune)

Profile: Dr. Surya received his Ph.D. from Andhra University. His research areas are air-sea interactions in the Indian Ocean, Coupled Ocean-Atmosphere Dynamics, Climate Variability and Prediction, Physical Oceanography, Coupled Climate Modelling, Ocean Modelling and Physical Oceanography. Dr. Surya was a gold medalist at Andhra University with several awards in his career in Japan and India, including the Frontier Research Center for Global Change Award for outstanding research in climate studies. He also received the award for Highly Cited papers from Thompson Scientific.

Prof. Santosh Ansumali, (Associate Professor at Jawaharlal Nehru Center for Advanced Scientific Research)

Profile:

I am an Associate Professor at Engineering Mechanics Unit, JNCASR Bangalore. I am also Ramanujam Fellow of Department of Science and Technology (DST) India since 2009. Before joining JNCASR, 2005-2009, I was an assistant Professor at School of Chemical and Biomedical Engineering, NTU Singapore. I did my Post-doc at Institute of Energy Technology, ETH Zurich (Switzerland) in year 2004-2005. My PhD was from Institute of Polymer Physics, Department of Materials ETH Zurich (Switzerland) in year 2000-2004. My MS was in Chemical Engineering, IISc Bangalore (1998-2000) and BE in chemical Engineering from BIT, Sindri (1994-1998). My research interest is mainly in Mesoscale methods for computational Fluid Dynamics, High Performance Computing and Computational Kinetic Theory. I look for algorithms and physical models which are fundamentally suited for high performance computing. I have authored 46 papers in various international journals.

Title of Talk: Stencil Computing for Exa-Scale Era

Abstract:

With the expected arrival of Exascale computing by the end of this decade, direct numerical simulations of complex phenomena such as turbulence with a large number of degrees of freedom is a distinct possibility. The primary challenge for such simulations is successful management of ultrafast floating point computation with slow data movement and communication bottlenecks. This problem is compounded by the presence of random, transient hardware errors and delays. Since this trend is expected to continue in the foreseeable future, It is thus imperative that algorithms and codes exploit extreme parallelism and rely on careful data movement. In this talk, I will present a recently developed framework for decoupling communication with computation for stencil computing [1]. I will highlight a new data structure for stencil computing [2], which allow efficient utilization of available memory bandwidth. Finally, I will highlight how these ideas put together are showing possibility of fluid dynamics simulations without any explicit turbulence modeling approach.

Ms. Akshara Kaginalkar, (Associate Director and Head of the Atmospheric Group of C-DAC)

Profile:

Akshara Kaginalkar is an associate director at C-DAC, Pune and leads the Computational Atmospheric Science (CES) group, which has initiated a wide range of computational earth projects using the PARAM series of super computers. The CES group is engaged in the application of HPC to Atmospheric, Oceanic and Environmental Sciences and their research and technology employs a complete computational environment, modeling and simulation services, data analysis and processing tools and HPC related skilled manpower resources. The CES group is expanding its wings, looking to enhance its research and development activities in areas like Real-time Weather Forecasting using the WRF model “Anuman.” Real Time Weather Forecasting is very important for weather scientists, operational meteorologists, and aviation/ transport industry as well as policy makers, and the CES group has developed an automated workflow for real time weather simulations and a number of different weather forecast products useful as a decision support for various user communities. The weather forecasting system hosted at “http://rtws.cdac.in” provides hourly weather outputs with 72 hours lead time. This group also works Army, Navy and Air Force projects requiring weather and wind predictions, and it has teams working in Extended Range Monsoon Prediction, including research to improve the seasonal forecast of the Indian Summer Monsoon using Global as well as Regional Numerical models.

IRUS2: Life Science & HPC: Issues in computer science, applied mathematics, scientific data management, visualization, and informatics

Prof. Shyam Sunder, (Professor, Computer Science Department, TATA Institute of Fundamental Research, Mumbai)

Profile:

R K Shyamasundar is a Fellow of IEEE, a Fellow ACM, a Distinguished ACM Speaker, and an IEEE Distinguished Speaker. He is currently Senior Professor and JC Bose National Fellow at the Tata Institute of Fundamental Research. He has more than 300 publications, 8 books, 8 international patents & 3 Indian patents. Thirty five students have completed their Ph.D. under his guidance, and he has served on IEEE Standards and as consultant to ESPRIT projects. He did post-doctoral work under the legendary Turing Laureate Professor Dr. Edsgar Dijkstra and was a Distinguished Visiting fellow under Royal Academy of Engineering at the Computing Laboratory of University of Cambridge. He has served as Faculty/Staff at IBM TJ Research, Eindhoven University, State University of Utrecht, Pennsylvania State University, University of Illinois, University of California, San Diego, IRISA, Max Planck Institute, IBM Research India etc. He was Founding Chair of FSTTCS and founding President of IARCS. He serves on the Governing Council of IIIT Allahabad, IIIT, Jabalpur, CSIR Centre of Fourth paradigm (earlier CMMACS), Bangalore and serves on the Technical Advisory Board of BSE (Bombay Stock Exchange) and IDRBT, Hyderabad. He has served on IEEE Esterel Standards Committee.

He is a Fellow of Indian Academy of Sciences, Indian National Science Academy, National Academy of Sciences, National Academy of Engineering and a Fellow of the Academy of Sciences of the Developing World (TWAS), Trieste, Italy. He is also a Fellow of the Computer Society of India and IETE. He was recently named Distinguished Alumnus of Indian Institute of Science.

Dr. Rajendra Joshi, (Well-known industry research expert, currently working at C-DAC as associate Director of Bio Informatics group involved in several projects from Cloud Services to cancer research using HPC)

Profile:

Dr. Rajendra Joshi holds a Master’s degree in Biochemistry from the University of Pune and a Doctorate in Biochemistry from National Chemical Laboratory (University of Pune). He has been associated with the area of Biotechnology & Bioinformatics for the about 28 years. He has over 20 years of experience in the area of Bioinformatics. He is primarily responsible for building a strong bioinformatics group at C-DAC and is presently serving as an Associate Director and Head of the Department at C-DAC. He has set up Bioinformatics Resources and Applications Facility (BRAF) which serves as a nodal point for all researchers in life sciences who require high speed computing. The BRAF facility plays an important role in the Indo-US collaborations. Similarly, he has been responsible for setting up the Indo-UK collaboration with the Biotechnology and Biological Sciences Research Council (BBSRC, UK) on molecular dynamics simulation of large macromolecular assemblies. His major area of expertise is in the use of high performance parallel computers for biological research. His unique strength is in the form of good knowledge of biology and parallel computing. His main research interests include molecular dynamics simulations of nucleic acids & proteins, genome sequence analysis, metabolic pathways and development of specialized computational tools for applications to biology. He is a principal investigator for funded projects and other research projects, leading to a number of publications in international journals and conferences. He is actively involved in collaborative programes and national scientific meets. He has numerous publications, articles and invited talks to his credit. His career objective is to build an environment where much of biology could be understood using high performance parallel computers.

Prof. U.S. Bhalla, (Professor and Dean of National Centre for Biological Sciences (NCBS), TIFR, Bangalore)

Profile:

U.S. Bhalla is currently Professor and Dean at NCBS Bangalore where he has been since 1996.His area of research is computational neuroscience of memory and sensory processing. Prof. Bhalla did his BA (Physics) at Cambridge, PhD (Biology) at California Institute of Technology, and Post-doc at Mount Sinai School of Medicine. He works on various editorial and institutional boards and is the India representative to the International Neuroinformatics Coordination Facility. He is a Fellow of Indian National Science Academy and Indian Academy of Sciences and prestigious Bhatnagar awardee.

Title of Talk: Challenges in Multiscale Modelling of Brain Function.

Abstract:

The brain is both a supremely complex and capable computational entity, as well as the greatest computational modelling challenge. The human nervous system is itself vast in terms of interactions and numbers, with roughly one hundred billion neurons and one thousand trillion interconnections. Thus brain modeling must address vast datasets, similar to such fields as weather modeling. However, the brain differs from many other grand computational challenges in that the algorithms themselves are diverse, heterogeneous, and multiscale. Neuronal modelling requires algorithms that describe multiple scales and processes that include electrical and chemical signaling, genetic and structural change, mechanical processes, all coupled tightly together. In this talk I will discuss the development of MOOSE, the Multiscale Object-Oriented Simulation Environment, which we have used for brain modelling. I will emphasize the computational challenges and opportunities that we have encountered, and that we would enthusiastically seek collaborations to address. These range from numerical and algorithmic, to data management, interoperability, and various forms of parallel computation.

Prof. Pushan Majumdar, (Professor and Senior Research Scientist from Indian Association for the Cultivation of Science, Kolkata)

Title of Talk: Parallelizing Lattice QCD calculations on multicore CPUs and GPUs

Abstract:

In this talk we will look at the computational problem of the theory of strong interactions in particle physics and why it needs PetaFlops of computing power. We will look at some attempts to parallelize such computational tasks using openMP, openMP + MPI, CUDA and openACC programming environments and compare the different approaches.