Big Data, Small Testing ?

– by Prof. Jayant Haritsa, IISc

IEEE Bangalore Section

Outline of the talk

Big Data has become the buzzword of choice in recent times, especially in the software industry. The accompanying hoopla has spawned frenetic claims foretelling the development of great and wondrous solutions to Big Data challenges. However, there is very little said about the testing of such systems, an essential prerequisite for deployment. In this talk, we will discuss the research challenges involved in the testing process, especially from the database perspective. We will also present CODD, a graphical tool that takes a first step towards the effective testing of Big Data deployments through a new metaphor of “data-less databases”. CODD is currently in use at industrial and academic institutions worldwide.

About the Speaker

Prof. Jayant Haritsa

Jayant Haritsa is on the Computer Science faculty at the Indian Institute of Science, Bangalore, since 1992. He received a BTech degree from the Indian Institute of Technology (Madras), and a PhD degree from the University of Wisconsin (Madison). His research interests are in database system design, analysis and testing.

Keynote 2: The Challenge of Performance: Software is (Still) the Key
– by Dr. Sunil Sherlekar

Outline of the talk

Abstract: The challenge of achieving performance for parallel computation will be addressed with emphasis on scientific computing. The key issues will be brought out and possible solutions will be outlined. Most of the solutions will be in software but some possibilities in hardware will also be discussed.

About the Speaker

Dr. Sunil Sherlekar

Sunil Sherlekar is the Chairman and CEO of SankhyaSutra Labs Pvt Ltd, a company incubated in Jawaharlal Nehru Centre for Advanced Scientific Research (JNCASR) Bangalore. The company is focused on scientific computing with an initial emphasis on Computational Fluid Dynamics. He is also proposer of and a member of the Technology Mission for Indian Railways announced in the Railway Budget in 2015.

From Sept. 2010 to April 2015, Dr Sherlekar was the Director of Parallel Computing Research at Intel Labs in Bangalore. Earlier he was the Founder & Head of Research at Tata Computational Research Labs in Pune (2006-2010), the Head of Embedded Systems R&D at Tata Consultancy Services (2002-2006), the CTO at Sasken Communication Technologies (1992-2002) and on the faculty of Computer Science & Engg. at IIT Bombay (1982-1992).

Dr Sherlekar has a B. Tech. (Elect. Engg.) a M. Tech. (Computer Science & Engg.) and a Ph D. all from IIT Bombay. He has published several papers in the areas of Electronic Design Automation and VLSI Signal Processing, High-Performance Scientific Computing and a book on VLSI Signal Processing. He was an Associate Editor of IEEE Trans on VLSI, on the Steering Committees of ASPDAC and International Conf on VLSI Design and on the Executive Committee of India Semiconductor Association.

Dr Sherlekar’s current areas of research interest lie in modelling of physical systems and design of algorithms and data structures for extracting optimal performance from modern evolving hardware.

Dr Sherlekar is a Fellow of the Indian National Academy of Engineering, an Adjunct Professor at IIT Bombay and an Adjunct Professor of the Academy of Scientific & Innovative Research (set up by the Council of Scientific & Industrial Research, Govt. of India).

Invited Talks:

High Performance networks in multicore/manycore computing era

– by Atul Bodas

IEEE Bangalore Section

Outline of the talk

With emergence of manycore and multicore computing, traditional programming paradigms are changing. Running an application over hundreds of thousands of computing core and extracting performance now equires the application developers to be aware of underlying architecture of the system.

It is, however observed that these efforts are centered largely around the compute elements of the system, with network component being assumed to be completely transparent, needing no active involvement of the application developer. This assumption however is erroneous, considering various challenges involved in network design and making efficient use of the resources.

This talk is aimed towards making users aware of challenges involved in design of contemporary HPC network design, and debating whether time has come for the application developers to become “network aware” as well so as to get maximum benefit and performance out of their applications.

About the Speaker

Atul Bodas

Atul Bodas works in the HPC-Technologies group at C-DAC, Pune and has over 20 years of experience in the area of hardware system design.He has been a key member involved in the design of indigenous high performance network family called PARAMNet which has been backbone of last several generations of PARAM systems.Currently he is involved in design and development of next generation of PARAMNet called “Trinetra”

He is an alumnus of Indian Institute of Science, Bangalore.

Bits to Barrels – New ways of Computing in the Energy Industry

– by Dr. Chiranjib Sur

Outline of the talk

In this presentation, I will give an overview of the kind of non-conventional computational effort which is required in the oil and gas industry to meet the energy demand for modern era. The presentation will outline problems from different areas – ranging from micro to macro scale. I will touch upon areas like material simulation in quantum / classical scale, fluid structure interaction, numerical techniques to model oil reservoir and also about geophysical imaging of subsurface. Some of the problems can be classified as purely high performance number crunching and some are driven by data and analytics.

These problems are very challenging to solve on their own and it certainly needs attention from experts in cross discipline – both from domain and computer science.

I will also talk about other community efforts in this areas which might be of interests to the community.

About the Speaker

Dr. Chiranjib Sur

Dr. Chiranjib Sur, a seasoned leader in high performance scientific computing, is a computational astrophysicist by training with 15+ years of experience in academia and industrial R&D.

Chiranjib has worked in wide variety of areas starting from laboratory astrophysics, material science, computational chemistry, complex fluid flow, scientific visualization, numerical optimization, applied mathematics, machine learning, system software development for HPC systems and few others. He has 3 granted patents and he has published more than 45+ technical papers in different international scientific journals, books and conference proceedings. He has also been plenary and keynote speakers in many national and international conferences and workshops. He serves the editorial board as review editor of Frontiers of Computational Physics Journal in addition to his services as regular referee for the American Physical Society Journals, Institute of Physics (UK) journals and European Physical journal.

Currently, Chiranjib is serving as general chair and member of the steering committee of the IEEE conference on High Performance Computing, Data and Analytics (HiPC – Previously he served as vice-general co-chair and workshop chair at HiPC for last few years. He is also involved as a program committee member for other international computing conferences in India and also engaged in shaping up the HPC community in India. He has leading the effort of creating the industry academia consortium for computational sciences, data and analytics.

Chiranjib is the 2003 recipient of “ISCA Young Scientist Award” from the Indian Science Congress Association. His biography is listed in Marquis and International Who’s Who publications.

He is a member for the technical committee for parallel processing (TCPP) of IEEE Computer Society, Indian Science Congress Association (ISCA), the American Physical Society (APS), ACM, Society of petroleum engineers (SPE) and a senior member of IEEE.

After spending many years at high performance computing lab at IBM, Bangalore, the Ohio State University, USA and Indian Institute of Astrophysics, Bangalore, he is currently working as a senior scientist at the computational centre of excellence at Shell Technology Centre, Bangalore.

Chiranjib lives in Bengaluru with his daughter, wife, Royal Enfield and a home loan!

Using Machine Intelligence for Healthcare

– by Dr. Geetha Manjunath

Outline of the talk

Extreme digitization in every field of life is driving Data Analytics to be in the core of any new innovation that improves quality or efficiency. In addition, analytics applications in healthcare domain can create tremendous societal impact. It is also a data rich environment with a lot of heterogeneity in the available data – physiological measurements, radiology images, claims, clinical notes and so on. Healthcare poses several interesting technical challenges to analytics algorithms due to large volumes, velocity, imbalance in data, high levels of accuracy expected and considerable missing information in the input, among others. In this talk, I will share some examples of novel machine intelligence algorithms that we have developed towards improving outcomes in critical care and cancer screening.

About the Speaker

Dr. Geetha Manjunath

Dr. Geetha Manjunath is the Co-founder and CEO of NIRAMAI Health Analytix, a startup developing a novel breast cancer screening solution. Geetha has over 25 years in IT research and has created many innovative prototypes, patents, publications and new products. Till end of 2016, she was the head of Data Analytics Research at the Xerox Research India. Before that, she was a Principal Research Scientist at Hewlett Packard Labs India and a Technical Lead at C-DAC Bangalore.
Geetha is a gold medalist from the Indian Institute of Science (IISc), Bangalore from where she holds a Master’s Degree and PhD from Computer Science and Automation Department. She was the winner of 2009 NASCCOM IT Innovator and 2010 MIT TR Grand Challenges for technologist. She is the author of a book on Cloud Technologies called “Moving to the Cloud”. She is currently the Chair of IEEE Computer Society, Bangalore Chapter, and on the Governing Board of CCICI. She holds 11 US patents to her credit.

Green Software Development and Green Programming


Outline of the talk

The importance of saving energy in general need not be over-emphasized.Computing equipment of various kinds, e.g., mobile phones, laptops, portable games, desktops, servers, and large data centres, all deserve to be consideredseriously as candidates for saving energy. The techniques for saving energy that need to deployed at various levels, e.g., in a data centre and in an embedded device, will be certainly different. The former looks at work load characterization and the latter heavily depends on the compiler to do the job.

This talk will provide an overview of various techniques for saving energy in software from a compiler point of view and provide a glimpse of the important area of green programming.

About the Speaker


Y.N. Srikant received his B.E in Electronics from Bangalore University,and M.E and Ph.D in Computer Science from the Computer Science and Automation department of the Indian Institute of Science. His area of interest is compiler design. His most recent research includes timing analysis of programs, parallel programming languages and their compilers, and security analysis of Android apps. Srikant is a Professor in the Department of Computer Science and Automation at the Indian Institute of Science in Bangalore.

Efforts in realization of scalable quantum computing by exploiting quantum correlations and ultrafast dynamics of complex systems: vibrating molecules, colloidal clusters, and mixed liquids

– by Debabrata Goswami

Outline of the talk

We highlight the importance of the dynamical aspects of complex systems by using emtosecond laser pulses to probe and perturb the dynamics of infra-red dye molecules in several environments and confined conditions. Importance of obtaining the molecular description on relation between dynamics of solute molecule and solute-solvent interactions arises from the fact that the dynamic and static properties of dye molecules are strongly affected by the surrounding solvent molecules. Modulation of the immediate micro-environment of solute can tinker the dynamics of the solute.

In a similar vein, programmable ultrafast pulse shaping or photon flux confinement also result in spatiotemporal control of dynamics. These are demonstrated through femtosecond pump-probe studies or by invoking pulsed optical traps for monitoring complex systems. We will finally show how confinement enables control and how physics principles can be used to unravel the dynamics of complex systems and that it is, in fact, much simpler than one presumes a priori.

These efforts form the basis of our efforts in realization of scalable quantum computing by exploiting quantum correlations and ultrafast dynamics of complex systems: the main theme of the talk focuses on this very aspect of the possible robust implementation.

About the Speaker

Debabrata Goswami

Educational Details of Mr. Debabrata Goswami are as follows:
Ph.D. (Physical Chemistry), Princeton University, Princeton, NJ, USA, June 1994.
Dissertation: Control of Chemical Dynamics using Arbitrary Shaped Optical Pulses and Laser Enhanced NMR Spectroscopy
Thesis Advisor: Prof. Warren S. Warren
Thesis progress Adjudged Excellent and recipient of Hoechst Advanced Technology Division Industrial Affiliates Fellowship
M.A. (Physical Chemistry), Princeton University, Princeton, NJ, USA, 1990.
M.Sc. (Chemistry), Indian Institute of Technology, Kanpur, India, July 1988
Thesis: Fluorescence Quenching of Aromatic Amino Compounds by Chloromethanes.
Thesis Advisor: Prof. S.K. Dogra
Bachelor of Science (Hons in Chemistry), Jadavpur University, Calcutta, India, June 1986.
Currently Mr. Goswami is working as a professor in Indian Institute of Technology, Kanpur, India.

Memory Architecture for Exa-scale Computing: Challenges and Solution Directions

– by Sparsh Mittal

Outline of the talk

Memory system plays the primary role in determining performance, reliability and power consumption of all computing systems, ranging from hand-held embedded systems to massive supercomputers. The ever-increasing data-intensive nature of state-of-the-art applications demands significantly higher performance within the constraints of power and cost, beyond what conventional memory technologies and architectures can provide. For example, more than 100 hours of video and 250K photos are uploaded every minute to YouTube and Facebook, respectively.

Similarly, future exascale supercomputers (i.e., those capable of 1000 petaflop computations per second) will require at least 100 petabytes of main memory, which, if designed with today’s DDR3 DRAM, would consume 52MW power, which is far more than the 20 MW power budget mandated for the entire system. This talk will highlight the challenges, such as power consumption and reliability, in designing memory systems for next-generation exascale computing systems. It will also point out possible solution directions, such as emerging non-volatile memory, heterogeneous memory systems, near-threshold computing, etc.

About the Speaker

Sparsh Mittal

Sparsh Mittal received the B.Tech degree in electronics and communications engineering (ECE) from IIT, Roorkee, India and the Ph.D. degree in computer engineering from Iowa State University (ISU), USA. He has worked as a post-doctoral research associate at Oak Ridge National Lab (ORNL), USA for 3 years. Dr. Sparsh was the topper of ECE batch in B.Tech final year and has received fellowship from ISU and performance award from ORNL.

He has published more than 60 papers in top conferences and journals. His research has been covered by several technical news websites, e.g., InsideHPC 1, InsideHPC 2, InsideHPC 3, Primeur Magazine, StorageSearch,, TechEnablement, ScientificComputing, SemiEngineering (semiconductor engineering), and HPCWire. He has given invited talks at ISC Conference (Germany), University of Michigan and New York University. His webpage is .

Cryptanalysis using Supercomputers, FPGAs and GPUs

– by Dr P. V. Ananda Mohan

Outline of the talk

About the Speaker

Dr. P. V. Ananda Mohan

Dr P. V. Ananda Mohan obtained his Ph.D degree in Electrical Communication Engineering from Indian Institute of Science, Bangalore in 1975. He was with I.T.I. Limited, in R&D from 1973 to December 2003. Since January 2004, he was with Electronics Corporation of India Limited, Bangalore as Advisor (Telecom) and in November 2004, appointed Executive Director (Technical).

His research interests are in the area of Analog VLSI design, VLSI architectures, Cryptography, and he has published in these areas in referred international journals and conferences. His books include Switched Capacitor Filters: Theory, Analysis and Design published by PrenticeHall (London) in 1995, Residue Number Systems: Algorithms and Architectures published by Kluwer Academic Publishers in 2002 and Currentmode VLSI Analog Filters: Design and Applications in 2003 by Birkhauser.

He is a Fellow of IETE, and the chair of IEEE CAS Chapter, Bangalore. He was the Associate Editor of IEEE Transactions on Circuits and Systems Part I during 20002003. He is an Honorary Editor of IETE Technical Review. He taught Mixed Signal Design for a few semesters at the Indian Institute of Science, Bangalore. He is also the current secretary of IEEE Bangalore Section.