My research group works on two different areas
• Spatial Audio and Computational Sound Scene Analysis:
I have been fascinated by the way humans effortlessly use spatial sound for making sense of the environment. A desire to understand the mechanisms used and to provide machines with the same capabilities has been a driver
in this research. The use of computing, mathematical physics and experiments have characterized this work.

• Large Scale Scientific Computing
In scientific computing I focus on very large problems using approximation algorithms and parallel computing. The work has been driven by applications in audio, speaker ID, vision, physics, and machine learning. The focus here is on extending algorithms based on the fast-multipole method, and the use of data-parallel architectures to achieve orders of magnitude speed-ups.

My Carbon group within CSAIL MIT is working on a new operating system for massive multicores and cloud computing called Factored Operating System (FOS). We are also working on multicore architectures for 1K cores. Postdoc positions are available in both these areas.

Intra-die process variations in nm technology nodes pose significant challenges to robust design practices. Geometric variations along with random dopant fluctuation effects have had significant impact on Logic/Memory functionality/yield. The inaccuracies in the models and variabilities in the process are more pronounced and key is to understand the variability effects. Innovative analysis techniques as well as methodologies are needed to counteract the variability issues. Our team is heavily engaged in this activity.

Smart statistical and numerical techniques are key for understanding the variability in VLSI circuits. These techniques need to be highly efficient to be used by designers. New algorithms to achieve speed-up in these techniques are essential. Extension of these smart algorithms beyond the VLSI domain would be critical to achieve accurate and computationally practical would benefit such fields (e.g application to medical field). We are exploring new algorithms and their applications to fields other than VLSI.

As the VLSI circuits scale power and performance remain the top issues. Developing efficient circuit techniques based on emerging devices (FinFet derived, nano-devices) needs extra attention.
We are involved in fabrication of such circuits. Also we are engaged in modeling of 3-D structures and the thermal management of multilevel domain.

Scaling and manufacturing of alternate memories such as MRAM and PCM would help to replace DRAM. Variety of activities from modeling to fabrication are explored in our team

I am broadly interested in topics around cloud computing, including software for datacenter management, cluster file systems, techniques for managing data center power, hybrid cloud technologies, programming frameworks, operating systems, and cloud applications. As part of my research, I manage a 200-server cluster designed to enable experimentation with prototype system software, so our research facility provides ample opportunity for empirical evaluation.

The main focus of my work is on energy efficient computing. Energy consumption has already become one of the most important design parameters in electronic systems. In large scale data centers it is because of the high cost of powering and cooling the equipment. We are currently working on design of VM level energy management strategies that integrate measurement of power, temperatures and workload characteristics, while monitoring QoS of applications running within VMs. In addition, we are evaluating how to integrate SmartGrid information with VM level energy management strategies and how to leverage accurate predictors of the availability of renewable resources that we developed. This work is funded by a large center on Multiscale Systems (www.musyc.org).

We recently also won an NSF Expedition on variability. The work here focuses on how to address variability issues in general purpose systems due to environmental, manufacturing and vendor differences. The current focus is on cooling and thermally aware scheduling techniques for 2 and 3D integrated server and mobile systems.

I am directing Data Mining Group in the Center for Computational Science at the University of Miami. Overall goals of the group is to develop methods for for understanding, spatial/temporal, heterogeneous data.

Limited bandwidth, increase in power dissipation at longer communication distances create a major communication bottleneck in high-performance computing (HPC) systems, affecting not only their performance, but also their scalability. My research interests include designing interconnects for high performance computing systems at all levels: on-chip, chip-to-chip and rack-to-rack.

Come join a large well-funded research effort in cloud computing! Over 10 Carnegie Mellon University professors and 4 Intel Labs researchers are teaming up to drive a common research agenda in cloud computing. Contrary to the common practice of striving for homogeneous cloud deployments, we are exploring the use of specialization as a primary means for order of magnitude improvements in efficiency (e.g., energy), including the design of new platform configurations based on emerging technologies like non-volatile memory. Other research themes include automation at cloud scale (e.g., resource allocation/scheduling, problem diagnosis), big data analytics beyond search (including over live data feeds), and new paradigms for meshing client devices and cloud. Our research lab is a great place to do research. It combines the resources of a well-funded industrial research lab (e.g., our lab’s cluster has 1500+ cores and 600+ terabytes of storage) with the academic freedom of a university. As for me, I am the Intel PI for the cloud computing efforts. I have over 120 publications (cited 10,000+ times), including co-authoring award-winning papers at ICDE, ISCA(2), NSDI, PLDI, and SIGMOD, as well as 13 other papers that were selected for “best papers” journal issues for their respective conferences (including ICFP, PODC, PODS, SIGCOMM, SPAA, and VLDB). I have served on 50 program committees for international conferences (including program chair, vice-chair, etc), and have helped place junior collaborators on a number of top committees. I am on the editorial board of the Journal of the ACM, and an ACM Fellow.

My research concerns the development of efficient algorithms and software
for matrix algebra problems This work began with the development of
innovative memory-efficient algorithms and, more recently, has moved
toward tools to aid in programming of matrix algebra software. I have
recently been collaborating with experts in compiler technology, focusing
on compilers that create fast numerical software. Our initial focus
has been on making efficient use of the memory hierarchy on a single
processor but we are moving into multicore and GPU implementations.
I am also interested in usability of scientific software. To that end,
I am working with collaborators on a tool to automate the construction
of numerical software. Given a problem specification, the tool will
find and tune appropriate routines for its solution.

I was co-developer of an award-winning, NSF-funded undergraduate
curriculum in high-performance scientific computing and have continued
to work on innovative approaches to education in my field. I have
also conducted research on factors that influence women’s interest in
computer science.

I have broad interests in systems spanning applications, system software, networking, and architecture. Current interests include: 1) Cloud/Grid applications, system software, and architecture; 2) Computer Architecture for Exascale computers; and 3) Programming models and tools for post-Moore’s Law computing substrates.

HP Labs Immersive 3D

HP Labs is conducting research using large display walls, 2D and 3D. HP has multiple product and services offerings in the so-called “Big Walls” area ranging from digital signage to large telepresence facilities (such as HP’s Halo). Our research extends this by following two main, intertwined branches. The first examines a new style of 3D entertainment, and the other targets commercial and industrial uses for 3D enabled operations centers, brainstorm or war rooms, and control centers. We are conducting research in:
• Computer vision, image processing, and recognition,
• Robust multi-imager and multi-camera modeling and calibration,
• 3D data visualization,
• High performance computation, transport, and imaging on hardware accelerated platforms (CPUs and/or GPUs),
• Novel human-big wall interaction modalities across heterogenous systems, and
• Local and remote collaboration technologies.

We have publicly shown our Immersive 3D Entertainment efforts with events at the 2010 Sundance Film Festival and the Earth Wind and Fire Concert at 2011 CES in Las Vegas (links listed below). HP’s goal is to experiment with non-standard aspect ratios (e.g. 3:1 for concert stage, 5.5:1 for basketball court) to allow the audience to experience the event as though you are seated at the best seat in the house. This is a break from the traditional movie view from the director’s eyes and allows the audience to “look around” and soak up the scene which works so well in 3D. To do this, we are experimenting with large multi-projector 3D displays (Pluribus), multi-imager camera capture (Herodion), and advanced digital image pipelines (Pericles). We conduct in-situ laboratory studies at entertainment events such as concerts, sports events, fashion shows, and other gaming.

We have built an experimental operations center with multiple 2D and 3D walls and other services. Together with many HP customers, we are exploring novel ways to use Big Walls to address issues in areas such as disaster recovery, emergency response, hospital wards, city monitoring and management, mergers and acquisition planning, product planning, and supply chain management. We are creating a next generation capability by using Big Walls together with mobile devices, touch surfaces, 3D data visualization techniques, and techniques for displaying, manipulating and visualizing large amounts of unstructured data. In addition to our research, we participate in the real world by providing support for HP’s worldwide operations centers and the HP supported gallery at the Newseum due to open in 2012.

Links to entertainment reviews
HP 3D Live: CES Earth Wind and Fire concert

http://blogs.forbes.com/oliverchiang/2011/01/08/ces-hp-believes-in-3-d-too-but-on-a-larger-scale-much-larger/

http://www.monstercable.com/events/ces2011/3dexperience.asp

http://h20435.www2.hp.com/t5/The-Next-Bench-Blog/Earth-Wind-and-Fire-Going-Large-Live-and-in-3D-at-CES/ba-p/60745

http://www.hardwaregeeks.com/index.php/site/comments/hp_streams_concert_live_in_3d/

http://venturebeat.com/2011/01/14/hp-streams-earth-wind-fire-in-live-3d-on-a-huge-screen/

2010 Sundance Film Festival

http://philmckinney.com/archives/2010/01/sundance-and-3d.html

http://h20435.www2.hp.com/t5/The-Next-Bench-Blog/Stories-in-3D/ba-p/52743

HP Newseum announcements

http://www.hp.com/hpinfo/newsroom/press/2010/100930a.html

http://www.newseum.org/news/2010/09/hp-announcement.html

My research focuses on the development of mathematical models and algorithms for estimating evolutionary history in Biology and Historical Linguistics. The main objective is to develop methods that produce much more accurate estimations of evolutionary history than can be obtained using existing tools. Our group is distinguished from many other groups in computational biology due to our focus on ultra-large datasets, with up to 500,000 sequences. We use real data and perform simulations to evaluate the performance of methods that we develop. This research area inolves mathematics, probability, statistics, computer science, and intensive collaborations with domain specialists. My current research is funded by two grants from the National Science Foundation, one an Assembling the Tree of Life (ATOL) grant for simultaneous estimation of alignments and trees, and another for estimating species trees from gene trees. I also have an active interest in metagenomic analysis. No background in Biology is needed. However, good mathematical intuition, and skill in algorithm and software development are important.

Design, implementation and evaluation of distributed services and applications that adapt and exploit social characteristics; algorithms for efficient computation of social graph metrics (e.g., centrality measures, triad census); experimental systems approach combined with theoretical and analytical evaluations.

I work mostly in two areas Computational Geometry and Computational Biology.

My research group creates new software and algorithms which enable robots to effectively assist physicians and automatically accomplish healthcare tasks. We focus on developing motion planning algorithms and physically-based simulations with applications to robot-assisted surgery, treatment planning, medical image registration, personal assistance, and physician training. Our research spans the following areas:

(*) Motion Planning for Healthcare Robotics: Our objective is to compute actions that enable a robot to automatically accomplish clinical or assistive tasks. Using sampling-based and geometric approaches, we develop algorithms to automatically maneuver medical devices such as robotic surgical assistants or steerable needles around anatomical obstacles to difficult to reach sites, enabling new surgical and interventional procedures. Our research identifies and exploits parallels between medical problems and traditional robot motion planning problems while explicitly considering the complexity and uncertainty inherent in healthcare applications.

(*) Physically-based Medical Simulation: Human soft tissues are heterogeneous and have nonlinear properties, resulting in complex deformations during clinical procedures. Using finite element methods and mesh maintenance algorithms, we are developing simulations of soft tissues and their interaction with medical devices. These simulations can assist physicians in registering diagnostic and treatment images obtained at different times, and can also be used for interactive physician training and procedure planning.

I have broad interests in large-scale distributed systems. I’d be delighted to collaborate with somebody on either of my two main activities:

1. Cluster management – Google’s term for the systems that allocate work to our fleet of computers, solving what are effectively large-scale bin-packing problems under a raft of constraints, failures, and uncertainty. We are pushing the envelope of sophistication and capability in managing clusters of machines, and I’m open to collaborating on a wide range of topics, including scheduling, failure modeling, configuration management, and a raft of automation approaches.

2. Using SLAs (Service Level Agreements) to improve our ability to delegate control to automated decision-making systems when they face trade-offs. Work here includes control mechanisms, control systems, feedback loops, and system design. SLAs need a way to specify the consequences of meeting/not meeting their objectives, and economic feedback mechanisms seem a particularly powerful way to achieve this.

As you might imagine, Google provides a great many really exciting opportunities to try these ideas out at significant scale, on real production systems, with real-world problems and requirements. It’s an ideal proving ground for real systems work in support of a wide range of uses, including cloud computing. And we are looking to publish more papers in these areas, so that would be an explicit goal of supporting this program.

Please contact me if you are interested; I’d be happy to discuss opportunities. I’m based in Mt View, California.

Our research interests lie in the cross-section of petascale distributed systems, eScience, Grids, Clouds, and collaborative computing with a focus on large-scale data-intensive distributed applications. Our research especially responds to the urgent need of scientists across a broad range of disciplines who have large-scale data generation, analysis, sharing and collaboration requirements. We develop a diverse set of algorithms, models, and tools for mitigating the data bottleneck in petascale distributed computing systems as well as supporting a broad range of data-intensive and dynamic data-driven applications.

Our current funded research activities can be summarized in these four subareas:

Data placement scheduling. We are working on priority-based data placement scheduling to provide capabilities such as planning, scheduling, resource reservation, job execution, and error recovery for petascale data transfer tasks. Specific tasks that we are working as part of this component includes development of priority-based scheduling with advanced reservation and provisioning; data aggregation and caching; integration with peer-to-peer and real-time data streaming technologies; and support for early error detection, classification, and recovery.

End-to-end data throughput optimization. In a data-intensive petascale system, effective use of available network throughput and optimization of data transfer speed is crucial for end-to-end application performance. Specific tasks that we carry as part of this component include development of application-level models to predict the best combination of protocol parameters for optimal network performance, including the number of parallel data streams, protocol buffer size and I/O block size; integration of disk and CPU speed parameters into the performance model to predict the optimal number of disk (data striping) and CPU (parallelism) combinations for the best end-to-end data throughput; and development of an estimation service for advanced reservations and provisioning.

Integration with workflow planning. The end-to-end performance of petascale distributed applications will depend on the integration of data transfer and scheduling services into higher level workflow planning and management services. Specific tasks that we carry as part of this research component includes development of a model to choose the best data access method specific to application; integration of the data placement scheduler with the workflow planning systems; and incorporation with the existing storage management systems.

Distributed storage management. We have been working on a novel distribute storage system which will enable transparent handling of underlying data sharing, archival, and retrieval mechanisms, and make data available to scientists across all participating sites for analysis and visualization on demand. The system provides scientists with simple uniform interfaces to store, access, and process heterogeneous distributed data sources. The stored data is catalogued using a novel cross-domain metadata framework which allows queries spanning multiple application domains.

My research focuses on two main topics and their interdependence: the role of physical constraints in processor design, and the architecture of future multi/manycore platforms.

Despite continuing availability of plentiful transistors thanks to Moore’s Law, physical constraints—chiefly temperature, power delivery, energy efficiency, and reliability—now prevent straightforward scaling of processor performance. New design-time and run-time techniques are needed to address these limitations, as well as improved modeling and capabilities. Our group at UVA has a long track record of high impact work in these areas.

These physical constraints, combined with the memory wall, will constrain the types of designs that are possible, requiring both novel combinations of cores and novel memory hierarchies. Our group is exploring these issues, especially focusing on the memory hierarchy, how to organize cores around bandwidth limitations, requisite programming support, and how to incorporate these techniques within power, thermal, and reliability constraints.

The goal of the Galois project is to develop programming notations, compilers and runtime systems for programming multicore processors. We focus on irregular algorithms, which are algorithms in which the key data structures are sparse graphs, trees, sets, etc. This class of algorithms is very general, and includes mesh generation, refinement and partitioning algorithms, SAT solvers, machine learning algorithms, n-body methods etc. Our work has appeared recently in top conferences like PLDI 2011, POPL 2011, OOPSLA 2010, and ASPLOS 2011. Our group has six PhD students and three post-docs, and we are always looking for new talent!

In the area of high performance computing, we are exploring the Exascale architectures. Various teams in HP Labs are looking in broad aspects of Exascale, but to me of particular interest is systems software, management, and sustainability (power in particular). As an expansion of this topic, I am exploring HPC in the Cloud: spectrum from general purpose Cloud support (shared resources and commodity interconnects) to dedicated special purpose hardware with high end interconnects. Next, I am exploring Cloud computing in its broadest sense as well as deployment on the Open Cirrus, the cloud computing testbed (http://www.computer.org/cms/Computer.org/ComputingNow/homepage/2010/0410/rW_CO_OpenCirrus.pdf). Finally, I pursue (Cloud) Sustainability research with the dashboard as an artifact.

Prof. Balazinska is developing tools and techniques for data-intensive scalable computing, cloud computing, sensor data management, stream processing, and scientific data management. Her research focuses on building high-performance data management systems but also building tools for making these systems easier to use (e.g., SQL auto-complete, performance debugging, etc.)

I am interested in a variety of topics related to database management system technology, mostly focused on database system architecture, algorithms, and performance evaluation. We are currently in year two of a large parallel semi-structured data management and analysis project at UC Irvine, ASTERIX, and I am interested in mentoring CI Fellows whose interests closely overlap our work. Roughly speaking, ASTERIX can be thought of as semi-structured data management meets parallel database systems meets Map/Reduce (and friends). We have completed a first version of the data-intensive run-time system for ASTERIX, called Hyracks, and are now building the upper layers of ASTERIX (including a new semi-structured data model and query language) as well. The ultimate goal is the creation of a robust, open source software system that will be shared with the research community. An appropriate CI Fellow would be systems-oriented, excited about building something “real” that others can use, and have experience with Hadoop and/or parallel DBMS technology and/or scalable distributed storage management (including data consistency issues, e.g., transaction and consistency models for large-scale, cluster-based data management). Because we are building a system, we are interested in CS Fellow candidates who are interested in implementation as well as design and evaluation.

In my research, I investigate and develop compiler technology in order to automate domain-specific performance and parallelization transformations that are currently applied by hand or not at all. My current and previous research can generally be categorized into performance improving program transformations for regular and irregular applications and domain-specific analysis and program transformation. The ultimate goal is to raise the level of abstraction for scientists who develop computational simulations of physical phenomena while enabling powerful interfaces for performance programmers that enable the orthogonal mapping such simulations to parallel systems. See http://www.cs.colostate.edu/~mstrout for more information about the research we are doing.

At Colorado State University in the High Performance Computing research group, a CI Fellow would have the opportunity to create and direct research and help supervise students. We have active collaborations with the BioChemistry, Math, and Statistics departments hear at CSU. We also are actively collaborating with researchers at Argonne National Laboratory, Berkeley, Cray, LBL, and Intel. I look forward to hearing from interested candidates.

Formal Methods / High Performance Computing / Software Engineering / Concurrency

Our main interests are to develop scalable formal analysis methods for high-performance computing systems.
Of particular interest are formal dynamic analysis methods for MPI, shared memory threads, and CUDA/OpenCL, as well as hybrid codes using more than one API. Efficient test generation in these domains as well as coverage enhancing analysis methods are of interest. The Center for Parallel Computing at Utah (http://www.parallel.utah.edu) provides ample opportunities for collaboration including new faculty hire in formal analysis of concurrency.

We’re developing general systems for scalable Bayesian inference. We’re exploring a mixture of sampling and point estimation strategies including Hamiltonian Monte Carlo and variational methods.

General purpose tools under development include a general posterior sampler (along the line of BUGS), multiple imputation for missing data, and post-stratification for prediction.

We’re particularly interested in multilevel regression and factor models, with applications to prediction problems in epidemiology, climate modeling, and social science. I’m also interested in large scale probabilistic RNA alignment, expression estimation and pathway modeling.

Other team members include Andrew Gelman, Ben Goodrich, Matt Hoffman, and Michael Malecki.

Our goal is to develop methods that will allow us analyze and ultimately to modify the computations that cells have evolved to carry out in response to signals from their environments.

My main research interests are in the areas of
- High-performance processors
- Memory hierarchy, prefetching, cache coherence
- Multi-cores and multiprocessors
- Power and temperature management
- Compiler and architecture co-design

My research focuses on the interaction between the architecture and software of computer systems and underlying hardware implementation challenges. These challenges include power, reliability, and variability issues across embedded and high-performance computing systems. A basic tenet of my research is that architecture design must be cognizant of these implementation issues, and that multi-layer solutions spanning circuits, architecture, and software can provide significant advantages.

It is now widely recognized that current commercial many-core systems are simply not good enough: most programmers can’t handle them. Therefore, alternatives must be developed. Anticipating this problem over a decade ago, the Explicit Multi-Threading (XMT) framework has been under development at the University of Maryland. XMT is a general-purpose many-core computing platform with the vision of a 1000-core chip that is easy to program but does not compromise on performance.

XMT is built to support the PRAM theory of parallel algorithm, which is second in its wealth only to the serial algorithms. Since four decades of parallel computing research provided no real alternative to the PRAM, the XMT project sought to draft specifications for the general-purpose many-core desktop of the future, by first inventing hardware and software support for the abstractions developed by PRAM algorithmics — a task deemed impossible by architecture researchers prior to the accomplishments of the XMT project.

A 2010 status report of XMT appears in U. Vishkin, Using simple abstraction for reinventing computing for parallelism, CACM, January 2011. Order of magnitude speedups, dramatic advantages on teachability from middle school to graduate courses have been demonstrated. And favorable student ranking for achieving speedups relative to standard platforms have been demonstrated.

So far, XMT has spanned applications, parallel algorithms, compilers, HW/SW and education of parallelism. Research opportunities building on this promising foundation include now also CS education, bioinformatics, machine learning and other applications, security, OS, and SW architectures.

There is so much more to the potential of many-core parallel computing than the horizons of commercial hardware offer!

My research and educational career straddles the border between computational theory and mechatronic engineering. Motivated by applications in confined spaces, my group pursues a comprehensive program in mechanism design, path planning, motion planning, and estimation. These research topics are important because once the robot is built (design), it must decide where to go (path planning), determine how to get there (motion planning), and use feedback to close the loop (estimation). Already, we have directly applied this body of work to challenging and strategically significant problems in diverse areas such as surgery, manufacturing, infrastructure inspection, and search and rescue.

Many of the research fundamentals support the development of snake robots, highly articulated mechanisms that can thread through tightly packed spaces reaching locations that people and conventional machinery otherwise cannot. We have developed snake robots for minimally invasive cardiac surgery; recently, we completed our first in human procedure. Current work includes further mechanism development for natural orifice surgery, and prescribing estimation/filtering approaches, based on Kalman and Bayes filtering, to map the internals of the body, i.e., it is SLAM on the inside.

We are also addressing the motion planning of snake robots and all underactuated systems. Our approach takes recourse to the fundamentals, drawing from advanced concepts in differential geometry to prescribe gaits. Examples of results in this work are applying Stokes Theorem to the local form of the connection on shape spaces to efficiently design gaits. Recently, we have shown that many biological systems, including fish and lizards, can be modeled this way.

By taking recourse to the fundamentals, we have been able to address other problems such as multi-agent planning. Recently, we have developed an efficient provably complete optimal multi-agent path planner that can plan paths for 40+ robots in large spaces (note that the size of the configuration space makes it impossible for A* to even search one step, let alone complete a path). We have also developed an algorithm for multi-agent manipulation. Current work includes applying these techniques to distributed manufacturing. We are also working with a biologist to use these concepts to model swarms.

It is the excitement of working with students that continues to draw me to academia. I am certain that a casual tour of my lab reveals a feeling of energy and productivity. My students, both graduate and undergraduate, work hard to provide fresh new insights within the framework of mathematical and experimental rigor endowed by my research program.

My interests are in the application of information and computing technology in organizations, particularly unconventional applications in the service sector. My work has mostly been in applying both quantitative and qualitative modeling to various aspects of an organization, so that we can gain insights from the large volume of data that we are now able to collect. Possible topics of joint investigation include mathematical/empirical models for business performance scorecards, models for organization design, numerical solutions to Markov models.

My group focuses on verification of embedded and cyber-physical systems, studying the question “how can we build computerized controllers for physical systems that are guaranteed to meet their design goals?” These questions are of crucial importance in many areas, including automotive, aeronautics, railway, mobile robotics, factory automation, and medical devices. After all, our society cannot afford to have these systems malfunction. In our research, we have developed powerful logic-based verification techniques that help producing reliable complex systems, e.g., in aeronautical, railway, and automotive applications. We have developed KeYmaera, the first theorem prover for hybrid systems. We also developed the first verification technique for distributed hybrid systems, which combine the challenges of hybrid systems with those of distributed systems. We have further developed the first logic and compositional verification technique for stochastic hybrid systems. My grou p also works on statistical model checking techniques. There are plenty of exciting challenges ahead in these and related directions of research.

I’m interested in advancing the state of the art in developing environments that combine data management, computation, and collaboration to reduce the barriers to use of large-scale resources and to enable cross-disciplinary research and industrial design integrating heterogeneous data and models. I am looking for individuals interested in research coupled with application in support of scientific research, engineering design, or education; we have a number of projects involving collaborations with domain researchers in science, engineering, and the humanties that could provide concrete requirements and interested user communities for individual research topics.

1. Fast, large-scale graph analysis with emphasis on social networks
2. Distributed-memory algorithms with PGAS on clusters
3. Performance analysis, with emphasis on systematic mechanisms for signature capture, bottleneck detection, and refactoring support for HPC applications

Professor Xu’s research interest lies in the development of machine learning models and optimization algorithms for the problems in the field of Computational Biology and Bioinformatics. He is currently working on the following topics: protein sequence/structure alignment, homology detection, protein structure prediction, protein side-chain packing, protein-protein interaction prediction, and biological network analysis. Professor Xu has developed a popular protein structure prediction program RAPTOR, which performed very well in several CASP (Critical Assessment of Structure Prediction) events. His tree-decomposition algorithm for protein side-chain packing now is a major technique underlying the new version of SCWRL, the most widely-used side-chain packing program.

My research interests fall into the broad areas of operating systems, distributed systems, and parallel computing. The overarching goal of my research is to understand the (normal and abnormal) behaviors of complex computer systems and to manage them according to such understanding. The empirical work of my research focuses on server computer systems including, in particular, servers on multicore platforms, I/O-intensive systems, and multi-component network services. More information about my research are available at http://www.cs.rochester.edu/~kshen.

We have several interesting projects in computational science education that could use the efforts of a CIfellow to develop effective learning materials across the sciences and math, and with the use of parallel computing in the process.

My research interests are in the general area of
distributed systems. My work focuses on a relatively new paradigm of Many-Task
Computing (MTC), which aims to bridge the gap between two predominant paradigms
from distributed systems, High-Throughput Computing (HTC) and High-Performance
Computing (HPC). My work has focused on defining and exploring both the theory
and practical aspects of realizing MTC across a wide range of large-scale
distributed systems. I am particularly interested in resource management in
large scale distributed systems with a focus on many-task computing, data
intensive computing, cloud computing, grid computing, and many-core computing. I
was a CIFellow recipient in 2009-2010.

I have several active projects:

•            
dFalkon: distributed data-aware task execution fabric
(http://datasys.cs.iit.edu/projects/index.html#dFalkon)

•            
FusionFS: Fusion distributed File System
(http://datasys.cs.iit.edu/projects/FusionFS/index.html) 

•            
ZHT: Zero-Hop Distributed Hash Table
(http://datasys.cs.iit.edu/projects/ZHT/index.html) 

•            
Falkon: Fast and Light-weight tasK executiON framework
(http://dev.globus.org/wiki/Incubator/Falkon)

•            
Swift: Fast, Reliable, Loosely Coupled Parallel Computation
(http://www.ci.uchicago.edu/swift/main/)

dFalkon is a distributed data-aware execution fabric
supporting both HPC and MTC workloads. The execution fabric is fault tolerant by
having all compute nodes participate in the job submission and handling process;
work stealing is used to achieve efficient distributed load balancing. The
fabric guarantees job execution and dependencies, and relies on an underlying
scalable distributed storage system for inter-process communication. Data-aware
scheduling maximizes data-locality by scheduling computational tasks close to
the data. Computations are overlapped with I/O to reduce wasted resources and
hide latencies. The fabric is elastic based on application demands, and supports
compact task representation to alleviate task submission bottlenecks for common
patterns.

FusionFS is a distributed filesystem that co-exists with
current parallel filesystems in HEC, optimized for both a subset of HPC and MTC
workloads. FusionFS is a user-level filesystem that runs on the compute resource
infrastructure, and enables every compute node to actively participate in the
metadata/data management. Distributed metadata management is implemented using
ZHT. The data is partitioned and spread out over many nodes based on the data
access patterns. Replication is used to ensure data availability, and
cooperative caching delivers high aggregate throughput. Data is indexed, by
including descriptive, provenance, and system metadata on each file. FusioFS
supports a variety of data-access semantics, from POSIX-like interfaces for
generality, to relaxed semantics for increased scalability. 

ZHT is a zero-hop distributed hash-table, which has been
tuned for the specific requirements of HEC (e.g. trustworthy/reliable hardware,
fast networks, non-existent "churn", low latencies, and scientific computing
data-access patterns). The primary goal of ZHT is excellent availability, fault
tolerance, high throughput, and low latencies.

Falkon enables the rapid and efficient execution of many
independent jobs on large compute clusters. Falkon combines three techniques to
achieve this goal: (1) multi-level scheduling to enable dynamic resource
provisioning; (2) a streamlined task dispatcher able to achieve
order-of-magnitude higher task dispatch rates than conventional schedulers; and
(3) performs data caching and data-aware scheduling.

Swift is a parallel-programming tool specifically designed
to address the fast and reliable execution of large-scale scientific
computations on distributed systems, and the concise specification of these
applications.

http://datasys.cs.iit.edu/

My research focuses on two broad areas. First, Internet measurement and the application of these measurements in timely problems, such as the detection of network neutrality violations, the diagnosis of performance problems, and the adaptive transport of video streams.

Second, I am working on applications of network science in biology, neuroscience and climate science.

Currently funded research projects:
1. Evolution of the Internet ecosystem.
2. Applications of network measurements in performance monitoring and diagnosis.
3. Internet video technologies.
4. Economics of network interconnections.
5. Applications of network science in biology, neuroscience and climate science.

Electrical-engineering (ee) degrees. Learning in fields such as art/architecture.
Widely published in archival journals. Three edited books. Authored encyclopedia articles.
Doctorate: dissertation in stochastic control (ee); mathematics and statistics minors; reading knowledge of Russian and French. Now use Google-translate to write Russian.
Rand Corporation employment and UCLA projects on applied problems ranging from operations research to biomedical computing.

Emeritus Professor since 1994.

My research is in power efficient computing. Power efficiency can be achieved on various level from algorithm, architecutre, microarchitecture and circuits. I enjoy learning new things and would not mind working with researchers from different backgrounds.

VLSI CAD for More Moore (nanolithography beyond 22nm) and More-than-Moore (3D-IC, nanophotonics, etc.);
Physical design and technology co-optimization;
Computational lithography;
Vertical integration of architecture/CAD/circuit/technology;
Design/automation of emerging technologies

Developing scalable data mining algorithms for processing massive volumes of data
Statistical machine learning
Text analysis
Data mining in streaming environments and distributed platforms
Time series analysis

For lots of applications like speech recognition and machine translation, it is recommended that the training set should be similar to the test set. But real life isn’t like that. What can we do when the test set doesn’t match the training set? In many practical settings the test document will be unique in lots of unanticipated ways (author, topic, speaker, genre). We need methods that are more robust to mismatches between the training data and the test data. We also need methods that can adapt to the unique properties of the test document.

1. GPU computing (GPGPU)
2. Massively parallel algorithms
3. Heterogeneous computing
4. Computational Intelligence, especially accelerated by GPUs
5. High-performance computing.

My research interests lie in data mining, high performance computing, and architecture-conscious algorithms. I am especially interested in designing algorithms and systems support to process and analyze large data sets, while ensuring efficient utilization of modern and emerging computer architectures.

I am currently interested in distributed algorithms deployed on a complex network of coupled processors. Other interests include using parallel architectures such as GPU to speed up image processing operations.

Large-scale business analytics and Seismic Imaging

1. Large-scale graph analysis on current and emerging HPC systems with emphasis on social network analysis.
2. Performance analysis and tuning for HPC applications with focus on systematic method, knowledge mining, refactoring on massively parallel systems such as blue-gene
3. Fast algorithm implementation and performance debugging with PGAS languages

My group has developed middleware, FG (http://www.cs.dartmouth.edu/FG/), which helps to mitigate latency in data-intensive programs. With FG, the programmer structures the computation as one or more pipelines (not constrained to be linear pipelines), where stages work on buffers asynchronously. This work came out of real implementations of out-of-core algorithms for the Parallel Disk Model, but it applies to much more than the PDM. I’m looking for a creative researcher to help us extend what FG can do and to help us incorporate FG into real applications.

I am also interested in developing and implementing algorithms for reducing the number of high-latency operations, on models such as the PDM. Because implementing such algorithms is as important as designing them, this project would be a good opportunity for a CIFellow who is interested in algorithm engineering.

GUI for molecular modeling software and highly parallel computation with GPUs for materials modeling

My research focuses on machine learning with emphasis on developing fast scalable algorithms for massive data analysis. Towards this end our lab develops optimization algorithms and focuses on graphical models, structured prediction, and kernel methods. Application areas include computer vision, social network analysis. I am open to explore any area that is fun and is sufficiently challenging!

My research involves the design, analysis, and practical realization in software of combinatorial algorithms for problems in the sciences and engineering. Recent work has involved: Algorithms and software for variant graph coloring problems to enable Automatic Differentiation; exact and approximation algorithms for matchings (for applications in scientific computing) and edge cover (to analyze flow cytometry data and identify leukemia); and the design of multithreaded combinatorial algorithms on emerging many-core architectures. My work is based in algorithmic theory, but is motivated by practical problems, and leads to software for massive graphs and data sets.

My research interests span distributed systems, networking, operating systems. Current projects include:

- Energy measurement and management in mobile computing: in today’s portable, battery-operated devices, energy must be a first class optimization metric for operating systems and applications. Building on our previous work on Quanto (OSDI’08 with Prabal Dutta, Philip Levis and Ion Stoica), we are interested in operating systems abstractions that allow applications, developers, and device drivers to understand energy usage, and apply this knowledge for energy-aware optimization, admission control, and scheduling.

- Proactive datacenter networking: modern datacenter and enterprise networks are increasingly dynamically programmable. We are investigating the question of how knowledge of near-future workloads (e.g., by looking at an upcoming queue of large batch jobs) can be exploited to reconfigure the network (including endpoints) in a proactive way, to increase performance, isolation, and fairness.

- Tracing distributed systems: as the distributed systems we build become more complex, so does the task of understanding how they work, and, especially, how they fail. We are interested in the implications of causal tracing as a first-class concept in distributed systems design and integration. This builds on our previous work on X-Trace (with George Porter, Ion Stoica, Scott Shenker, and Randy Katz).

In addition to these problems, I am also interested in programming models and frameworks for large scale data intensive parallel computing (MapReduce and beyond), including understanding what classes of algorithms can be cast efficiently into existing frameworks, and what new abstractions can enable new applications.

I have also worked extensively on networking protocols and architectures for wireless sensor networks, and am interested and open to collaboration in any of these areas.

Our goal at the Systems Research Laboratory (SyLab) is to conduct research in the area of Operating Systems. While we are interested in all problems related to Systems, lately, we have been developing new capabilities for storage systems and virtualized data centers. Some of the capabilities that we have developed recently address energy, performance, and self-management for storage systems and resource isolation and performance guarantees within virtualized systems.

More information about our work can be obtained at:

http://sylab.cs.fiu.edu/

My primary research interests are in computer graphics and animation, especially using physical simulation for computer animation. I am also interested in scientific computing, numerical methods, computational physics and computational geometry.

My research interests are related to two broad thrusts. First thrust is the area of nonrigid motion analysis (especially, elastic and fluid motion), with biometrics and security applications. Second thrust is in the area of biomedical image analysis related to tissue segmentation in MR, CT, PET and microscopy images. I am also interested in high performance issues of machine learning algorithms.

We study the design and fabrication of nanostructures as applied specifically to the fabrication of future computing and sensor systems: devices-to-computer architecture. The terms ‘nanocomputing’ or ‘molecular computing’ refer to the fabrication techniques (e.g., self-assembly) that have the potential to create devices with critical dimensions near the molecular scale (i.e., < 10nm). However, defects introduced during self-assembly require a change in the way we design and build these systems.

Self-assembly is a bottom-up fabrication technique that can be used to achieve molecular scale resolution. The goal is to use these structures to integrate active nanoelectronic devices into a fully self-assembled circuit technology – and to study the new forms of computer architecture that the technology enables. To do this we have adopted a broad and vertical research approach to cover topics in the synthesis and design of DNA nanostructures, nanoscale device and circuit modeling, and studies of emerging computer architectures.

Work in this area requires a deep interest in pushing the frontiers of computing and a traditional background in hardware/software design, computer architecture, or systems. Techniques specific to DNA nanotechnology and the "device side" are acquired here through hands-on laboratory work and analysis.

Our research thesis is to develop large-scale data-intensive analytics platform that combines massively parallel data management with massively parallel data analytics, to extract insights from a combination of structured, unstructured, static and streaming data. We also research on novel applications in industrial contexts, such as retail intelligence, consumer social media, environmental sensing and energy production and operations management, that depend on such an analytics platform for scalabiility and performance.
We are particularly interested in candidates who have experience and knowledge in one or more of the following areas:
• Data mining for time series analysis, real time analytics, incremental algorithms, and event stream processing.
• Analytics over mobile social media and information extraction from text; Web 2.0 applications and consumer analytics.
• Implications of multi-core, parallel co-processors, and flash devices on dat a management and analytics
• Large scale data warehousing (data integration, query optimization, workload management), design of robust query processing algorithms and adaptive indexing

matrix theory and computation, large-scale scientific simulations, image data analysis, image reconstruction, algorithm-architecture co-design, mathematical software

A range of parallel computing: applications to large scientific problems (such as climate
modeling and high-energy physics), performance analysis, and parallel algorithms
for physically realizable models such as mesh-connected computers.

Our group is mainly interested in the research and development of VIEW, a well-known scientific workflow management system in the field. Our next thrust is to develop VIEW on top of the Cloud.

Most of our PhD graduates found faculty positions in other universities.
Therefore, this is an ideal environment for people who like to pursue an academic career.

My primary research interests are in
data mining/machine learning, high performance computing and database systems.
In our lab we seek to develop efficient and novel
algorithms for managing and analyzing complex data. Our recent research
is particularly motivated by
applications that arise in the area of
network science (specifically biological networks and social networks).
Below we briefly describe two projects in these areas, for others
please refer to the personal and laboratory web pages listed.

1. Architecture Conscious Algorithms and Systems:
Here we have been looking at ways
in which various algorithms (XML indexing, Network motif mining, Frequent
pattern mining) can be re-designed to fully exploit the capabilities
of current day architectures ranging from GPUs to
multicores to supercomputing systems. Of particular interest
is the development of an effective infrastructure enabling such algorithms to
scale to very large data stores .

2. Algorithms and Systems for Network Science:
Here we seek to unravel common principles, events, algorithms and tools that
govern network behavior across different domains ranging from social
networks to biological networks. Of particular interest here are not just
algorithms for module discovery, link discovery,
anomaly detection and event detection
but also usable systems infrastructure that can enable
researchers to effectively
query, visualize,
and analyze such networks under various trust, probabilistic and
provenance models.

If you are interested to learn more about our activities in these areas please feel free to contact me as noted herein.

In the past I have focused on high performance computers. I and my colleagues developed some of the first prototype computers that exceeded 200 MHz. However, in the past decade I have focussed on various aspects of low power computing. My research group developed the “Intelligent Energy Management” system used in many ARM cores. I was a co-inventor on the original razor patents, a circuit technology that is being deployed in many next generation low power cores to combat the growing effects of variation that occurs in smaller technology nodes. My research group has developed a series of ultra-low power signal processors targeted at mobile wireless baseband processing. They evolved into a commercial prototype developed by ARM, which led to a commercial spinoff, Cognovo. They will commercialize the processor and provide wireless software stack.
My group were among the first to propose replacing DRAM with non-volatile memory to reduce memory power, cost and footprint. His group has also been active in exploring new server architectures that combine “near threshold” technology and 3D chip stacking to drastically reduce system power consumption. They recently complete the tape-out of a 128-core multiprocessor that integrated the cores, caches and DRAM into a single 3D stack. The fabrication run was funded by a DARPA grant.
Finally, my group is exploring ways to build low power compact interconnect fabrics. We recently completed a compact 128 x 128 crossbar architecture that can sustain a bi-section bandwidth of about 1.3 terabits / second while consuming less than 130 mW.

My main research areas are visualization, geometry processing, and high-performance computing.

I am interested in designing and implementing large-scale distributed systems that are robust and easy to write applications against. I have worked on a variety of projects including a distributed persistent object store, a secure distributed serverless file system using Byzantine fault tolerance, diagnostics in wireless networking, object-relational mapping framework, and a distributed auto-partitioning lease manager for datacenter servers – the last two pieces of work shipped as part of commercial products/large-scale online services. I am currently working on another highly-scalable distributed system that can make certain aspects of Web programs much more efficient and easier to program.

The computational aspects of the research include signal analysis, modeling and optimization of procedures, improved methods of modeling to provide robust conclusions from experimental data.
The lab’s experimental research interests focus on the structural biology of protein domains in intracellular signal transduction, including SH2, SH3, kinase, phosphatase, PH domains, and many others and how natural ligands interact with them. An additional area of interest is the development of NMR and related methods for structural biology. Projects include segmental labeling, other novel isotopic labeling methods, and detection of protein-protein interaction surfaces and dynamics.

The main research theme of my group is to understand the relationship between protein sequence, structure, and function, which is of fundamental importance in life science and the health care industry. We address important questions and problems in proteomics, structural biology, and biomolecular dynamics, using computational biology and bioinformatics approaches. In particular, we develop methodologies and techniques for massively parallel supercomputers such as the IBM Blue Gene, and other high performance computing (HPC) platforms such as grid computing and cloud computing. The group pursues a broad but well grounded approach to leverage and expand upon our existing techniques for scientific and technological advancement.

The current research projects in our group include: (1). Development of novel methods for parallel multi-scale modeling of complex biological systems. (2). Free energy perturbation (FEP) methods for large scale protein-protein, protein-ligand binding affinity predictions. (3) Influenza modeling with both statistical models using bioinformatics tools and physics-based models using molecular dynamics simulations. (4) Protein folding, misfolding, and aggregation with massively parallel molecular dynamics simulations. (5) Interactions of nanoparticles with biological systems and the related mechanism of nanotoxicity. The group is also actively engaged in research projects on modeling of GPCR, and proteomics/biomarkers.

My recent research efforts have focused on energy-efficient data management systems, query processing on new processor and storage technologies, main-memory transaction processing, and column-oriented databases (see my web page for more details: http://nms.csail.mit.edu/~stavros/).

In addition to continuing pursuing research directions in those topics, I am also interested in distributed transactional stores on main memory and/or flash, high performance data analytics on Hadoop clusters, and hardware/software co-design for energy-efficient data centers.

Parallel computing

I work in the areas of computational electronics and quantum transport theory. My group uses a variery of numerical techniques to address the transport of charge and heat in semiconductor nanostructures. Some of our recent projects include:

(1) Electronic and thermal properties of nanostructures. Nanostructured thermoelectrics
(2) Monte Carlo simulation of quantum cascade lasers
(3) Decoherence in nanostructures
(4) Quantum Computing
(5) Global modeling of carrier-field dynamics in conductive media using EMC/FDTD
(6) Flexible electronic systems (e.g. semiconductor membranes and ribbons)
(7) Solid-state-based quantum information processing. Spintronics

For a detailed description of current and recent projects, please visit http://homepages.cae.wisc.edu/~knezevic/

My research is in the broad area of parallel and distributed computing, including cluster computing, grid computing, cloud computing, and volunteer computing. Within this area I focus particularly on questions pertaining to scheduling and resource management, going from theoretical contributions (novel complexity results, novel scheduling algorithms) to practical contributions (system designs and implementations, application deployment). A key technique for conducting quantitative research in this area is simulation of distributed systems. Consequently, some of my research also explores the issues of speed, scale, and accuracy for the simulation of distributed systems. All my recent publications are available on my personal Web page.

My research interests fall at the intersection of nonlinear dynamics, artificial intelligence, and control theory. I am deeply intrigued by the process of modelling: what you can learn — and accomplish — by describing something in the language of mathematics. My students and I have worked on a variety of analysis and synthesis problems whose solutions, at their core, are rooted in the modelling process. The applications involved have ranged from computer architecture to internet attacks, and meltwater ponds on the arctic ice sheets, and the mathematics involved has ranged from topology to time-series analysis, but the underlying questions are the same: what is the right abstraction to use for a given problem, when does it work, how does it fail, and what does all of that enable? Using the right mathematical representations, it becomes obvious, for instance, that computers are often chaotic, that fluids can be mixed by exploiting sensitive dependence on initial conditions, and that denial-of-service attacks can be handled gracefully if one works with their nonlinear nature.

Dynamical systems theory gives us a great way to think about human movement. In collaboration with Jessica Hodgins, we are using this mathematics to model and simulate dance. Along with scientific papers, this collaboration has produced a performance piece entitled “Con/cantation: chaotic variations,” which involves a human dancer and computer animations projected on three 10? by 10? screens. One potential postdoc project would be to explore whether movement — dance and other types — “lives” on some lower-dimensional manifold, and what techniques might be useful in finding that manifold. Another potential postdoc project would focus on the cognitive neuroscience angle of this problem, working with an international group of scientists that is just forming to explore how movement, dynamics, and the perception of beauty interact.

The focus of our AI work is on automating scientific reasoning. In collaboration with software engineers and climate scientists, we have developed computer tools that help geoscientists deduce the ages of landforms from cosmogenic isotope data. This tool is not simply a number cruncher; rather, it uses AI techniques to capture the expert’s knowledge and help him or her make sense of complex relationships in sparse, noisy data. We have just initiated a similar project involving another forensic scientific reasoning problem: the deduction of age models from ice and sediment cores. This would be a particularly appropriate postdoc project for someone whose interests fall at the intersection of computer science and climate science, as well as for someone who is interested in the techniques and ontologies involved in the automation of scientific reasoning.

The current projects in the Advanced Operating Systems group at IBM Research aim at exploring innovation on operating systems and hypervisors to:
(1) advance dynamic resource management in cloud computing environments;
(2) advance support for problem determination;
(3) advance system software support for analytical applications;
(4) leverage high performance computing platforms to advance large-scale commercial applications.

Please contact me for details.

Very large parallel databases on cloud computing platforms. Data integration of genomic databases. query processing on energy constrained sensors. Efficient algorithms for RDF statistics collection.

The polyhedral model is a formalism for reasoning about compute-and data-intensive kernels that take up a huge part of most programs. Such computations arise in many applications in signal and image processing, scientific and engineering simulations, dense linear algebra, bio-informatics, numerical analysis, etc. In fact, most of four of the Berkeley “dwarfs” or motifs can be described very succinctly in the polyhedral model.

In the new era of multi- and many-core, programmer productivity is just as critical as performance (if not more so). The polyhedral model provides a framework for both, optimally compiling high-level equational programs to parallel architectures, and also improving programmer productivity.

I am seeking a postdoctoral researcher to work in this broad area. The specific topic can be determined based on mutual interest, and may range from low level (e.g., FPGA platforms) through discrete nonlinear optimization, program transformation systems to semantics of equational programs.

The Mathematical Sciences Department at the IBM T.J. Watson Research Center is engaged in basic and applied research in several areas of scientific computing, high-performance computing, algorithms, and optimization. We are seeking postdoctoral researchers in the areas of highly parallel graph algorithms, preconditioners, and other core methods to aid the development of massively parallel scalable solvers for large sparse systems of linear equations.

Solving large sparse systems of linear equations is the core computation in a large number of applications in science, engineering, and optimization. With the rapidly increasing number and complexity of scientific applications that model time dependent physical phenomena, there is an urgent need for robust, high-performance, highly-parallel, practical general purpose sparse linear solvers because of the limitations of the current iterative solver software libraries and the asymptotically superlinear time and memory demands of direct solvers. The development of such solvers requires advances in some key enabling algorithms and preconditioning techniques that are self adaptable to a variety of sparse linear systems and machine architectures. In this context, some of the topics that we plan to explore in the near future include parallel approximate and exact algorithms for graph matching (particularly, maximum edge-weight matchings for general and bipartite graphs), parallel gra ph partitioning and mesh distribution, use of machine learning for preconditioner tuning, self adapting preconditioners, preconditioners based on random sampling, composite preconditioners and solvers, etc.

Prof. Vitter’s research interests include the design and mathematical analysis of algorithms, especially dealing with massive data. He has worked extensively in data compression, string algorithms, external memory algorithms and I/O efficiency, computational geometry, caching and prefetching, machine learning, incremental algorithms, and order statistics. His work on the analysis of algorithms deals with the precise study of the performance of algorithms and data structures under various models. His work on I/O-efficient methods for solving problems involving massive data sets has helped shape the subfield of external memory algorithms. He is actively involved in developing efficient indexing methods for text and techniques for space-efficient processing of massive data. Other work includes sorting, information storage and retrieval, geographic information systems and spatial databases, clustering and geometric optimization, random sampling, and random variate generation.

Titles of current projects:
* Robustness of Simulations in Model Based Design Environments
* Robust Testing for System Validation
* Temporal and Modal Logics for Dynamical and Hybrid Systems
* Hybrid System Synthesis from High Level Specifications
* Task and Motion Planning for Mobile Robots
* Natural Language Interfaces for Robotics

As a part of the Exascale Computing Lab at HP Labs in Palo Alto, I am currently working on next-generation data centers. In particular, our project on “data-centric data centers” is examining how “big-data” applications are going to drive future system architectures. My interests in this space includes Infrastructure-as-a-Service systems, non-volatile memories (Memristor, PCM, and Flash), and distributed frameworks such as Hadoop/MapReduce, HBASE, and other scale-out NoSQL systems.

I enjoy building systems and therefore all this work is based on real implementations deployed in our research data center and the OpenCirrus cloud testbed. The majority of our research is open and is published in the traditional top-tier venues for the systems’ community (NSDI, FAST, SOSP, OSDI, etc.).

My research areas include (1) the development of information data management systems in support of clinical genotyping and personalized medicine, and (2) computational genomics in clinical and non-clinical organisms. In addition, I have experience in commercializing research findings in see-stage companies.

Heterogeneous many-core system, architecture

High Performance Reconfigurable Computing

Hardware acceleration technologies (e.g., FPGA, GPGPU), programming model and applications

Hardware design (e.g., image processing, cryptography)

Cache policy design in Solid-State Drives and emerging data storage devices

1. Virtual organizations as socio-technical systems
2. Collaboration via large-scale, megapixel displays (e.g., OptiPortals)
3. Impact of cyberinfrastructure on scientific and engineering output
4. Technology-mediated social participation

My research group focuses on computer architecture, parallel processing, computer systems performance analysis, and high-performance storage systems. We have a particular emphasis on the interaction of software and compilers with computer architecture, and the interaction of computer architecture and circuits. Current projects include: novel architectures for new technologies, stochastic computing, heterogeneous parallel computing (e.g. conventional processors with GPUs and Cell processors), high-performance scientific computing, data de-duplication, incorporating flash memory into scalable systems. We work closely with researchers in geophysics, databases, compilers, and operating systems.

Current Research Interests: My current research principally revolves around the broad topic Network on Chip (NoC), which has emerged as the communication backbone for multi-core chips. With my graduate students and collaborators I am working on the following projects.
•On-chip wireless communication network: Recent research has established characteristics of silicon integrated antennas for intra- and inter-chip communication. Moreover excellent emission and absorption characteristics leading to antenna like behavior in carbon nanotubes (CNTs) are observed recently. In this project we are working on the design methods for wireless NoCs with different types of on-chip antennas.
•Reliable and Low Power NoC: We have designed a family of joint crosstalk avoidance and multiple error correction codes (CAC/MEC) and demonstrated how a low power and reliable NoC can be designed by incorporating these CAC/MEC codes.
•Three dimensional (3D) NoC: NoC has emerged as the communication backbone for the multi-core chips. The performance improvement arising from the architectural advantages of NoCs will be significantly enhanced if 3D ICs are adopted as the basic fabrication methodology. The amalgamation of two emerging paradigms, NoC and 3D IC, allows for the creation of new structures that enable significant performance enhancements over more traditional solutions. In this project we are investigating characteristics of 3D NoCs.
•Network-on-chip based hardware accelerators for Biocomputing: The gap between data generation and data processing is rapidly widening in biocomputing applications, and to close this gap it is imperative to assimilate the latest of breakthroughs in the Integrated Circuit (IC) design community into mainstream biocomputing research. Integrating huge number of processing cores on a single chip can help realize orders of magnitude improvement in performance and eventually will bridge the gap between data generation and data processing. In this project our aim is to design NoC-based hardware accelerators for different biocomputing applications, like sequence alignment and phylogenetic tree construction

Dr. Ferris’ research is concerned with algorithmic and interface
development for large scale problems in mathematical programming,
including links to the GAMS and AMPL modeling languages, and general
purpose software such as PATH, NLPEC and FATCOP. He has worked on
several applications of both optimization and complementarity,
including cancer treatment plan development, radiation therapy,
video-on-demand data delivery, economic and traffic equilibria,
structural and mechanical engineering.

Bioinformatics: Ontologies and Protein Structure Prediction
Assistive Technologies: non-visual access to the web and to complex data
Declarative Languages: logic programming and constraint programming
Parallel Processing: parallel execution of search problems
Knowledge Representation: planning, action languages, multi-agent systems

N-body solvers, which calculate pairwise interactions among a large set of particles, are a vital tool in many simulations of physical phenomena. There is an opportunity to significantly increase the power and the applicability of this technology through the development of algorithms and software of unprecedented simplicity, efficiency, and generality. The basis for such an advance is a relatively obscure approach known as the multilevel summation method (MSM), which is s flexible unified methodology based on a hierarchy of grids (with the possibility of finer grids being localized) and well suited for modern computer architectures. Indeed, we expect multilevel summation to perform an order of magnitude better than other methods for important classes of problems, such as simulations of macromolecules, and expect it to be a formidable competitor in other situations.

The calculation of pairwise interactions and the solution of discrete elliptic equations are the time-limiting steps of applications that consume vast amounts of CPU cycles. Molecular dynamics, in particular, can require months of computer time. The project envisioned here is motivated by problems in computational molecular biophysics, which is being transformed by increasing computing power into a quantitative science with predictive value. And though the MSM will benefit many applications, there is a particular application that makes development of the MSM truly compelling: the use of the spherical and the generalized solvent boundary potential methods for modeling very large systems for which full atomic detail is not needed at a long distance from the active site. Such boundary potentials are implemented in molecular simulators but their use is limited due to the high cost of using standard methods for nonperiodic boundaries. This application is of great interest for very large-scale simulations on proteins that play a crucial role in human disease.

There are several key parts to this work: One is to achieve higher accuracy, using techniques inspired by those that are employed in competing fast N-body solvers. The second is to achieve accelerated performance on current and emerging computer architectures, e.g, using OpenMP or OpenCL (for multicore, multiprocessor CPUs and GPGPUs) and/or MPI (for clusters) for ensembles of smaller particle systems and/or for large particle systems. A third important goal is an adaptive version, e.g. using a pruned oct-tree, for problems with nonuniform particle distributions.

My current research interests are in two broad areas: the automatic analysis of software (either static or dynamic) for verification or bug finding, and in programming highly parallel machines, in particular programming heterogeneous machines with complex memory hierarchies.

I am interested in the interface between discrete and continuous computations. The theory of real computation is a current challenge, and its complexity theory is in infancy.

I am also interested in the algorithmic challenges in applications from computational geometry, graphics, algebraic geometry and continuous mathematics. In the last 20 years,
we have understood a great deal of how to solve the nonrobustness problems in basic geometric algorithms, but much tougher challenges lie ahead in nonlinear geometry, and in transcendental computations.

I’m a glaciologist who studies various parts of ice sheets with numerical models. These models require solutions of various PDEs on various platforms. One of my interests is a development of large-scale ice-sheet model as a part of a Global Climate Mode.

http://www.princeton.edu/aos/people/research_staff/sergienko/index.xml

Our group (currently including five PhD students, two M.S. students and an undergraduate) has a six-year history of developing a robuts, transparent checkpointing package: DMTCP ( Distributed MultiThreaded CheckPointing; dmtcp.sourceforge.net ). DMTCP is now experiencing over 100 downloads per month. We directly checkpoint the binary executable with no modification to the kernel (entirely in user space). We transparently checkpoint and restart most programs, including OpenMPI, Matlab, python, perl, bash, scheme, lisp, etc. We are collaborating with the teams for Condor, Maple, SCIRun and others to support checkpointing for those environments. In the area of supercomputing, we plan to support Infiniband (in addition to current support for TCP/IP), and such supercomputers as Cray XT and IBM Blue Gene. ==========================================

In the last year, we have been developing a reversible (time travelling) debugger (URDB: Universal Reversible Debugger; urdb.sourceforge.net ) that adds reversibility gdb, the Matlab debugger, the perl debugger (perl -d), and the python debugger (pdb). The foundation is the support of DMTCP for checkpointing entire debugging sessions (e.g.: gdb and target process). This research platform opens up many exciting possibilities, such as a program that can communicate with its past self and make decisions. One of the novelties implemented today is the ability to do a binary search on the lifetime of a program and stop at a point in time where a given expression changed from a “good value” to a “bad value”. This allows a user to run until an error is detected, then construct an expression representative of the underlying fault that caused the bug, and then moving quickly to the point in the lifetime of that process where the expression changed from a good value to a bad value.

There are two areas of interest:

1) Database Middleware – Techniques for fault-tolerance query processing in mobile environments, where data sources themselves are mobile hosts.

2) Energy-Efficient, Data Intensive Data Management for the Clouds – Novel DBMS architectures designed for Cloud environments, that can scale to thousands of hosts and minimize the amount of energy their consume during query processing.

The Community Grids Laboratory, led by Prof. Geoffrey Fox, is looking for researchers to lead projects in Cloud computing (both systems design and large scale parallel application development), parallel computing, message-oriented middleware, and experience with streaming systems of all types (including sensor Webs and audio-video systems).

For a snapshot of lab activities, see our presentations and publications at http://grids.ucs.indiana.edu/ptliupages/presentations and http://grids.ucs.indiana.edu/ptliupages/publications/. We have a distinguished publication and funding record in parallel computing, Cloud computing, the application of Web technologies to scientific problems, and publish/subscribe systems.

The Community Grids Lab is part of the recently refunded Pervasive Technology Institute. We work closely with the Indiana University Research Technologies division, which provides world-class scientific computing infrastructure. Lab facilities are located in the IU’s new Incubator facility.

High performance computing, MPI, generic programming, C++, scientific computing, numerical analysis, parallel programming, large-scale systems, computational photography

My work at Google focuses on improving the programmability and performance of web applications delivered over the internet. Recent years have seen an explosion in size and scope of web applications to the extent that they are now straining the capabilities of today’s web execution frameworks. I am investigating the evolution of these frameworks to better handle greater client-side compute requirements, scalability, programmability, and security.

Understanding the structure of interacting black hole spacetimes, computational black hole collisions, prediction of gravitational wave signals, data reduction for gravitational wave detectors, scalable computing and computational scaling infrastructure, information security.

I seek to develop programming systems, i.e., languages, libraries, compilers, and development tools and environments, that are (a) natural and easy to use, (b) efficient to run (particularly over very large data and many machines), and (c) actually used in practice by real developers getting real work done. Some of my recent work at Google includes libraries and run-time systems supporting easy yet efficient massively data-parallel pipeline computations, and data synchronization infrastructure for easier development and deployment of web-based applications.

Fault Tolerance, Reliability, Computer Architecture, Power management, Performance and power analysis, Soft Errors, System Design, Low power circuits, system design methods