The research activities of the CCT are organized into five Focus Areas:
- Core Computing Sciences
- Coast to Cosmos
- Material World
- Cultural Computing
- System Science and Engineering
These are broad, and sometimes overlapping, areas where faculty from diverse departments (Mathematics, Computer Science, Physics, Civil Engineering, Oceanography and Coastal Sciences, Petroleum Engineering, Mechanical Engineering, Electrical and Computing Engineering, Music, Business, etc.) collaborate in multidisciplinary projects. Our REU students will learn how to use some of the nation's largest supercomputers, may participate in the setup and management of large-scale simulations, and may take on an important role in the analysis and visualization of the simulation results.
Below, you can click on each focus area to learn more about the research, faculty involved and the projects in which you can participate.
Core Computing Sciences
The Core Computing Sciences focus area covers research in areas at the heart of computational science: computational mathematics, frameworks
and application toolkits, middleware and system level software, data and network middleware, scientific visualization and human-computer
interfaces, and portals and science gateways. Core Computing Science includes faculty in computational mathematics (Blaise Bourdin,
Susanne Brenner (lead), Li-yeng Sung, Shawn Walker, Xiaoliang Wan, Hangchao Zhang), distributed systems, network applications, computational frameworks (Jian Tao, Frank Löffler), scientific visualization (Bijaya Karki).
Some possible research areas for undergraduate research projects are:
Computational Frameworks [Mentors: Steve Brandt and Frank Löffler]. The computational frameworks group at LSU develops the Cactus Framework, a parallel computing environment used for large scale simulations in numerical relativity, coastal science, petroleum engineering and other scientific domains. Cactus is a modular framework where different modules contribute different computational capabilities, such as providing a coordinate system, a black hole evolver, or a method for writing efficient data. In addition to supporting a range of important scientific applications, Cactus has been used to demonstrate how new computational technologies, such as grid computing or Web 2.0, can be integrated into scientific workflows to provide new capabilities. REU students working in this group will learn how modern scientific software is developed and supporting using a range of collaborative and development tools. Potential student projects include developing tools to provide verification of data, including metadata into Cactus modules so that simulations can be self-described, and developing applications for visualization and monitoring of simulations.
Numerical Analysis of Liquid Crystals [Mentor: Shawn Walker]. Liquid Crystals (LCs) are abundant in modern technology, such as LCDs (displays). Research into LCs continues to increase because of increasing numbers of applications. LCs are easily actuated by electricity, magnetism, heat, light, etc., leading to new technological innovations such as highly tunable lasers, smart films/windows, and LC elastomers. Indeed, there is growing interest in the soft matter community to create colloidal particles with a designable valence (i.e. mesoatoms), which is an important goal in material science. Thus, LCs are a potential tool to creating new materials, as well as "switchable" materials.
Projects here would involve modeling, simulation, and analysis of liquid crystals coupled to electric fields, including the effects of colloidal particles, and possibly controlling the arrangements or shapes of particles. This would also involve learning and developing the necessary numerical tools (such as finite element methods) as part of the project.
Moving Contact Lines and Surface Tension Control of Droplets on Substrates [Mentor: Shawn Walker]. The development of engineered substrates has progressed to a very advanced level, which allows for interesting dynamics (and control) of droplets on these substrates. Applications are in the realm of micro-fluidics, such as lab-on-a-chip and droplet lenses that use electro-wetting. A recurring theme in this area is moving contact lines (i.e. the motion of the 3-phase contact line where the phases liquid, solid, and gas meet) and how to deal with them.
Projects would involve learning and developing finite element methods to capture moving contact lines as well as the optimal control of substrate surface tension to direct the motion and shape of droplets in a desired way.
Forensic Data Restoration and Repair [Mentor: Shane Li]. In this project, you will develop a graphics and 3D modeling system to facilitate the restoration and repair of forensic geometric evidence. You will operate a 3D scanner to digitize fragmented physical objects into digital models, then you will develop/use 3D geometric modeling techniques to reassemble and repair the damaged model. If the final restoration is accurate, the physical object's prototype can be generated through 3D printing.
Medical Image Analysis for Identification, Segmentation, and Tracking of Abnormalities [Mentor: Shane Li]. This project will develop image processing and analysis techniques to help identify abnormal tissues or tumors from CT or MRI scan images. You will develop/use an effective image analysis computer program to suggest suspicious abnormality, and develop an image editing program to allow users to visualize and refine this identification.
Autonomous Robotic Exploration and Mapping [Mentor: Shane Li]. This project aims to design a vision system for an autonomous robot or vehicle so that it can "see" the surrounding environment to guide its exploration in an indoor environment. We will use an iRobot, which carries a webcam or a Kinect, and develop an image processing program to simultaneously construct a map of the environment and localize itself in this map. This could lead to autonomous robotic exploration and inspection to complicated or hazardous environments.
Development and Application of AtomVis System [Mentor: Bijaya Karki]. AtomVis is an interactive atomistic visualization system based on space-time multi-resolution approach. The system has widely been used for visualizing atomic position-time series generated by molecular dynamics simulations to gain insight into the structural and dynamical behavior of the simulated material system. In this REU project, we will further develop the system for visualizing atomic magnetization. The simulated data thus consist of time-varying magnetization values on per atom level. The magnetization property is a scalar quantity (can be -ve or +ve value). So the main idea is to visualize/render magnetization of each atom, and how it evolves over time. We will improve/implement various features related to data loading, data processing, data rendering, and user interfaces. The intended development involves the use of C++, OpenGL, GLSL, GLUI/Qt, HDF5 with which the original AtomVis system has been implemented. We will also make application case studies by visualizing the atomic position and magnetization data for simulated silicate materials.
- The Core Computing Sciences focus area covers research in areas at the heart of computational science: computational mathematics, frameworks and application toolkits, middleware and system level software, data and network middleware, scientific visualization and human-computer interfaces, and portals and science gateways. Core Computing Science includes faculty in computational mathematics (Blaise Bourdin, Susanne Brenner (lead), Li-yeng Sung, Shawn Walker, Xiaoliang Wan, Hangchao Zhang), distributed systems, network applications, computational frameworks (Jian Tao, Frank Löffler), scientific visualization (Bijaya Karki).
Coast to Cosmos
Researchers in the Coast to Cosmos focus area seek better methods to predict the behavior of natural and engineered systems, at scales ranging from nanometers to millions of kilometers. Coast to Cosmos embraces a wide range of disciplines, including faculty doing research in numerical relativity (Steve Brandt, Peter Diener, Jorge Pullin, and Frank Löffler), coastal modeling (Q. Jim Chen, Lead, Steve Brandt, Scott C. Hagen, Haosheng Huang, Dubravko Justic, Jun-hong Liang, Giulio Mariotti, Celalettin Özdemir, Robert Twilley, Frank Tsai, Zuo "George" Xue), petroleum engineering (Mayank Tyagi, Karsten E. Thompson), Chemical Engineering (Krishnaswamy Nandakumar, Karsten E. Thompson) and computational fluid dynamics (Q. Jim Chen, Krishnaswamy Nandakumar, Celalettin Özdemir, Mayank Tyagi). The diverse problem space, high performance computing requirements, and attention to real-world problems provide opportunities for undergraduate researchers to address challenging, relevant, cutting edge problems in the physical and earth sciences, engineering, and applied mathematics.
Computational Astrophysics: Astrophysical simulations of black hole and neutron star systems. [Mentors: Steven Brandt, Peter Diener, Frank Löffler] The numerical relativity group at LSU is researching astrophysical systems such as the collisions of black holes and neutron stars that produce gravitational waves that may be detectable by ground-based detectors such as LIGO. The group uses numerical models to simulate these systems, solving Einstein's Equations on some of the largest academic supercomputers in the USA with codes written using the Cactus Framework. REU students working in this group will be immersed in a vibrant, interdisciplinary and international collaboration. Potential student projects include running and analyzing simulation codes, investigating the efficiency of the simulations to reduce the time taken to run them, extending the Cactus Computational Toolkit to make better use of accelerators, and visualizing output data to display the resulting gravitational waves emitted by the astrophysical systems. The group also performs numerical simulations of loop quantum cosmology to study the nature of the big bang.
Numerical Simulations of Loop Quantum Gravity [Mentors: Peter Diener, Jorge Pullin, Parampreet Singh]. The project consists in developing numerical treatments for situations with high symmetry in loop quantum gravity. This includes cosmological models and also models with spherical symmetry. Loop quantum cosmology predicts in simple models that the Big Bang is replaced by a "bounce" from a previous universe. A major research focus today consists in validating these predictions of simple models in situations of increasing complexity, hence the need for numerical simulations. As similar prediction concerns the singularity that occurs inside black holes, it gets eliminated through a tunneling to another region of the universe. Here it is also desirable to confirm the prediction in a simulation in which a black hole is formed from collapsing matter. Undergraduate projects could consist in running codes and evaluating their efficiency and help in the development of new codes.
Numerical Simulations of Tsunamis, Storm Surges, Ocean Waves and Fluid-Structure Interactions [Mentors: Q. Jim Chen, Steve Brandt, Kelin Hu, Jian Tao]. More than 53% of the U.S. population lives in coastal regions, which are home to a wealth of natural and economic resources. Many megacities with more than 8 million people each are located in the coastal zone worldwide. Coastal flooding caused by geo-hazards (earthquakes and landslides) and atmospheric-oceanic hazards (tropical cyclones and extra-tropical storms) has killed millions of people living in coastal regions in the last two hundred years. This threat to humanity increases as coastal populations in the world continue to grow and the sea level continues to rise. The coastal modeling group in CCT specializes in the development and application of state-of-the-art numerical models for the study of coastal inundation hazards. Computer models, such as CaFUNWAVE solving Boussinesq-type equations, Delft 3D, a coastal modeling suite, and Computational Fluid Dynamics codes based on OpenFoam have been utilized to simulate the generation, propagation and inundation processes of tsunamis and storm surges as well as the impact of ocean waves on natural and built infrastructure. REU participants will work closely with the research team on model optimization, model setup, deployment on HPC systems and analysis of simulation results of real-world inundation hazards.
Mathematical and computational modeling of multiphase flow in porous media for improved oil recovery [Mentor: Mayank Tyagi]. Flow of oil, gas and water in porous media is typically modeled using the relative permeability concept. Understanding these flow mechanisms is crucial for improved oil recovery techniques for both on-shore as well as offshore wells. However, the relative permeability approximation is inappropriate in several scenarios leading to unphysical results. As a part of this research effort, mathematical expressions for the relative permeability functions will be derived from first principles using multiphase flow details at the appropriate pore-scales of the media.
Cyberinfrastructure-enabled parallel Reservoir Simulator [Mentor: Mayank Tyagi]. Reservoir simulation is an integral part of asset management and production schedule decisions. With increase in computational resources and network capabilities, it is now possible to perform detailed simulations to understand the influence of uncertainties in model parameters as well as construct models based on inversion techniques that provide better predictions when compared to the observed data. This research project involves design of a scalable parallel reservoir simulator and efficient workflow managers using cyberinfrastructure and high performance computing platforms.
Geothermal Reservoir Simulation [Mentor: Mayank Tyagi]. Geothermal energy is becoming an attractive option primarily due to its abundance and possibly less detrimental environmental impacts as compared to conventional non-renewable resources such as coal and natural gas. However, there are several challenges in extracting heat efficient from the geothermal systems such as hot dry rock or enhanced geothermal systems including fracture characterization, understanding of flow and heat transfer in fractures and efficient energy generation from the extracted heat. This research project involves both experiments and simulations about fractures in rocks and the fluid flow through these fractures.
Simulation of inertial flows using Lattice Boltzmann Method for high productivity gas wells [Mentor: Mayank Tyagi]. Lattice Boltzmann Method (LBM) has matured as an alternate simulation technique to classical Navier-Stokes equations in the Computational Fluid Dynamics area. In particular, LBM is fundamentally sound in the scenarios where continuum approximation breaks down. In this project, the effect of grain geometries on the inertial effects in single-phase flow for high productivity wells is computed using the LBM on high performance computing (HPC) platforms. Images of porous media are obtained using X-Ray imaging or similar techniques. LBM has several attractive features including the near-ideal weak scaling on HPC platforms. Macroscopic properties such as permeability and beta-factor to account for inertial flows in porous media are computed from the results of LBM simulations.
Computer models of chemical processes [Mentor: Krishnaswamy Nandakumar]. Dr. Nandakumar's research group uses computer models of chemical processes to understand the basic physics and chemistry taking place inside the process equipment and use that to innovate the process equipment design methodologies. Specific projects that are available this year include (1) "CFD modelling and design of novel bioreactor design", (2) "CFD modelling and design of fractal mixers", (3) "CFD modelling and design of granular flow in cavities". More about our group can be found at http://che.epic.lsu.edu/
- Researchers in the Coast to Cosmos focus area seek better methods to predict the behavior of natural and engineered systems, at scales ranging from nanometers to millions of kilometers. Coast to Cosmos embraces a wide range of disciplines, including faculty doing research in numerical relativity (Steve Brandt, Peter Diener, Jorge Pullin, and Frank Löffler), coastal modeling (Q. Jim Chen, Lead, Steve Brandt, Scott C. Hagen, Haosheng Huang, Dubravko Justic, Jun-hong Liang, Giulio Mariotti, Celalettin Özdemir, Robert Twilley, Frank Tsai, Zuo "George" Xue), petroleum engineering (Mayank Tyagi, Karsten E. Thompson), Chemical Engineering (Krishnaswamy Nandakumar, Karsten E. Thompson) and computational fluid dynamics (Q. Jim Chen, Krishnaswamy Nandakumar, Celalettin Özdemir, Mayank Tyagi). The diverse problem space, high performance computing requirements, and attention to real-world problems provide opportunities for undergraduate researchers to address challenging, relevant, cutting edge problems in the physical and earth sciences, engineering, and applied mathematics.
The Material World focus area covers the modeling and understanding of all materials, including biological, chemical, and condensed matter
materials. Faculty from Biological Sciences, Electrical and Computer Engineering, Chemistry, Chemical Engineering, Civil and Environmental Engineering, Computer Science, Geology and Geophysics, Mathematics, Mechanical Engineering, and Physics (Dana
Browne, Michal Brylinski, Les Butler, Bin Chen, Theda Daniels-Race, Francisco Hung, Mark Jarrell (lead), Bijaya Karki, Michael Khonsari, Robert Lipton, Dorel Moldovan, Juana Moreno, Dimitris Nikitopoulos, Georgios Veronis, George Voyiadjis, Grover Waldrop, Jianwei Wang) belong to this area. A
few examples of research projects are:
Strongly Correlated electronic systems [Mentors: Mark Jarrell, Juana Moreno]. This project is motivated, in part, by a variety of complex emergent phenomena, including high-temperature superconductivity and quantum criticality. These correlated materials form the basis of future high-tech devices and their proper theoretical understanding is paramount for technological progress. Interest is also driven by the rapid development of new computational approaches to simulate the many-body problem.
Design of metallic photonic crystal structures for thermophotovoltaics [Mentor: Georgios Veronis]. In this project, we will exploit emerging opportunities in the area of metallic photonic crystals for Thermophotovoltaic (TPV) solar cells. The success of this project may lead to a fundamental breakthrough in the area of solar energy conversion, by paving the way towards efficiencies approaching the thermodynamic limit.
Molecular simulation studies of organic salts inside nanopores [Mentor: Francisco Hung]. This project will provide students a hands-on, introductory experience in molecular simulations, and how it complements experimental work in nanotechnology. Students will perform short molecular dynamics simulations of ionic liquids confined inside carbon with nanopores of different sizes and shapes. Ionic liquids (ILs) are organic salts that are in liquid state near room temperature. ILs have been proposed as alternative electrolytes for energy storage in electrochemical double-layer capacitors, and as base materials for the synthesis of non-toxic nanoparticles for biomedical applications. We aim at understanding how changes in the pore size and shape affect the molecular-level properties of the confined organic salt. These molecular-level properties affect macroscopic variables, such as capacitance, resistance, magnetic and optical properties.
Accelerated Physics and Chemistry codes using GPU (Graphics Processing Unit) [Mentors: Mark Jarrell, Juana Moreno, Ram Ramanujam]. The student will have access to a new GPU cluster and would work with faculty in Physics and Chemistry to implement simple GPU accelerated calculations in OpenCL, CUDA, or with code written with PGI compilers.
LAMMPS on GPUs for Biomaterial Transport [Mentors: Dimitris Nikitopoulos, Dorel Moldovan, Ram Ramanujam]. Graphics Processing Units (GPUs) are pervasive in computer gaming environments but also provide a relative inexpensive platform for fast highly parallelized scientific computation. The objective of this REU project is to go through the necessary implementation steps to run LAMMPS on GPUs, and solve a simple test problem of transport of a bio-molecule in an aqueous environment. Since implementation of LAMMPS, which is a very widely used code for Molecular Dynamics Simulation, on GPUs is still in the development stage, performance and other technical issues still need be evaluated. Part of the effort of the student in this project will be directed to collecting data facilitating this evaluation based on the test problem. The REU student will have the opportunity to interact with a graduate student and a post-doctoral associate in addition to faculty and some members of the LAMMPS development team at Sandia National Labs. The bio-molecule transport test problem is highly relevant to bio-medical applications such as drug delivery and bio-analytical diagnosis.
Molecular dynamics simulation study of self-assembly of Span 80 micelles [Mentor: Dorel Moldovan]. Surfactant molecules are important in a large variety of processes such as: biological, as in carrier structures of molecules across cell membranes; commercial, as in detergents and stain removers; and food industry, as in emulsifiers. In this project, using a realistic all-atom inter atomic interaction model, we will perform large-scale molecular dynamics simulations of Span 80 (sorbitan monooleate) surfactant self-assembly in water at concentrations above the critical micelle concentration. The ultimate goal of this study is to develop an atomistic understanding of the mechanism of surfactant aggregation into micelles. Throughout the duration of the project the students will be introduced to the basics of using GROMACS simulation package and the visualization software VMD.
Computer-aided drug design [Mentor: Michal Brylinski]. Currently, drug development programs routinely include a strong computational component. Such cost effective computer-aided drug discovery is especially appealing in an academic setting, which lacks industrial capabilities for high-throughput screening of large compound libraries; however, it is not trivial and poses several formidable challenges. One of the most promising modeling techniques used in structure-based drug design is virtual screening by molecular docking. In this project, REU students would have an opportunity to learn about current trends and approaches to drug docking and binding affinity prediction. Moreover, they would participate in the development of novel algorithms for virtual screening that offer improved accuracy and effectively utilize modern heterogeneous high-performance computing platforms, e.g. these powered by massively parallel graphics processors.
Sign Learning Kink Path Integral Method [Mentors: Frank Löffler, Juana Moreno, Mark Jarrell]. Exact solution of the Schrödinger equation for atoms, molecules, and materials is hindered by numerical difficulties associated with the notorious "sign problem". This project will use a novel algorithm that "learns" as a Monte Carlo simulation progresses and eventually the learned information is used to overcome the sign problem. The summer project will focus on exactly solving the Schrödinger equation for a number of molecules. The student will learn how the computer can be used to implement modern methods in Chemistry and become part of an interdisciplinary team of researchers.
3D Printing and X-ray Interferometry/Tomography [Mentor: Les Butler]. Butler and his group are using X-ray interferometry/tomography coming from a newly-constructed systems to examine 3D printing. For the REU student, the project offers a tangible object combined with advanced imaging methods and data analysis. For example, the Stanford Bunny is easily printed in acrylonitrile butadiene styrene (ABS) or PLA. From the X-ray absorption image, questions develop about the quality of the 3D print (e.g., what are the parameters that affect it?), which leads to a discussion of point clouds and iterative closest point analysis. From the X-ray dark-field image (essentially a small-angle scattering image), we can see print layer defects, and this leads to a discussion of printhead tracking and curvature. How do we compare the stereolithography (STL) print file with the tomography data volumes?
ABS printed Stanford Bunny, X-ray absorption, and X-ray dark-field image. The last is a unique imaging modality available with X-ray grating-based interferometry. The dark-field signal is ascribed to print defects, namely micron-sized voids between print layers.
Projects: Strong-field ionization rates in molecules; first-principles X-ray absorption in insulators; and nonlinear optical response properties [Mentor: Kenneth Lopata]. Strong-field and above ionization excited states in molecules, materials, and at interfaces underpin a wide range of important physical processes such as spectroscopy, photochemistry, and high harmonic generation, to name only a few. Modeling these processes can be challeging for density functional methods, however, due to deficiencies in the exchange-correlation function and finite basis setlimitations.
We are currently working towards extending time-dependent density functional theory (TDDFT) to capture above ionization states and dynamics using tuned range-separated hybrid functionals in conjunction with non-Hermitian real-time propagation.
- The Material World focus area covers the modeling and understanding of all materials, including biological, chemical, and condensed matter materials. Faculty from Biological Sciences, Electrical and Computer Engineering, Chemistry, Chemical Engineering, Civil and Environmental Engineering, Computer Science, Geology and Geophysics, Mathematics, Mechanical Engineering, and Physics (Dana Browne, Michal Brylinski, Les Butler, Bin Chen, Theda Daniels-Race, Francisco Hung, Mark Jarrell (lead), Bijaya Karki, Michael Khonsari, Robert Lipton, Dorel Moldovan, Juana Moreno, Dimitris Nikitopoulos, Georgios Veronis, George Voyiadjis, Grover Waldrop, Jianwei Wang) belong to this area. A few examples of research projects are:
The Cultural Computing focus area includes faculty from Art, Music, Business, and Mass Communication (Brygg Ullmer(lead), Jesse Allison, Stephen Beck, Chris Branton, Rudy
Hirschheim, Robert Kooima, Xin Li, Hye Yeon Nam, Derick Ostrenko, Gabriele Piccoli, Lance Porter, Susan Ryan, Sonja Wiley-Patton). The Laboratory for
Creative Arts and Technology has remained an entity within Cultural Computing, with dedicated laboratories for audio and video. The
Cultural Computing focus area is responsible for the Red Stick Animation Festival, which has been
held in Baton Rouge yearly since April 2005. This focus area also leads the AVATAR initiative, and is
developing new curricula for digital media. Possible projects is this area are:
Tangible Visualization [Mentor: Brygg Ullmer]. The tangible visualization group is focused on the design and deployment of new kinds of physical interaction devices and associated software systems to simplify, strengthen, and extend computer visualizations, especially in collaborative environments.
Laptop Orchestra [Mentor: Steve Beck]. Each player manages the activities of his laptop computer, and uses the laptop as a virtual musical instrument. Instruments can be performed by physical gestures using keystrokes, mouse clicks, sweeps of the trackpad, or through a variety of game interfaces. Virtual instruments can also talk to each other over a wireless network, so that the control over one laptop influences other laptops in the network. In the Laptop Orchestra, students will write their own computer code in order to compose a music piece and work in teams to perform these pieces.
Mobile Music [Mentor: Jesse Allison]. Music creation on mobile devices is a broadening field of exploration within computer music. New instrument or interactions with sound that can engage in gesture based control, interactive moving image representations of music and sound, sensors, geo-location, augmented reality, networking communication between performers, and social computing networks are broadening our understanding of what an instrument, performer, and ensemble may be. Interaction designers, programmers, musicians, composers, and researchers in many fields collaborate to create musical instruments, sonic environments, explore acoustic ecology, or any other aspect of sound and music related to mobile computing.
Value extraction from financial digital data streams (DDS) [Mentor: Gabriele Piccoli].Equipping decision makers with relevant and timely information is an enduring problem in organizations. Our group has been daily harvesting earning calls transcripts from traded companies and managed a continuously growing database. The application of text mining and topic modelling techniques to this textual DDS allowed us to elaborate a new confidence index for describing and perhaps predicting economic trends. The next phase in this project consist of extracting and visualizing relevant decision-making data from this DDS. Interested students will help with the design and testing of the front-end, contributing to prioritizing decisions about relevant content to display and to the design of the user interface.
Measuring tourist attraction service levels [Mentor: Gabriele Piccoli]. Online reviews are increasingly guiding customer's decisions in all aspects of life - from the purchase of product and services, to restaurant selections, to travel and tourism. But online reviews can also serve as a service quality monitoring system for businesses and organizations. Our group has been studying how to use the digital data stream of online reviews as a service quality monitoring system. In this project we intend to apply our approach to tourist attractions in order to empower state and local officials with a direct and simple quality monitoring system.
High-performance Interactive Visualization and Electroacoustics [Mentor: Derick Ostrenko]. High-performance Interactive Visualization and Electroacoustics or HIVE is a 448-core cloud based platform currently being built at the CCT. HIVE is a tool for collaboration between artists and developers seeking to build new forms of digital expression using high-performance computing. Projects created using HIVE can take many forms including web applications, games, animations, audio, or a combination thereof. Given its ease of scalability and provisioning HIVE is also a sandbox for defining new practices in cluster-based rendering, sonification, web graphics, and interactive media.
Interactive Image Systems [Mentor: Robert Kooima]. The Interactive Image Systems group designs and implements real-time 3D visualization software. Recent projects have produced tools for multi-giga-pixel planetary data rendering, photo-realistic planetary rendering algorithms, and omni-stereoscopic virtual reality image capture and display. Ongoing projects pursue omni-stereo video capture and encoding, as well as a variety of topics related to perception in virtual reality with head-mounted displays.
- The Cultural Computing focus area includes faculty from Art, Music, Business, and Mass Communication (Brygg Ullmer(lead), Jesse Allison, Stephen Beck, Chris Branton, Rudy Hirschheim, Robert Kooima, Xin Li, Hye Yeon Nam, Derick Ostrenko, Gabriele Piccoli, Lance Porter, Susan Ryan, Sonja Wiley-Patton). The Laboratory for Creative Arts and Technology has remained an entity within Cultural Computing, with dedicated laboratories for audio and video. The Cultural Computing focus area is responsible for the Red Stick Animation Festival, which has been held in Baton Rouge yearly since April 2005. This focus area also leads the AVATAR initiative, and is developing new curricula for digital media. Possible projects is this area are:
System Science and Engineering
This focus area addresses the challenges and opportunities for
advancing computer system concepts and design. This area includes faculty from Computer Science and
Electrical and Computing Engineering (Ram Ramanujam (lead), Hartmut Kaiser, David Koppelman, Jay Park).
The Advanced Computer Architecture Laboratory supports this area of research.
The core funded projects in this focus area are related to ParalleX
and Advanced Networking.
Parallex [Mentor: Hartmut Kaiser]. The Ste||ar research group is investigating advanced parallel computer architecture and programming environments to eliminate constraints and program petascale-class machines in ways that incorporate multiple elements effectively. In particular, ParalleX is designed to make areas such as synchronization, scheduling, manual data layout, and messaging more efficient.
Advanced Networking [Mentor: Jay Park]. The Advanced Networking group has worked on Future Internet and Cloud Computing by developing high speed network protocols and high performance computing applications. REU students can have hands-on experience of the state of art technology on computing and networking at the CRON 10Gbps network testbed (http://www.cron.loni.org).
Big Data research project with Hadoop over Supercomputers [Mentor: Jay Park]. Our projects address big data challenges with software frameworks (i.e., Hadoop and Girape), cloud computing and high performance computing clusters. For example, we develop software solutions to assemble genome sequencing data (De Novo Genome Assembly), data mining from peta bytes of data (twitter, facebook, etc.), and accelerate molecular dynamic simulations.
Cybersecurity: Targets [Mentor: David Koppelman]. The Cybersecurity Laboratory is a facility for testing methods of attack against computers, defensive measures against such attacks, and related issues. To facilitate such investigations the laboratory will develop a library of typical computer configurations, called targets. A target is essentially a snapshot of a real-life computer (referring to the mix of software and data, as well as how correctly everything was set up) at some moment in time. Lab users will check out a target (copy the target to one of the lab's machines), perhaps apply some defensive measures, and then attack it. An undergraduate can help with the identification and collection of these real-life targets, and with developing a friendly user interface for the library.
Compiler Research [Mentor: Ram Ramanujam]. The Tensor Contraction Engine (TCE) project is a domain-specific optimizing compiler developed to facilitate the rapid automatic generation of efficient parallel implementations from high-level mathematical specifications. Because of the high-level view, the TCE can perform optimizations that are extremely difficult or impossible once the computations are translated to a language such as C/C++/Fortran. The PLUTO and PrimeTile efforts enable a broader application of transformations, such as tiling and fusion, that were critical to performance optimization for tensor expressions in the TCE project. A significant advantage of the polyhedral compiler model used by Pluto is the integrated treatment of arbitrary compositions of transformations on (affine) imperfectly nested loops with multiple statements, and reasoning about their cumulative effects through algebraic cost models. REU students will benefit significantly from working on these projects.
- This focus area addresses the challenges and opportunities for advancing computer system concepts and design. This area includes faculty from Computer Science and Electrical and Computing Engineering (Ram Ramanujam (lead), Hartmut Kaiser, David Koppelman, Jay Park). The Advanced Computer Architecture Laboratory supports this area of research. The core funded projects in this focus area are related to ParalleX and Advanced Networking.