Page not found

International HPC Summer School 2015

The sixth International Summer School on HPC Challenges in Computational Sciences is now accepting applications for this summer’s session, which will be held June 21-26, 2015, in Toronto, Canada. Graduate students and post-docs from Canada, Europe, Japan, and the U.S. are invited to apply. The...

Modeling Efficient Organic Photovoltaic Cells

<h4>Modeling Efficient Organic Photovoltaic Cells</h4><p>The quest for higher efficiency organic photovoltaic cells (OPVs) is dependent on furthering the current understanding of fundamental device physics. At the heart of OPV operation is the diffusion of bound electron-hole pairs, or excitons, and their dissociation into free charge carriers at a donor-acceptor interface. The characteristic length an exciton can diffuse is called the diffusion length (LD). Knowledge of LD for OPV active materials of interest allows for intelligent device design where no excitons are generated more than a diffusion length away from a donor-acceptor interface. There are currently no methods available for the direct measurement of LD, instead one-dimensional models describing exciton motion with the diffusion equation are used to connect a value of LD to experimental data. Further complicating matters, not all device designs can be understood by the simple 1-D diffusion equation. This project seeks to create three-dimensional models of exciton diffusion at the molecular instead of device level. By employing the Monte Carlo method, exciton diffusion at the molecular level can be modeled by a series of energy transfer events between molecules instead of as diffusion across an active layer. The rates of energy transfer events are described by photophysical parameters which are easily obtainable for the OPV active materials of interest. The better understanding of exciton diffusion enabled by this work will guide the development of future high-efficiency OPVs.</p>
Group name: 

Critical Phenomena in Topologically Ordered Systems

<h3 class="red">Critical Phenomena in Topologically Ordered Systems</h3><p>The aim of this research project is to improve understanding of certain types of phase transitions that cannot be described by the standard (Landau) paradigm typically used to study phase transitions across which a symmetry is broken. Specifically, these researchers are studying a family of lattice Hamiltonians which can realize transitions between phases of differing topological order. Since topological order is not described by any local order parameter, the analytical techniques applicable in the symmetry-breaking case fail, and numerical studies are crucial in understanding these systems&#39; behaviors. Simultaneously, they are exploring the (often complex) phase diagrams of these systems. The primary tools currently employed for these studies are high-order perturbative series expansion methods, and Lanczos exact diagonalization.</p><p>Return to this PI&#39;s <a href="">main page</a>.</p>
Group name: 

This two part tutorial will first introduce you to the concept of interactive high performance computing, as distinct from batch computing. We will cover the Citrix (Windows) and NICE EnginFrame (Linux) interactive computing environments hosted by MSI.

New Director for MSI

MSI is very happy to report that Jorge Vinals has accepted the offer to become the next MSI Director. He will also become a professor of physics. Jorge will assume the directorship on August 1, 2010. Jorge is currently the director of CLUMEQ, a Canadian Supercomputing Center led by McGill...

Analysis of Performance Determinants in NYC Taxicab Industry

<h4>Analysis of Performance Determinants in NYC Taxicab Industry</h4><p><span style="color: rgb(51, 51, 51); font-size: 14px; background-color: rgb(255, 255, 255);">The project explores NYC Taxi and Limousine Commission data on yellow cab trips in 2013. The principal question is to what extent the locus of rent creation resides with individual drivers and/or taxi fleets and leasing agents. The study employs various regression techniques with multidimensional categorical variables to separate the effects via variance and higher moments decomposition. MSI resources are required to estimate the regressions with tens of thousands of dummy variables on the set of 170 million observations.</span></p><p><span style="color: rgb(51, 51, 51); font-size: 14px; background-color: rgb(255, 255, 255);">Return to this PI&#39;s <a href="">main page</a>.</span></p><p>&nbsp;</p>
Group name: 

High-Performance and Big Data Research

<h3 class="red">High-Performance and Big Data Research</h3><p>This group&#39;s research during 2015 focused on the development of parallel shared-memory graph partitioning, ordering, and clustering algorithms that use the multilevel paradigm. Graph partitioning is used widely for parallel task scheduling and data distribution. Graph ordering is used reducing the amount of computation and memory required for sparse direct numerical methods. Graph clustering is a widely used technique for discovering relationships between data points by creating groups of unconstrained size with high internal connectivity. Access to MSI&#39;s HPC resources has been critical in the development of these algorithms as evaluating the scalability of the algorithms requires machines with a large number of compute cores, and many of the graphs/matrices in these domains reach massive size, requiring large amounts of memory.</p><p>The group&#39;s work in 2016 focuses on developing hybrid shared/distributed memory codes that can effectively utilize compute architectures composed of many multicore nodes. This work will be an extension of the researchers&#39; past work on shared and distributed memory graph partitioning, ordering, and clustering. Part of this will include ensuring their methods scale to very large numbers of processing cores. These methods will be required for partitioning, ordering, and clustering problems on the next generation of large petascale and exascale machines, which will have millions of processing cores.</p><p>Return to this PI&#39;s <a href="">main page</a>.</p><p>&nbsp;</p>
Group name: 

Design, Modeling, and Control of a HCCI Free Piston Engine

<h4>Design, Modeling, and Control of a HCCI Free Piston Engine</h4><p>A Free Piston Engine (FPE) is a device where the linear motion of the piston is not constrained by a crankshaft. The advantage of this setup is the ability to run with varying compression ratios, thus enabling multi-fuel capability and Homogeneous Charge Compression Ignition (HCCI). The nature of this device requires precise control over the fuel, air, and piston motion. To enable this, a deep understanding of the constituent processes and hardware subsystems is necessary. One of these processes is the in-cylinder mixing between fresh air/fuel mixtures and hot, residual exhaust gas. CFD analysis using FLUENT is used to characterize the behavior of in-cylinder fluid motion. CHEMKIN is used to simulate the chemical reaction inside the combustion chamber. The power generated by FPE can be captured and transformed to fluid power or electrical power using a linear pump or alternator. For the linear alternator, to find the induced voltages and currents and also the opposing forces, it is necessary to model its magnetic circuit. To do this Maxwell-2D was employed and 2D axisymmetric analysis of the linear alternator was performed. Based on the coil flux linkages data obtained from this analysis, a Simulink model was developed using MATLAB which makes it easier to integrate the linear alternator model into FPE model and perform a thorough analysis of the device.</p><p><span style="font-size: 14px; line-height: 1.5;">A bibliography of this group&rsquo;s publications is attached.</span></p><p>Return to this PI&#39;s <a href="">main page</a>.</p><p>&nbsp;</p>
Group name: 

Brain Science

<h3 class="red">Brain Science</h3><p>The Brain Sciences Center is an interdisciplinary research institute and training center that focuses on the mechanisms underlying the active, dynamic brain in both health and disease. The Center&#39;s neuroinvestigators collaborate on a wide variety of research studies including: schizophrenia, Alzheimer&#39;s Disease, alcoholism, mechanisms of cognitive function, memory and learning, control of movement, and musical analysis of brain signals. These investigators represent a cross- section of academic disciplines including biophysics, clinical psychology, cognitive psychology, electrical engineering, music, neurology, neurophysiology, neuroscience, psychiatry, radiology, scientific computation, and statistics. Advanced technology is essential to tackling the complex nature of the brain, so Center investigators must employ a multitude of functional neuroimaging and other leading-edge research methods in their work including: magnetoencephalography (MEG), functional magnetic resonance imaging (fMRI), large-scale computer modeling, neurophysiology, sonification, experimental psychology, single cell recordings, and statistical analysis.</p><p>Return to this PI&#39;s <a href="">main page</a>.</p>
Group name: 

September 2011

A "jet in cross-flow" is a jet of fluid that exits an orifice to interact with the surrounding fluid that is flowing "across" the orifice. The term "high-speed jets in crossflow" implies that either the jet or the crossflow is supersonic. These are very complex flows, as shown in the image above...