Program

We have an exciting program, built around the theme of ‘HPC Serving Society’. We’ll be exploring this theme through sessions focusing on a wide range of research areas that rely on advanced computing, including Bioinformatics, Big Data & Analytics, Computational Materials Science, Ocean & Atmosphere Modelling, Computational Chemistry, Data Cave Visualization, and more.



You can find the list of posters presented at the conference here.



Computational Chemistry

Computational chemistry uses efficient computer programs to calculate the structures of molecules and solids and simulate the of behaviour dynamical systems. It complements information available from experiments and provides insight into chemical phenomena, which in turn sometimes suggests additional experiments. In this session, speakers will describe new advances in the methods employed by computational chemists and a broad range of recent applications.


Plenary Speaker Axel D. Becke

Department of Chemistry, Dalhousie University

becke_axel_highres

Dr. Becke is the Killam Chair in Computational Science at Dalhousie University. He is an international leader in the development of novel quantum molecular theories and computational methods, and one of the most cited physical scientists in the world. He has received numerous awards; most recently the Theoretical Chemistry Award of the American Chemical Society.



A Density-Functional Theory for Covalent and Noncovalent Chemistry

Density-Functional Theory is an approach to computing the structures and properties of atoms, molecules, and solids that has, over the last three decades, become the most popular tool in computational physics and chemistry. Advances by Dr. Becke’s lab in the theory and its implementation are largely responsible for this popularity. Dr. Becke’s talk will outline the salient developments, and the role of computational resources in their success.

Andre Bandrauk

Département de chimie, Université de Sherbrooke

High Performance Supercomputer Simulations of Nonperturbative Laser-Molecule Interactions. Modern laser technology allows for the generation of ultrashort laser pulses with intensities exceeding the internal electric field in atoms and molecules. The interaction of such pulses with atoms and molecules leads to new highly nonlinear, non-perturbative regimes of laser-matter interaction requiring high level simulations based on large-dimensional spatio-temporal partial differential equations. The main spin-off of research in this new field, mainly theoretical, has been the design of new ultrashort pulses, the attosecond pulse, for the ultimate visualization and control of the “quantum” nature of the electron in molecules.

Alex Brown

Department of Chemistry, University of Alberta

Computations on Small to Large Molecules. A variety of computational techniques for determining geometries, bonding (electronic structure), and physical as well as photophysical properties of molecules is critical to interpreting and predicting modern experimental measurements and syntheses in chemistry. In this talk, Professor Brown will highlight his computations on small to large molecules to help understand chemical problems.

James Gauld

Department of Chemistry & Biochemistry, University of Windsor

High Fidelity Substrate-Assisted Catalysis in Aminoacyl-tRNA Synthetases. Aminoacyl-tRNA Synthetase (aaRS) enzymes are ubiquitous essential enzymes, having a central role in many physiological processes, in particular protein synthesis. Despite the fact that they are thought to be very ancient, they have been observed to have an outstandingly high fidelity. How this is achieved, however, is poorly understood.

Erin Johnson

School of Natural Sciences, University of California, Merced

A DFT-Based Approach to Modeling Dispersion Interactions. Density-Functional Theory (DFT) methods have attained widespread use in chemical and materials physics research due to their accuracy and efficiency. This talk will present a DFT-based approach to modeling dispersion interactions and survey a range of applications, including surface adsorption, chiral crystals, thermal contraction in polymers, and friction of graphene.

Jing Kong

Department of Chemistry, Middle Tennessee State University

High Performance Computation of Density Functional Theory. Modelling of molecules with density functional theory (DFT) is a major application area in high performance computing. In this talk, Dr. Kong will demonstrate how ten-times speed-up can be achieved for DFT calculations without incurring errors through developing new algorithms.

Nelaine Mora Diez

Department of Chemistry, Thompson Rivers University

Deuterium Isotope Effects on Acid-Base Equilibria at Ambient and Hydrothermal Conditions. Heavy water is used as a neutron moderator and coolant in CANDU reactors. Due to the presence of acidic species in the medium and the isotopic nature that alters the acidic character of these species by deuterium isotope exchange, the pipes that carry heavy water are prone to corrosion. While this corrosion can be minimized by applying chemical treatments that neutralize the medium, current operating procedures rely on data generated from light water under hydrothermal conditions since very little data is available in heavy water. Our research involves the computational determination of the pKa values of 73 weak organic and inorganic acids in both H2O and D2O at ambient and hydrothermal conditions using DFT and continuum solvation methods. The calculated difference of pKa between D2O and H2O at ambient conditions for the studied acids is in very good agreement with experimental values, which provides confidence in the calculated pKa difference at hydrothermal conditions. In addition, both pKa differences follow a very interesting pattern depending on temperature and pressure which is further investigated.

Jason Pearson

Department of Chemistry, University of Prince Edward Island

Chemical Data Open Storage System with Search Tools. HPC is an indispensable tool in the chemical sciences, among many other research fields. However, only pieces of the enormous amounts of valuable data generated are preserved through publication. Dr. Pearson is leading a joint undertaking with data specialists in the UPEI library and computational chemists at UPEI and MUN, to develop an open model system for storing and performing sophisticated searches of vast amounts of computational chemical data.

Christopher Rowley

Department of Chemistry, Memorial University

Modelling Biophysical Chemistry with High Performance Computing. Living things are comprised of a vast set of cellular components like proteins and membranes. The molecular level interaction of these components ultimately leads to the behaviour of these systems, however these interactions are so complex, computer simulations are needed to understand them. Professor Rowley’s group uses massively-parallel molecular dynamics study how molecules cross cell membranes, how proteins fold into their functional state, and how drugs react with their targets.

Dennis Salahub

Department of Chemistry, University of Calgary

Multiscale Modeling of Chemical Reactions in Complex Environments. Dr. Salahub will describe recent advances in modeling enzymatic catalysis, electron transfer between proteins and/or heavy oil upgrading, focusing on how several techniques, from quantum chemistry, to molecular dynamics, to stochastic network analysis, have to be brought to bear on the problem and on issues of HPC parallel processing.

Stacey Wetmore

Department of Chemistry & Biochemistry, University of Lethbridge

It was the Best of Times, It was the Worst of Times: A Day in the Life of your DNA. Despite containing all information necessary for life, our DNA can become damaged by exposure to external sources (X-rays, UV sunlight) or by natural processes (errors when DNA is copied). Since DNA repair has been correlated with the prevention of life-threatening illnesses, research in Professor Wetmore’s laboratory uses computer calculations to understand both how DNA is damaged and how this damage is repaired in our bodies.

And contributed talks from

Eric Bushnell, Dalhousie University

A computational study of the similarities and differences of dithiolene- and diselenolene-catalysts

Corey MacDonald, Dalhousie University

The Catalytic Formation of Leukotriene C4: A Tale of Two Orientations



Modelling the Atmosphere and Oceans

Understanding how the atmosphere-ocean system works and predicting how it will change over hours to weeks to decades is a major challenge facing the scientific community. Coupled with this is the growing need to develop practical applications, such as efficiently extracting offshore oil and gas, providing warnings of oceanic “dead zone” occurrences, and helping coastal communities adapt to climate change. This session will review progress and challenges in the development of the highly complex, and computationally demanding, numerical models that are used to make predictions and projections. Practical applications of the modelling will be described.


Plenary Speaker Gilbert Brunet

Deputy Director of Weather Science of the UK Met Office

Gilbert Brunet

Dr. Brunet is responsible for numerical models and their application in operational forecasting and consultancy. Prior to his leadership role at the UK Met Office Dr. Brunet was head of the Meteorological Research Division at Environment Canada. Dr. Brunet is an international leader in dynamical meteorology, numerical weather prediction and intraseasonal variability. He is also presently the Chair of the Joint Scientific Committee of the World Weather Research Programme of the WMO.


The Challenges of Weather and Environmental Prediction

Internationally, the increasing demand for accurate high-impact weather and environmental predictions is indisputable. It has led to significant investment in developing sophisticated mathematical and physics based models, high performance computing, high-speed telecommunication, remote sensing, ground-, space- and aircraft-based measurement technologies. Significant applied physics research and development challenges remain to be met before acceptable meteorological and environmental forecasts with increased economic and societal values can be produced worldwide from urban to planetary scales and all relevant time scales. An historical perspective and future challenges of this multi-scale and seamless prediction problem will be presented.

Gary Brassington

Centre for Australian Weather and Climate Research, Australian Bureau of Meteorology

Operational Ocean Forecasting – An HPC application. An ocean forecast system comprises two computationally intensive components: a high-resolution model and a data assimilation system. The system produces large data volumes requiring data storage with live access and server side processing. This talk will present the technical requirements for the Australian Bureau of Meteorology’s new global ocean forecast system.

Ricardo de Camargo

Departamento de Ciências Atmosféricas, Universidade de São Paulo

Using HPC in weather and climate studies. Dr. de Camargo’s talk will emphasize the relevance of HPC for weather and climate studies, considering the evolution of computational resources, as well as the dynamical representation of the atmosphere and the ocean on their particular frequencies of oscillation. Particularly in Brazil, homemade solutions worked very well for a while, considering their cost.

Fraser Davidson

Research Scientist, Fisheries and Oceans Canada

Developing Ocean Prediction Systems to deliver services: the CONCEPTS approach. Ocean Prediction involves melding oceanographic data and numerical models that represent our understanding of physical ocean circulation, temperature and salinity. This effort needs important team work in developing data gathering/aggregation systems, oceanographic data assimilation, ocean model configuration development as well as implementation of the operations and product dissemination. All elements of ocean prediction from research through development to operational stages demand significant high performance computing resources. This talk focuses on specific tasks in ocean prediction that rely on high performance computing clusters as well as supercomputers within the inter departmental “Canadian Operational Network of Coupled Environmental PredictTion Systems” (CONCEPTS) MOU activities.

Katja Fennel

Department of Oceanography, Dalhousie University

Effects of Anthropogenic Perturbations of the Global Carbon and Nitrogen Cycles. Professor Fennel will discuss how the fundamental physical and chemical properties of the ocean are being altered and how these changes will affect marine ecosystems. Numerical ocean models that accurately simulate physical, chemical and biological processes are key tools in her research. Given the prospect of global climate change, its anticipated impact on the global economy and our quality of life, and the important role of the ocean in the climate system an improved predictive understanding is urgently needed.

Richard Karsten

Department of Mathematics and Statistics, Acadia University

Mathematical Models of Ocean Circulation. In-stream tidal power is on the verge of a rapid development in Nova Scotia. Successful deployment of turbines requires a careful matching of the turbine to the tidal flow at the deployment site. In this talk, we’ll show how HPC resources are being used to accurately model tidal flows and optimize the design of future tidal farms.

Kevin Lamb

Department of Applied Mathematics, University of Waterloo

Modelling Internal Solitary Waves: Computational Challenges. Internal solitary waves in the ocean are short high frequency nonlinear waves that are formed from the nonlinear-dispersive evolution of the large-scale internal tide. Instabilities in these waves can resuspend bottom sediments and mix heat and nutrients. This talk will discuss the computational challenges of simulating the evolution of these waves and their impact on their environment.

And a contributed talk from

Aaron van Donkelaar, Dalhousie University

Global Estimates of Fine Particulate Matter from Satellite



Bioinformatics

Bioinformatics applies computational techniques to help us understand ourselves and the living world around us. The field is now seeing explosive growth thanks in large part to new technologies that can generate enormous amounts of biological data, in particular DNA sequences. These new data streams offer us innovative ways to understand the role of our genes in health and disease, the forces driving biodiversity in an age of habitat loss and climate change, and the heretofore under-appreciated role of microbes in many systems. To realize the full potential of bioinformatics, we need to develop new tools and techniques that can rapidly process these large sequence data sets in light of the processes that shape organisms and their genomes. This session will focus on the intersection between computational biology, algorithms, and high performance computing.



Plenary Speakers Frank Dehne and Ashkan Golshani

Departments of Computer Science and Biology, Carleton University

Frank Dehne

Dr. Dehne is the Chancellor’s Professor of Computer Science at Carleton University in Ottawa, and Fellow of the IBM Centre For Advanced Studies Canada. His research interests include the design and implementation of efficient parallel algorithms, the interrelationship between the theoretical analysis of parallel algorithms and the performance observed on current parallel architectures, and the use of efficient parallel algorithms for large-scale data analytics and computational biology/bioinformatics.

Ashkan Golshani

Dr. Ashkan Golshani is an Associate Professor of Biology and the Director of the Systems Biology Group at Carleton University. His main area of research includes high throughput investigation of protein-protein interactions and genetic interactions in model organisms. He is currently a natural sciences member of the CIHR Advisory Board for the new open funding schemes.



Protein Interaction Prediction Engine (PIPE) Project

This talk will discuss PIPE, a multi-disciplinary project at Carleton which began in 2002, with the aim of predicting protein interactions. Using HPC, the latest version of PIPE (MP-PIPE) enabled the research group to perform the first ever complete scan of the entire Human protein interaction network in a massively parallel computational experiment. The implications for understanding human cell function will be significant.

Robert Beiko

Faculty of Computer Science, Dalhousie University

New Generation of Bioinformatics Techniques. Ten years on, the human genome remains an important source of discoveries. But we are only beginning to appreciate the diversity of the microorganisms that live inside us, and their roles in health and disease. Taking stock of this “microbiome” requires DNA sequencing on an unprecedented scale, and demands a new generation of bioinformatics techniques to make sense of the data.

And contributed talks from

Cristina Baciu, The Hospital for Sick Children, Toronto

Bioinformatics Analysis to Uncover Predisposition to Pediatric Medulloblastoma

Carol Gauthier, Compute Canada, Université de Sherbrooke

Genetics and Genomics Analysis Platform (GenAP)



Computational Materials Science

Numerical modeling and simulation in Materials Science plays a vital role in our understanding of the complex and subtle relationship between the constituent atoms and molecules of matter and its bulk properties. Spanning theory and experiment, numerical modelling and simulation is fundamental to the ongoing development and application of new materials and processes that are key to existing and potential future technologies. The two sessions on Computational Materials Science will focus on the application of large scale computing techniques used to model the diverse range of phenomena that characterise many materials of current scientific and technological significance.


Plenary Speaker Nikolas Provatas

Department of Physics, McGill University

Nik Provatas

Nikolas Provatas is a Professor of Physics at McGill University and holds a Canada Research Chair in Computational Materials Science. His research is at the interface of condensed matter physics and materials engineering. It uses high-performance computing, principles of non-equilibrium thermodynamics and experiments to understand the fundamental origins of length scale selection in microstructure evolution in engineering materials and materials processes. These include systems undergoing crystallization from a melt or amorphous phases, particle precipitation, second phase formation, grain growth kinetics and reaction-diffusion processes in heterogeneous materials.


Computational Modeling as a Predictive Tool for the Development of Advanced Engineering Materials

Advances in high performance computing and Big Data science are now making it more realistic to think about designing materials from the atomic scale up by manipulating the non-equilibrium processes that control the evolution of their microstructure. This talk will review the success that approaches such as phase-field, phase-field-crystal, DFT and molecular dynamics modeling have had, and illustrate advances that have been enabled by high performance computing and their potential for industrially relevant applications in advanced materials.

Richard Bowles

University of Saskatchewan

The Thermodynamics and Dynamics of Confined Hard Sphere Systems. Colloidal particles confined within narrow, quasi-one-dimensional channels are found in a wide array of biological and engineering systems that include the passage of proteins through cellular microtubules and the filling of carbon nanotubes with fullerenes. We use molecular dynamics simulations of hard spheres confined within a hard cylinder to study the thermodynamics and dynamics of these types of systems. When the density of the system is low, the fluid is isotropic, but at high densities, the particles pack into complex helical structures that vary with the diameter of the channel. For wide channel diameters, the transition between the fluid and helical crystal occurs through a first order phase transition, but as the channel narrows, the transformation occurs continuously as a function of density. We also find that topological defects in the helical structure play an important role in the particle dynamics of these fluids.

Hong Guo

Department of Physics, McGill University

Device simulation from atomistic first principles. Significant progress has been achieved in fabrication of nanostructures where quantum transport of charge/spin is closely coupled to atomic details of the device material. A serious challenge to device physics is to develop high performance computational tools for quantitative and material specific prediction of nonequilibrium quantum transport. Dr. Guo will explain nanoelectronic device simulation from the point of view of atomic first principles. Then, using the newly discovered monolayer two-dimensional transition metal dichalcogenides (TMDC) as an example, he will report a set of new device phenomena that are related to electric control of spin, quantum transport mediated by the spin-valley degrees of freedom, and the notion of valleytronics. He shall conclude by summarizing his efforts, the status and challenges in developing the atomistic technology computer aided design (A-TCAD) tools for emerging nanoelectronics.

Martin Plumer

Department of Physics and Physical Oceanography, Memorial University

Modeling New Paradigms in Magnetic Recording. Modeling the magnetic processes involved at both the nanometer and atomic scale can provide valuable guidance towards the evaluation and optimization of new paradigms under consideration in the magnetic recording industry. A brief overview of recent progress and challenges in the simulation of thermally assisted magnetization reversal in recording media will be reviewed.

Ivan Saika-Voivod

Department of Physics and Physical Oceanography, Memorial University

Some Liquids Freeze When You Heat Them. Normally a liquid freezes when you cool it. However, there are a handful of liquids that do the opposite and crystallize when you heat them. Dr. Saika-Voivod’s group studies a model, formulated originally to understand the anomalous properties of liquid water, and find that it exhibits just this behavior at particular pressures.

Robert Wickham

Department of Physics, University of Guelph

Characterizing the structure and dynamics of the disordered micelle regime in diblock copolymer melts through simulation. Dr. Wickham’s group uses large-scale, fluctuating, dynamical field-theoretic simulations to characterize the fluid-like structure and dynamics of disordered micelles, which are self-assembled, nanometer-scale macromolecular aggregates, and which are challenging to study analytically. Ultimately, the group aims to study nucleation of a micelle nano-crystal, and to explore the possibility of forming a “micelle glass”.

Josef W. Zwanziger

Department of Chemistry, Dalhousie University

Computational Design of Broad-band Zero Stress-Optic Glass. While most optical glass becomes birefringent when mechanical stress is applied, so-called Zero Stress-Optic (ZSO) glass self-cancels this property. Using high performance computing Dr. Zwanziger’s group has designed families of ZSO glass, which they have then successfully synthesized and tested in the lab. Dr. Zwanziger will explain how computational studies are advancing this area, including the very new subject of broad-band ZSO glass.

And contributed talks from

Laura Albrecht, Dalhousie University

Investigating Local Stability in Water Wires Using Atomic Energies

Sergey Kazachenko, Queen’s University

Algorithms for GPU based molecular dynamics simulations of coarse grained fluids: Applications to
liquid crystals

Seetha Pothapragada, Stony Brook University

Particle-Based Simulation of Platelet Activation in Blood Plasma on Supercomputers



Big Data and Analytics

Big Data is transforming science, engineering, medicine, commerce and perhaps ultimately society itself. We are accumulating exponentially increasing amounts of data in diverse forms including text, images, video, graphs, clickstreams, and structured and semi-structured data. Extracting value from these massive, rapidly evolving datasets requires a judicious combination of high performance computing, algorithmic, and domain specific analytical techniques. This session will focus on challenges in HPC and Big Data Analytics including the design of novel algorithmic techniques, specialized big data processing environments, and Big Data applications.


Plenary Speaker Alexander Kreinin

Head of Quantitative Research, STSM, Risk Analytics at IBM Canada

Alexander Kreinin

Alex has been with Algorithmics since 1995. After IBM acquisition in 2011, he became STSM and Head of Quantitative Research, Risk Analytics. Alex has PhD in Probability and Statistics from the University of Vilnius (Lithuania). He published over 60 papers and 2 monographs. His research areas include Market and Credit Risk Modelling, Numerical Methods for Risk Management, Monte Carlo Methods, Calibration of Stochastic Models, Semi-Analytical Methods of Portfolio Valuation, Design of Numerical Algorithms and their Software and Hardware Implementation. Dr. Kreinin is an Adjunct professor in the Computer Science Department of the University of Toronto and has been affiliated with the “Masters of Mathematical Finance” program.
HPC in Financial Risk Management
Advanced computational strategies are at the heart of Financial Risk Management. The computation of the risk measures of a large financial portfolio usually requires huge computational resources that must be deployed in a timely manner. In this talk we consider the main computationally intensive risk management problems and discuss their potential HPC solutions.

Oliver Baltzer

Co-Founder and CTO, Analyze Re

Big Data Analytics for Risk Engineering. Risk Engineering is the discipline of utilizing predictive analytics and real-time market data to design portfolios of risk that target specific economic goals while satisfying regulatory, market, and other constraints. Risk Engineering is employed by insurance and reinsurance companies as part of their strategic planning and execution. Analytical processes around Risk Engineering are data-driven and computationally intensive. Analyze Re provides Software-as-a-Service solutions for Risk Engineering to the insurance and reinsurance industry. This talk discusses our experience in building such platforms on the basis of commonly available Cloud technologies and what aspects from HPC we have been able to leverage.

Khuzaima Daudjee

University of Waterloo

An Evaluation of Pregel-like Graph Processing Systems. The introduction of Google’s Pregel for large-scale graph data processing has led to the recent development of Pregel-like systems such as Apache Giraph and GraphLab. Dr. Daudjee will present a performance study of these systems that compares them experimentally by considering graph and algorithm agnostic optimizations on up to 128 Amazon EC2 machines, identifying areas of improvement for all systems.

Dominic Lam

Visiting Scientist, Princess Margaret Cancer Centre

Social Sciences & Humanities Research Computing

. The Big Data digital revolution offers parallels to the Industrial Revolution of the 19th century. Both extend our efficiency; one of our body and one of our brain. Today’s Big Data Analytics is breaking that parallels that it starts originate interaction with the human mind. Arguably, social and human activities form the most complex emergent phenomena and research in that domain will benefit tremendously from leveraging Big Data Analytics. 

This presentation will describe a set of Big Data Analytics projects in Social Sciences & Humanities research in Canada and will lead to a discussion on how the HPC support community can help enhance and expand these research activities.

Stan Matwin

Dalhousie University

Big Data Access – the Burning Issues. We are all familiar with exciting examples of fruitful Big Data applications using mobile or web usage data. And yet most of them are done using highly proprietary data and raise the issue of data access. In this presentation I will discuss some of the challenges of Big Data research, and their possible solutions. I will look first at some highly speculative ideas in big data governance. I will then focus on some of the existing cutting edge work in the Privacy-enhancing Techniques (PET), as tools that could potentially help to offer Big Data access solutions.



Oil and Gas Reservoir Modelling

Plenary Speaker K. Sampath

R&D Manager, Hibernia Management & Development

K. Sampath

Dr. Sampath’s interests include Arctic Engineering, Safety & Environment and reservoir performance. Prior to his current assignment, he was Reservoir Manager at ExxonMobil’s Upstream Research Company in Houston, Texas. Dr. Sampath has experience in reservoir engineering, reservoir simulation and reservoir management.




Enabling Improved Oil and Gas Exploration and Exploitation through Advances in HPC
The growth in demand for energy continues to drive the exploration of oil and gas resources to more challenging environments with an attendant increase in complexity of technology required for resource development. Advances in high performance computing is poised to facilitate efficient alternatives to existing scientific methods for subsurface imaging and modeling, while offering opportunities for new techniques to better account for underlying uncertainties towards improved project economics.

And contributed talks from

David Dubé, École Polytechnique de Montréal

An efficient parallelization of the radioactive particle tracking technique on graphic processing unit

Hui Liu, University of Calgary

Development of General Purpose Parallel Platform on IBM Bluegene/Q

Anup Radhakrishnan and Erkan Cakir, Memorial University

High Performance Cluster Computing for Fluid-Structural Analysis of an Experimental Rig to Test Submerged Flexible Ocean Structures



Data Caves and their Applications

With the ability to perform large scale and complex 3D numerical simulations came the difficulty of visualizing the results of these calculations. Data caves are 3D immersive environments that allow the user to determine both general regions where interesting events are taking place and specific results. As the technology matures, many other applications from art to psychology to medicine have been found for data caves. This session seeks to share the spectrum of these applications to help current users learn new methods and potential users think about how data caves might benefit them.


Christoph W. Sensen

Visual Genomics Centre, Department of Biochemistry and Molecular Biology, Faculty of Medicine, University of Calgary

Using a Virtual Reality System to Study Diseases and Developmental Patterns. The modeling of complex genetic disease progress and developmental patterns requires the integration of a large number of –omics data types. Most of the data produced, from complete or partial genomes over Proteomics to Metabolomics data, thus are only analyzed in isolation by specialists in the respective field. Their integration requires the development of Ontologies and visual models, which allow the creation of unified models. Using the CAVEman system as an example, we will demonstrate the use of a virtual reality system to study diseases and developmental patterns.

And contributed talks from

John Charles, Halifax Regional Municipality

Utilizing Data Cave Technology for Municipal Infrastructure Planning

Glen Hougan, NSCAD University

Data Cave to Word Cloud: visualizing new experiences

Mohamed Mediouni, Université de Sherbrooke

Orthopedic simulation in the CAVE: femoral diaphyseal fracture

Luke Murphy, Saint Mary’s University

Visualization of the Active Site of the Ethylene-Forming Enzyme using a Data Cave



Infrastructure and Technology

Behind all the research successes we’re showcasing at HPCS, is the computational infrastructure, resources, software, computational methods, and technology that make it possible. In this stream, we will be hearing from Computer Scientists, our Advanced Computing and Advanced Networking infrastructure providers in Canada and the US, the software platforms that enable innovation, and the industry partners who are supporting this community (and this conference!)


Technology Keynote Speaker Paul Crumley

IBM Research, T.J. Watson Research Center

paul crumley

Paul Crumley, IBM Research, T.J. Watson Research Center, is a Senior Technical Staff member. This title is given to IBM scientists and engineers who lead innovation and serve as a bridge between the R&D and customer worlds. After creating ways to scale IBM’s Blue Gene technology, Paul is now involved in bring the lessons learned in HPC to cloud. Prior to joining IBM, Paul was the Assistant Director at Carnegie Mellon University, Information Technology Center, where he lead the vision for the department’s R&D work. With his background in business and research communities, Paul is in a unique position to share lessons from distributed computing and HPC to leading edge big-data and analytics projects such as IBM Watson.

Techniques for HPC and Large Cloud Environments. Social interactions, Big Data, Analytics and the Internet of Things are pressing at the boundaries of our current computing infrastructure and technology. Mr. Crumley will review a number of capability systems which he has helped create, decommission or operate over more than two decades. The goal is to understand some unique features of these systems to help us consider what lessons and techniques we should bring forward to new HPC systems or large cloud environments as these platforms are the vehicles by which we will computationally explore new realms such as pharmacology, weather and energy management.


Technology Keynote Speaker Dean Koester

VP Sales, Data Direct Networks

dean koester

Unlocking the Commercial HPC Code. From a scientific and industrial focus, HPC is now moving to commercial organizations, presenting new challenges, and requiring a different approach for data management. In this presentation, DDN will discuss how cache-centric storage technologies can help resolve IO bottlenecks, while maintaining scale, performance, management and reducing TCO.

Steve Tuecke

Deputy Director, Computation Institute, University of Chicago

Campus Data Service using Globus. Globus provides software-as-a-service (SaaS) for research data management, including data movement, storage, sharing and publication. This talk will describe how to build a secure, scalable campus data service that delivers advanced data management capabilities to researchers while significantly reducing the development, integration, and operations burden on resource providers. Using Globus as an exemplar, we will illustrate why the user experience is critical to making these capabilities accessible to a broad range of users at national research and education institutions, leveraging intuitive web interfaces and hiding access complexities behind a sophisticated federated identity infrastructure.

Gregor von Laszewski

Assistant Director of CGL and DSC, Indiana University Pervasive Technology Institute

Lessons Learned from Cloud usage in FutureGrid. FutureGrid is a project to develop a high performance grid testbed that allows scientists to collaboratively develop and test innovative approaches to parallel, grid, and cloud computing. In this talk we will provide an overview of FutureGrid, identify some of the lessons that we learned, and summarize some of its main features. These include that (a) users can deploy their own software defined systems on resources within FutureGrid not only virtual machines, but also bare-metal with our Rain provisioning framework, (b) an advanced framework to manage user and project affiliation and propagate this information to a variety of subsystems constituting the FutureGrid service infrastructure, (c) access to a variety of IaaS services, (d) a metric system that allows us to create usage reports from our entire IaaS frameworks, (e) repeatable experiments can be created with a number of tools including Pegasus, Precip and Cloudmesh, and (f) Cloudmesh, a system to manage virtual machines in a federated view of multiple cloud environments.

Peter Bojanic

VP Software Development, Xyratex, a Seagate Company

From Spinning Disks, to HPC to Cloud – A Seagate Perspective. As HPC work-load rapidly extends beyond on-site computing to Cloud connected global scientific communities, this presentation provides key insights into latest storage developments, from technologies critical to the next 10x capacity increase, to high availability scale-out HPC storage, to open source Cloud initiatives driving the need for cost-effective compatibility strategies.

Bernard Cobb

Portfolio Lead, Modular Data Centers, HP Critical Facilities

Data Center Strategy and a New Alternative to Minimize your Capital Investment. A successful data center strategy aligns with your C-Suite Executive priorities and aligns your data center facility with the ever changing needs of IT The more capital tied up in your data center, the less capital available to invest in your business.

Scott Henwood

Lead Software Architect, CANARIE Inc

This Changes Everything: Accelerating Scientific Discovery Through High Performance Digital Infrastructure The needs of the scientific research community sit at the intersection of advanced technologies, big data, and global networks of collaboration. CANARIE’s Research Software program leverages national high performance digital infrastructure to facilitate the creation of software tools and services that equip researchers to maximize the potential of data to accelerate scientific research. The creation of a toolkit of reusable software services also reduces duplication in software development, and enables more research dollars to be focussed on research.

In this session, Scott Henwood will explain how CANARIE is changing the paradigm for research software development. The presentation will illustrate how this innovative model, based on the power of collaboration to maximize investments in resources, is resulting in the efficient development of new platforms that accelerate time to discovery. Recent ground-breaking and cross-discipline software services developed through this program will be showcased, highlighting the benefits of this collaborative initiative for the scientific community.

Todd Price

Associate Director of Product Management, Data Centre and Cloud Solutions, Bell Aliant

From Colocation to the Cloud – Why Hosting is more Strategic than you think. A strategic overview on the importance of Colocation in your Data Centre and Cloud evolution. Colocation has become a critical stepping point as customers evolve from an in house data room/data centre solution to an integrated, multifaceted hybrid cloud. Understanding the technology and key business decisions can help Universities and Enterprises move towards an optimal colocation and cloud solution. This presentation will touch on the business implications of key trends we are seeing in the Data Centre market place in 2014 and the recommended approaches for optimizing your data centre road map.

Thomas Armstrong

Solutions Architect, Fusion-io

Flash Memory in Big Data. Mr. Armstrong will discuss how flash can be leveraged to unleash maximum performance in Big Data processing workloads.

Vito Bongiorno

Manager, North America Pre-Sales, Cray, Inc.

Cray Solutions for Higher Education and Research. Cray has always been known for its high-end, high-performance, compute systems. It has a strong presence at the top national academic computation centers in the world. This talk will highlight some of the pioneering science done on these systems. The focus of the talk, however, will be on some of the less known products Cray brings to the academic research. This includes a standard cluster offering (the result of the Appro acquisition over a year ago) that combines high quality standard components with the software and services Cray is known for; and a full range of storage products, including a Cray-design Lustre appliance and a Hadoop product. The talk will cover Cray’s end-to-end solutions that are optimized for both compute and data intensive workloads.

Mike Bloom

Corporate Advisory Engineer (Canada), Mid-Tier Storage Division, EMC

Building Predictability in HPC Solutions. High Performance Computing struggles with the requirement of being able to provide predictable performance while being able to be deployed by individuals whose primary competency is usually analytics or another formal discipline. This session focuses on the problem of scaling, agility, and interoperability between applications that reside on different protocols in a high performance and data analytics context.

Marcel Bourque

Oracle Principal Sales Consultant, Oracle

Oracle innovations that cross the barriers between HPC and Enterprise computing. This session will look at Oracle innovations that cross the barriers between HPC and Enterprise computing, such as Infiniband in the enterprise, task specific systems, and high-performance, high-capacity, scalable storage.

Terry Dalton

Chair, ACORN-NS

Strengthening Nova Scotia’s Digital Infrastructure. ACORN-NS is the Nova Scotia Regional Advanced Network responsible for advanced network infrastructure development, operation and services throughout Nova Scotia for the research, education, health, and innovation communities, promoting the growth of a knowledge-based economy and society in Nova Scotia. Terry Dalton, Executive Director and Chair ACORN-NS, will give an overview of active projects and initiatives strengthening the advancement of digital infrastructure throughout Nova Scotia.

Roger Delisle

Manager, Research Partnerships Programs, NSERC

NSERC Partnership Grants – Supporting Research Collaborations between Industry and Universities. The Natural Sciences and Engineering Research Council (NSERC) empowers successful R&D partnerships between industry and post-secondary institutions. A brief overview of granting opportunities funding the whole spectrum of R&D activities, from short-term practical design solutions to 5-year strategic endeavours, will be presented along with examples of successful outcomes.

Don Etienne

Arista Networks

Next Generation Data Centre Network: SDN. Arista Software Driven Cloud Networking (SDCN), combines the principles that have made cloud computing the unstoppable force that it is: automation, self service provisioning, and linear scaling of both performance and economics, coupled with the recent trend in Software Driven Cloud Networking that delivers: network virtualization, custom programmability, simplified architectures, and more realistic price points to the table. This combination creates a best-in-class software foundation for maximizing the value of the network to both the enterprise and service provider data center; a new architecture for the most mission-critical location within the IT infrastructure that simplifies management and provisioning, speeds up service delivery, lowers costs and creates opportunities for competitive differentiation, while putting control and visibility back in the hands of the network and systems administrators.

Nick Ihli

Adaptive Computing

Speeding the Time to Discover with More Efficient Workflows – What’s New in Moab HPC Suite. Moab HPC Suite is trusted by the world’s largest HPC and Cloud environments. Come learn about the latest innovations in Moab HPC Suite that unifies data center resources, optimizes the analysis process and guarantees services to the business. These new features include Green Policy Configuration and Operation, Job Energy Usage Accounting, Data Staging Refactoring, Grid Operation Validation, High Throughput Computing and other market demanded capabilities to help your organization extract the insights necessary to make game-changing, data-driven decision to ensure a competitive advantage.

William Moore

Senior Account Executive, Net Direct Inc.

Commercial off-the-shelf (COTS) Servers + Original Design Manufactures (ODM) + Open Source Software = Cost Effective HPC. This session will focus on Net Direct’s experience with delivering and supporting, powerful, agile, industry standard COTS based HPC solutions.

Florent Parent, Lixin Liu, and John Sherwood

Calcul Québec, Westgrid and ACORN-NS

perfSonar as a National Network Monitoring Platform. Achieving high performance data transfers over high speed wide area networks is challenging. Proper software for data transfer must be available to researchers, network stacks need to be properly tuned, and networks must be monitored to find and solve potential bottlenecks. This talk will focus on network monitoring using perfSonar. We will present the coordinated efforts from Compute Canada, Canarie and ORANs to deploy perfSonar measurement infrastructure across Canada. We will also describe some real network problems that have been detected by perfSonar.

Lori McMullen

Executive Director, Canadian University Council of Chief Information Officers (CUCCIO)

Digital Leadership Infrastructure Council (DILC): Panel Discussion. This session will provide an update from the DILC and the key outcomes of the National Digital Infrastructure Summit held in January 2014. A brief presentation summarizing various topics including: the challenges of the present digital infrastructure (DI) ecosystem, the implications of the proposed changes to tri-council policies for researchers and institutions and the preliminary results of recent pilot projects.

Compute Canada Town Hall

An informal meeting for Compute Canada’s researchers and extended user community to discuss emerging issues on advanced research computing including the CFI cyberinfrastructure initiative and Compute Canada’s Sustainable Planning for Advanced Research Computing (SPARC). President and CEO, Mark Dietrich and Chief Scientific Officer, Dugan O’Neil will seek feedback on Compute Canada’s services and how they can be improved to enable your research. As the national platform for advanced research computing in Canada, Compute Canada senior leaders need to ensure its investments and services include both specific and common requirements for its researchers. This is your opportunity to have an impact on future planning and voice your thoughts on all things Compute Canada.

And contributed talks from

Bill Appelbe, ORION

Advanced Computing Needs and Opportunities – the ACTION Project

Mark Leggott, University of Prince Edward Island

Building a Research Data Management System with Islandora

Cristian Suteanu, Saint Mary’s University

Symmetry and scaling properties in natural time series – new tools for the detection and characterization of pattern change



Posters

First prize winner

  • Carl P. Romao (co-authors Josef W. Zwanziger, Mary Anne White), Finite Element Analysis of Thermal Stress in Thermomiotic Materials

Second prize winner

  • Maan H. Hani (co-author Rob Thacker), Explaining the Low Luminosity Tail of the AGN Luminosity Distribution

Third prize winner

  • Rehan Siddiqui (co-authors Brendan Quine,Caroline Roberts,Naif Al Salem,Catherine Tsouvaltsidis,Rajinder Jagpal), Efficient Detection of Cloud Scens from Spectral Data of Space-Orbiting Argus1000 Micro-Spectrometer

Other presented posters

  • Alejandro Allievi, Reducing Java I/O Bottlenecks Using MPI (FastMPJ) in Computational Fluid Dynamics Applications
  • Dijana Anzelj (co-author Cory C. Pye), Ab nitio Investigation of Pb (II) Complexes as Possible Corrosion Products in a SCWR
  • Victoria Berryman (co-author Russell J. Boyd), A Computational Investigation of Nitric Oxide Induced Oxidation of Oxyhemoglobin
  • Stephen Driscoll (co-authors Jean Burnell, Russell Boyd), A Computational Study of Cyclohexyne: the Search for Regioselective Cyclic-Alkyne Cyclization Reactions
  • Nathaniel Egan-Pimblett, Using the GEOS-Chem Adjoint Model to Understand Black Carbon Aerosol Emissions
  • Lanli Guo (co-author Jinyu Sheng), Numerical Simulation of Ocean Waves over the North Atlantic Ocean using WAVEWATCH III with Climate Forecast System Reanalysis
  • Xiang Jiang (co-author Daniel L. Silver), Using Vector Space Model of Semantics to Reduce Dimensionality of Text Data
  • Vasily Korabel (co-author Keith Thompson), The Assimilation fo Sea Level and Sea Surface Temperature into a ¼ degree North Atlnatic Model – Using Multivariate Ensenble Optimal Interpolation: The Importance of Tailoring the Ensemble
  • Ian Lumb (co-author Jim Freemantle), Towards Earthquake-Tsunami Causality via Data Science: Giraph-Derived Credibility Scores for Data from Twitter
  • Sedigheh Mahdavi, (Co-Authors L.A. James, T.E. Johansen), A General Review: Role of High Speed Computation in Oil and Gas Reservoir Simulation
  • Charles Marcotte Girard (co-author Alain Rochefort), Electronic Properties of CIGS Nanowires: a Theoretical Perspective
  • Reza Nakhjavani (co-authors Sahel Sharify, Ali B. Hashemi, Alan W. Lu, Cristiana Amza, Stephen Strother), Porting a Neuro-Imaging Application to a CPU-GPU cluster
  • Kyoko Ohashi (co-author Jinyu Sheng), Numerical study of the effect of the physical environment on the movement of the American Eel in the Gulf of St. Lawrence
  • Shiliang Shan (co-author Jinyu Sheng), Examination of Circulation and Particle Movement in a Submarine Canyon: Sable Gully, Using a High-resolution Ocean Circulation Model
  • Tarek Sherif (co-authors Nicolas Kassis, Marc-Étienne Rousseau, Reza Adalat, Alan Evans), BrainBrowser: Distributed, Web-based Neurological Data Visualization
  • Ian Short (co-author Eamonn Campbell), Modelling the Spectra of Red Giant Stars
  • Oliver Stueker (co-authors Sharene Bungay, Jason Pearson, Raymond Poirier, Peter Warburton), Retrievium – A Semantic Repository for Quantum-Mechanical structures
  • Jorge R. Urrego-Blanco (co-author Jinyu Sheng), Study of Sea Ice Dynamics in the Gulf of St. Lawrence Using a Coupled Ocean-ice Model
  • Monica Wachowicz (co-authors Maria Dolores Arteaga Revert, Rui Sun, and Yves Bourgeois), From 140 Characters to 140 Terabytes of Creative Ideas
  • Kai E. O. Ylijoki (co-authors Zachary E. Konway, Nicole Laidlaw, Shifat Sharmin, Stephanie Boudreau, Peter H. M. Budzelaar, Jeffrey M. Stryker, E. Peter Kündig), Computational Modeling as a Predictive Tool in Synthetic Chemistry
  • Mitchell Young (co-author Dr. C. Ian Short), 1.5-D Red Giant Stars in NLTE
  • Heng Zhang (co-author Jinyu Sheng), Numerical Study of Extreme Sea Levels Due to Storm Surges and Tides over the Northwest Pacific