Dr. Jason Bond, Auburn University

Dr. Bond spent two years as a postdoctoral research associate at the Field Museum of Natural History in Chicago where he studied Caribbean millipede systematics and evolution.   In 2001 he joined the Biology faculty at East Carolina University as a Professor of Biology and Director of the North Carolina Center for Biodiversity.   In 2011 he relocated to Auburn University to serve as Director of the Auburn Museum of Natural History and Biodiversity Learning Center. 

His research focuses on questions related to the pattern and process of evolutionary diversification in arachnids and myriapods.  His lab group has worked to develop approaches to evaluating species boundaries that seek to integrate genetic, spatial, ecological, and phenotypic data.

  Dr. Bond's talk will focus on the application of high performance computing to solving questions regarding organismal relationships and what these patterns tell us about how life has evolved on the planet.

 

Søren Holm Mogensen, Vestas Technology R&D

Mr. Mogensen has specialised within the field of numerical flow simulations. He has several years experience working with computer flow models. Furthermore, he has working experience in turbine load calculations and verifications according to IEC and Germanischer Lloyd standards.

Mr. Mogensen has been working in different companies with many aspects regarding flow regimes analysis and turbines verifications.Now is Specialist in Wind & Site Competence Centre, Vestas Technology R&D.

He has worked in different research projects, such as Energiberegninger: A project that made a database of existing wind farms, covering all different kinds of terrain, or  ENDOW, an EU funded project to examine turbine wakes regarding big offshore wind farms. Several people from research institutes throughout Europe joined the project.

 

 

 Mr. Mikael Haglund, IBM Sweden.

IBM's Chief Technologyst Mikael Haglund will introduce Watson's GTO.The Global Technology Outlook (GTO) is IBM Research's tool to communicate early identified significant technology trends, attempting to see future high impact technologies and game changing products & services. It looks specifically for disruptive changes to business / threshold crossings, and technologies with high potential to create new businesses.

The GTO helps drive IBM's $6B investment in R&D each year, and this year's GTO features both the game winning Watson and some potentially game changing hardware developments.

Dr. Daniel Hilding, Executive Director, DYNAmore Nordic

Daniel Hilding has been working with simulation and the finite element method since 1996, within both research and in the industry.
Daniel Hilding has over the years been working with software development for LS-DYNA, development of material models for metals, crash analysis, metalforming, fatigue (LCF), problems involving impact and penetration, and analysis and dimensioning of structures subject to blast loads.

Daniel Hilding will talk about HPC and LS-DYNA. LS-DYNA is a leading multiphysics simulation code for simulation based design in the engineering industry.

To meet the demands on realistic simulation, increasingly large HPC systems are employed by the LS-DYNA users. Topics of the talk include: current HPC systems and applications, current trends reagarding HPC and LS-DYNA and its impacton product development, and a number of show cases of state of the art analysis from various industries.

 

Tommy von Brömsen, Sveriges Riksbank

Mr. von Brömsen works in the research department of SverigesRiksbank, the central bank of Sweden. His role is to provide mathematical knowledge, primarily towards programming of economic models and the handling of large data sets. The models require high-performance computing to investigate the state of the economy.

Mr. von Brömsen hasan M.Sc. in Financial Mathematics from the University of Gothenburg / Chalmers, and a B.B.A. in Business Economics from the University of San Diego. Tommy played professional football in Allsvenskan, the highest division in Sweden.

Yoran Pichon, SCE solutions

Mr.Yoran Pichon is working at SCE Solutions on developing a new racecar:  The Le Mans Prototype Project. This as an end-project of his study, in combination with his work for CD-adapco.
Mr. Pichon is  a Engineering Student at the Ecole Centrale de Lille, specialized in Mechanics and Fluid Mechanics and wanting to begin a career within the motorsport industry.
Also a MSc student at Cranfield University, UK, within Motorsport Engineering and Management MSc course.

 

Mr. Mustafa Kavasoglu, ABB AB, corporate research.

Mr. Kavasoglu has a  M.Sc in Scientific Computing both from KTH and FAU Erlangen. He is working in the ABB Corporate research, with the role of building multi physic simulation models for ABB products which involves different fields and the interactions between them. The increasing complexity of the problems require high performance computing to meet both memory and computing power requirements.

Mr. Kavasoglu's speech has the title "Prediction of Transformer Load Noise" and will explain how transformers, as any other industrial products, have to fulfill various requirements on noise levels. Therefore, it is essential to use appropriate prediction tools in order to select a suitable strategy for noise control. The problem is a multi-physic problem involves electromagnetism, mechanics and acoustics and all necessary couplings between the fields.

 

Mr. Einar Rustad, Numascale.

Mr. Rustad has a background as CPU architect in Norsk Data, Dolphin, ACRI (France),  founder and VP of Scali and founder and VP of Numascale.

Numascale has developed NumaConnect to enable a collection of commodity servers to be used as an SMP. As the name NumaConnect indicates, this is an interconnect with hardware support for system-wide cache coherence which is a crucial feature for processor performance in a shared memory system.
As volumes of data grow at an exponential rate, there is a growing demand for large memory space to enable companies and institutions to extract vital knowledge from large amounts of data. For many purposes, this data analysis task must be performed in real-time or near  real time to be valuable. This requires data to be held in main memory rather than in orders of magnitude slower mass-storage.
Traditionally, large shared memory systems have been offered only by the large tier-1 system vendors at price levels that in many cases are prohibitive for mainstream use. Numascale's approach offers capabilities like mainframe SMPs at the price level of clusters.