Seminar: E-Science at Petascale - solving Multi-Scale Multi-Physics Problems arising in Engineering Design
Professor Martin Berzins from the University of Utah's Scientific Computing and Imaging Institute is giving a talk at BlueFern on Tuesday 10 December. BlueFern is part of NeSI at the University of Canterbury.
Details
Date: Tuesday, 10 December 2013
Time: 10:00 to 11:30 a.m. (morning tea 10am, talk 10.30am)
Location: Undercroft seminar room, University of Canterbury, Christchurch
Contact: Angela Armstrong angela.armstrong@canterbury.ac.nz
RSVP: Please RSVP for this event by contacting angela.armstrong@canterbury.ac.nz
Abstract
The computational capability to address a number of challenging multiscale multiphysics engineering design problems involving the use of petascale and (hopefully) exascale computers is considered in the context of the Uintah software framework developed at the University of Utah.
The problems considered are
- How do we transport explosives to avoid developing detonations?
- how do we design the next generation of large scale clean coal boilers? and
- how do provide a novel framework for materials by design?
The Uintah framework used in these studies allows the solution of large-scale, fluid- structure interaction problems through the use of fluid flow solvers coupled with particle-finite element based solid methods. Uintah's combustion solver tackles a broad and challenging class of turbulent combustion problems. A unique feature of Uintah is that it uses an asynchronous task-based approach with automatic load balancing to solve complex problems using techniques such as adaptive mesh refinement. At present, Uintah is able to make full use of present-day massively parallel machines as the result of three phases of development over the past dozen years. These development phases have led to an adaptive scalable run-time system that is capable of independently scheduling tasks to multiple CPU cores and GPUs on a node.
Particular challenges are include
- addressing multiple scales by using mesh refinement and coupling to molecular dynamics,
- solving large systems of equations at scale using iterative solvers and
- designing algorithms to address the challenges of radiation modelling
The approaches adopted to achieve present scalability are described, and their extensions to possible future architectures and challenging future problems such as multiscale materials by design are considered.
About the speaker
Martin is a Professor of Computer Science in both the School of Computing and in the Scientific Computing and Imaging Institute at the University of Utah. Before this he was at the University of Leeds in the UK where he was Professor of Scientific Computing and Research Dean for Engineering. He earned his PhD in the area of Mathematical Software and Numerical Analysis at Leeds in 1981. He has worked in the fields of mathematical software, numerical analysis and parallel computing with application to challenging problems in science and engineering. His present focus is on the application of petascale computers to the solution of combustion problems and multiscale materials by design problems.
