Tonight: November NWCPP Meeting - pizza update!

[Lloyd Moore]

UPDATE: Folks I would like to thank Kongsberg Underwater Technology for stepping up and providing pizza for the meeting tonight!!

 

I would like to welcome our very own Robert Goddard as the presenter for our November NWCPP meeting.

 

Time and Location:

November 16, 2011 at 7:00 PM, Microsoft Campus building 41/1511 (Townsend)

 

Title: Introduction to Scientific Computing

Abstract:
Computing in science and engineering involves modeling part of the physical
world. The inputs could be detailed measurements of the environment, or a
randomized set of potential measurements that might characterize some
environment in the future. The outputs could be images, or sound, or
predictions of whether a system will succeed or fail to perform its
function, or actions to make it more (or less) likely that some system will
succeed. The algorithms might involve solving a set of differential
equations, or integrating some function over a multi-dimensional domain, or
searching multi-dimensional parameter spaces to optimize some measure of
quality or fit, or sifting through noise to find a buried signal. Speed
requirements might be real-time (hard or soft), or much faster, or just fast
enough to fit into your project cost and schedule. The answers must be
accurate enough to accomplish the mission: Personnel training, or
performance prediction, or advance of scientific knowledge, or control of a
system or device, or life, or death.

Scientific computing typically involves a whole lot of arithmetic. So, I
will begin with a review of floating point arithmetic, specifically the IEEE
754 standard. (You might be surprised at how much you thought you knew, but
didn't.) Then I will move on to R. W. Hamming's five main ideas about
computation: Emphasis on the purpose of the computation, generalization of
algorithms into families, roundoff error, truncation error, and feedback
(stability). I hope to give enough practical information to help you avoid
the most common problems inherent in numerical computation, and to point out
resources that could help you go further.

Bio:
Robert Goddard is a physicist and software developer. For the last 31 years,
he has worked at the Applied Physics Laboratory of the University of
Washington, mainly on computer modeling of underwater sound. He is the
architect and team leader for the Sonar Simulation Toolset (SST), which
produces simulated underwater sound, suitable as input to sophisticated
signal processing systems (including human ears and brains), based on
user-specified descriptions of the undersea environment, the listening
system, and the sound sources and targets that might be out there. He has
also developed systems for control of measurement devices, data analysis,
visualization, modeling of quantum mechanical scattering, and optimization
of parameter values to fit observations.

Bob has been an active participant in NWCPP for most of its existence, and
is currently Treasurer.

 

Thanks,

Lloyd