|
Fall 2009
FLEET, INFINITY & MARINA
Adam Megacz (Berkeley) and Ivan Sutherland (Portland State)
October 9, 2009
5 – 6 pm
MEC 114
ABSTRACT
This talk describes a radically different architecture for computing called Fleet. Fleet
accepts the limitations to computing imposed by physics: moving data costs more
energy, more delay, and more chip area than the arithmetic and logical operations
ordinarily called "computing." Fleet puts the programmer firmly in charge of the most
costly resource: communication. Fleet treats arithmetic and logical operations as side
effects of where the programmer sends data.
Fleet achieves high performance through fine grain concurrency. Everything Fleet does
is concurrent at the lowest level; programmers who wish sequential behavior must
program it explicitly. Fleet presents a stark contrast to today's multi-core machines in
which programmers seek concurrency in an inherently sequential environment.
The Fleet architecture uses a uniform switch fabric to simplify chip design. A few
thousand identical copies of a configurable interface will connect a thousand or so
repetitions of basic arithmetic, logical, input-output, and storage units to the switch fabric.
The uniform switch fabric and the identical configurable interfaces will simplify many of
the hard parts of designing the computing elements themselves.
Both software and FPGA simulators of a Fleet system are available at UC Berkeley.
Berkeley students have written a variety of Fleet programs; their work helped to define
what the configurable interface between computing and communication must do. A
simple compiler configures both source and destination to provide flow-controlled
communication. We expect work on a higher-level language for Fleet to appear soon as
a Berkeley PhD dissertation.
Last year we built a 90 nanometer TSMC test chip, called Infinity, at Sun Microsystems.
Infinity demonstrated the switch fabric running at about 4 GHz. We now have a new test
chip, called Marina, also in 90-nanometer TSMC sponsored by Sun. Marina shows
correct operation of the configurable switch fabric interface. Together Infinity and Marina
give us confidence to build a complete Fleet. We seek participation from sponsors,
computer scientists, and hardware designers.
BIO
Ivan Sutherland is a Visiting Scientist at Portland State University where he and Marly
Roncken have recently established the "Asynchronous Research Center" (ARC). The
ARC occupies both physical and intellectual space half way between the Computer
Science (CS) and Electrical and Computer Engineering (ECE) departments at the
university. The ARC seeks to free designers from the tyranny of the clock by developing
better tools and teaching methods for design of self-timed systems. Prior to moving to
Portland, Ivan spent 25 years as a Fellow at Sun Microsystems. A 1959 graduate of
Carnegie Tech, Ivan got his PhD at MIT in 1963 and has taught at Harvard, The
University of Utah, and Caltech. Ivan is a member of the National Academy of
Engineering and the National Academy of Sciences.
Using RESTful Web Services and Cloud Computing to create Next Generation Mobile Applications
Jason Christensen
October 15, 2009
1:40 – 3:30 pm
MEC 114
Mobile computing has entered a new era with the broad availability of multi-modal, internet enabled, context enabled 3G devices. We are able to now create context aware applications that have much richer capabilities than were previously possible on mobile platforms. Our goal was to look at overcoming the limitation of current mobile applications using Context Aware Computing, RESTful Web Services, and Cloud Computing.
Using context-aware computing, we can incorporate user inputs, sensor inputs, and information from the web to enable an applications that are context aware and "adjust" to the user's current context, making for a rich user experience. This context-aware functionallity allows the application to ease the users input burden where appropriate. This helps overcome some of the input limitations of small form-factor devices.
Using RESTful web-services and cloud computing allow architectures to increase the computational, security, and storage capabilities of the traditional mobile application. The ability to be continuously connected to the internet, and the ease of consumption of RESTful web-services allows software architectures to be created that can use the device for human and sensor input, which is then passed off to a RESTful web-service for memory intensive and computationally intensive tasks.
Using Cloud Computing, architectures can be created that use cloud based authentication/authorization, and in which a mobile application can now store larger amounts of data than are typically possible in mobile applications. This ability to use cloud storage for mobile application storage greatly enhances the capabilities of modern mobile computing applications.
This session will be an overview of how to design and implement these architectures in order to modernize mobile applications. We will stay largely platform agnostic but ultimately implement in iPhone/Cocoa. The implementation is obviously secondary to a good design, and the design can be implemented on any multimodal context-enabled mobile device.
Fall 2008
Developing Software for the iPhone and iTouch: A Case Study
Jason Crawforth, Chief Strategy Officer, Mobile Dataforce
Pete Fransen, Senior Software Engineer, Mobile Dataforce
Dave Carter, Senior Software Engineer, Mobile Dataforce
12:40 PM
13th November, 2008
MEC 114
Come see the behind the scenes tour of a new software application using 3D modeling for iPhone and iTouch that will be premiering in the Apple iTunes store in a few weeks! Jason will provide the architectural highlights of the application. Pete and Dave will give an overview of the software development process for the iPhone/iTouch devices with specific details on the technical issues faced in developing their application.
Spring 2008
Don't like folding? Let your computer do it and cure cancer while you're at it!
Joseph Coffland, President, Cauldron Development LLC
9:40 AM
May 9, 2008
MEC 307
Folding@Home is a distributed computing project at Stanford University which aims to find cures for cancers and other diseases by simulating the behavior of proteins. Proteins are tiny molecules which carry out many critical functions in our bodies. Proteins start out as long chain molecules which self-assemble or fold into shapes that determine how they function. It is believed that diseases such as Alzheimer's, cystic fibrosis, Mad Cow disease and many cancers are caused by what is called protein misfolding. It is hoped that by better understanding protein folding, drugs can be designed which can prevent protein misfolding and cure the related diseases.
Although protein folding generally occurs within a few tens of microseconds, simulating these relatively short events accurately takes massive computing resources. This is why the Folding@Home project needs your help. By downloading and installing a small program you can donate the computing power of your home computer or laptop when you are not using it. Hundreds of thousands of computers are already connected to the Folding@Home network which in 2007 helped the project become the world’s first computing system to break the one petaFLOP barrier.
Industry View of Technology Trends & Careers
Ivan Lumala, Microsoft Corporation
3:40 PM
April 29, 2008
MEC 114
Ivan Lumala is an Academic Relations Manager at Microsoft Corporation. In this role, Ivan cultivates and nurtures strategic relationships between Microsoft and partner universities in six western states. Ivan has worked at Microsoft for the past 16 years in several technical and managerial roles, including 12 years on the Visual Studio team as a Software Design Engineer in Test and a lead Quality Assurance Manager. He holds a masters degree in software engineering from Seattle University, and an undergraduate degree in Computer Science from Seattle Pacific University. Ivan is also the co-founder of Leadership Advancement International
(LAI) www.laiweb.org, an organization with a mission to bridge the leadership divide in developing countries.
Run-time Adaptation in Future Multi-Core Architectres
Dan Conners, University of Colorado, Boulder
4:40pm
April 10, 2008
MEC 114
Emerging multi-core processor designs create an unparalleled computing
paradigm capable of advancing scientific areas, including medicine, data
mining, biology, physics, and earth sciences. These designs are the cornerstone
of all future general-purpose and high-performance computing
systems. However, a crisis is developing because the trends in multi-core
hardware technology have advanced far ahead of the advances in software
technology. Realizing the full potential of future multi-core machines requires
building new run-time systems capable of adapting thread-level behavior to
changing applications needs as well as the execution environment change. In
turn, multi-core architectures create can enable new services such as
run-time profiling, optimization, fault tolerance and security. This talk will
explore the topic of run-time adaptation in the future areas having significant
influence in computer systems design: performance of shared resources,
power related issues (energy/temperature) and fault tolerance/reliability.
Volta: Parting the Cloud
Wes Dyer, Microsoft
12:40pm-1:30pm
26th Feb. 2008
MEC 114
Writing data-intensive, asynchronous, multi-tier applications for the
web is difficult. It is done either by experts who piece together a
plethora of languages, tools, and frameworks or by depending on
technology that is not widely available. Volta radically simplifies web
programming while requiring only standards-based technology. Using
Volta, programmers can write web applications that are correct, clear,
secure, robust, and beautiful. This talk will introduce the technical
underpinnings of the Volta projects, and demonstrate practical
examples of its usage.
Bio: Wes Dyer is a member of the Volta project - http://labs.live.com/volta/
at Microsoft where he designs and builds programming languages and tools
to help developers with concurrency, distributed computation, web
programming,and data integration. He is also a member of the C#
language design team. Previously, he worked on the C# 3.0 compiler
adding support for functional programming and Language INtegrated
Querying (LINQ).
Spring 2007
Stock Markets & Algorithmic Trading
Paul Price & Brian Hill
Clearwater Advisors, LLC
Thursday, March 22nd, 2007
3:40pm - 4:30pm, MEC 114
Pizza and drinks will be served!
Balihoo – Search 2.0
Vince Martino – Chief Operating Officer, Alex Mueller – Director of Software Engineering, Mike Moser – DBA, Software Developer
Thursday, March 8th, 2007
3:40pm-4:30pm, MEC 114
Fall 2006
Immersive and Interactive Spaces: Computer Vision in Support of Human-Computer Interaction over Wide Areas
Dr. Christopher Jaynes,
Director of Collaborative Rendering Environment
Research Laboratory—University of Kentucky
Thursday, August 31, 2006
2:00 PM MEC 114
|
