Distributed Stream Processing with the DUP System
Christian Grothoff, PhD (UCLA)

It is now well established that the device scaling predicted by
Moore's Law is no longer a viable option for increasing the clock
frequency of future uniprocessor systems at the rate that had been
sustained during the last two decades.  As a result, future systems
are rapidly moving from uniprocessor to multiprocessor configurations
so as to use parallelism instead of frequency scaling as the
foundation for increased computational capacity.  The cost of
developing software that can utilize the capacity of these machines
typically far exceeds the cost of the hardware.

This talk will introduce the DUP System, a new domain-specific
language for productive parallel and distributed stream processing.
The primary goal for the DUP System is to enable productive
development of parallel and distributed stream-processing
applications.  In contrast to many other HPC languages, DUP generally
allows fast and inexpensive migration of legacy applications.  DUP is
also highly portable and allows developers to contribute components in
almost any language.


About the presenter:

Dr. Christian Grothoff is an Emmy-Noether research group leader at the
Technische Universitaet Muenchen.  He was a member of the core
development team for the X10 language and runtime system.  His thesis
on expressive type systems for object-oriented languages included a
design for a dependent type system for X10.  He is the maintainer of
several GNU projects, including GNUnet, GNU's framework for secure
peer-to-peer networking.  He recently created the DUP System to bring
parallel programming to the masses.
NUnet, GNU's framework for secure
peer-to-peer networking.  He recently created the DUP System to bring
parallel programming to the masses.