Chapel presentation, Steve Deitz
(10:30am – 12:00am)
Chapel is a new parallel programming language designed to improve the
productivity of users of large-scale supercomputers as well as
small-scale, multicore computers and workstations. Chapel aims to
vastly improve programmability over current parallel programming models
while supporting performance and portability at least as good as
today’s technologies. It has been under development at Cray
Inc. as part of the DARPA High Productivity Computing Systems (HPCS)
program. In the morning, an overview of Chapel will be presented with a
focus on its task- and data-parallel abstractions and its support for
systems with distributed memory.
Tutorial on Chapel and
X10, Steve Deitz and Marc Tajchman (2:00pm – 5:00pm)
In the afternoon, we will look at an example computation, Heat
Transfer, in both Chapel and X10 (a new parallel programming language
designed at IBM as part of the DARPA High Productivity Computing
Systems (HPCS) program1). We will compare and
contrast these two
languages with respect to this example computation, and we will present
live demos of this example. In addition, we will discuss the relative
merits of these languages in comparison to other parallel programming
technologies including MPI, UPC, and Co-Array Fortran.
Steve Deitz
is a Software Engineer at Cray Inc. He currently works on the design
and implementation of the Chapel parallel language in his role as
implementation lead for that project. Steve received his Ph.D. in
Computer Science & Engineering from the University of
Washington in 2005 for his work on extending the ZPL parallel array
language to handle more dynamic and advanced parallel computations. He
received a B.A. in Computer Science and Mathematics from Bowdoin
College in 1998.
Marc Tajchman is a Scientific Computing Engineer at CEA (French Atomic
Energy Commission). Working with a multi-disciplinary team of engineers
and scientists, he currently designs and develops numerical simulation
platforms. He is particularly interested by new languages and tools for
high performance computing. Marc received a PhD in Applied Mathematics
(University of Paris XI) in 1994 and a Construction Engineering degree
(ULB, Brussels, Belgium) in 1991.