You’re listening to the Sun Microsystem’s Podcast Network. Welcome to another edition of
Innovating@Sun with your host Hal Stern. Today’s topic: Fortress 1.0. And now, here’s Hal Stern.
Hello. Welcome to another edition of Innovating at Sun. I’m your host, Hal Stern, Vice President of Global
Systems Engineering, and with me today is Dr. Eric Allen, a principal investigator in the Programming Language
Research Group at Sun Labs, and Eric is going to give us an update on where we are with the Fortress language.
So, Eric, welcome to the show and tell us a little bit about your Fortress.
Hi. So Fortress is a new programming language designed for high-performance computing but with high
programmer productivity. So we’ve tried to bring in lots of new ideas in programming language design from the
past decade and apply that to the specific domain of high-performance computing in an attempt to improve the
productivity of programmers in that domain. So Fortress, it includes lots of innovative features such as syntax
that’s based very closely on mathematical notation which allows programs to look very much like the
mathematics that they’re implementing. There’s also lots of support for parallelism, in fact, implicit parallelism
at many levels. In fact, it’s very difficult to even write a Fortress program that’s not parallel in some sense. And
we’re expecting that that will be very useful for targeting multi-core architectures as well as super computers and
everything in between.
So you make reference here to both productivity as well as super computers. Fortress came out of some work we
were doing in the DARPA high-productivity compute system, HPCS. Again, I think a lot of people look to the P there
as performance, and this is really about making developers more productive with the systems that we see coming
out in the next couple of years.
That’s right. Yes. One of the goals that DARPA had for HPCS was to decrease the amount of time it to