HPC and Programmability

This post originally appeared on the Software Carpentry website.

Via Andrew Lumsdaine, a pointer to an interesting article in Communications of the ACM by Eugene Loh titled "The Ideal HPC Programming Language". A few key quotes:

These programmability studies began with a focus on programming languages, but the focus quickly shifted to other topics. Existing languages—notably Fortran...—proved remarkably adequate. Programming challenges stem mostly from other factors.

The DARPA HPCS program...sponsored the development of new programming languages: Chapel from Cray, Fortress from Sun, and X10 from IBM. Proponents of those languages would show early on how rewriting familiar HPC benchmarks in the new languages could reduce source-code volume substantially—tenfold reductions were not surprising—but rewriting these benchmarks even in Fortran achieved similar source-code reductions and corresponding improvements in expressivity.

In HPC, the mind-set is usually to program for performance rather than programmability even before establishing whether a particular section of code is performance sensitive or not.

Repeating some of these...studies on larger HPC programs would be interesting. In particular, it would be nice to move from self-contained programs that are small enough for one person to have written...to larger pieces of software...

The Cowichan Problems were designed to explore that last issue in an affordable way. If you're interested in helping us re-design them so that they better represent the full range of today's HPC applications, please let me know.