λondon HUG

HUG Update: Nested Data Parallelism in Haskell

May 1, 2007 on 5:29 pm | By Neil | In News, Meetings |

The following is the abstract of the talk that will be given at the inaugural meeting of the London HUG:

Nested data parallelism in Haskell
Simon Peyton Jones, Microsoft Research

There are many approaches to exploiting multi-cores, but a particularly promising one is the “data-parallel” paradigm, because it combines massive parallelism (on both shared and distributed memory) with a simple, single-control-flow programming model.

Alas, data-parallel programming is usually restricted to “flat” data parallelism, which is good for implementers but bad for programmers. In particular, all the parallelism must appear at a single point, which excludes useful paradigms such as divide-and-conquer, and inhibits modular programming.

I’ll describe the “nested” data parallel programming model, first developed in the 90’s. It is great for programmers but much harder to implement; as a result, it’s virtually unknown in practice. We are working on fixing this, by building a high-performance implementation in Haskell. I’ll tell you how it works and give you some numbers.

After Simon’s talk, I would like to propose an open discussion, kicking off with each person sharing some background about their interest in Haskell and level of experience (the answer “no experience yet” will be absolutely acceptable — we welcome everybody from newbies to gurus). It would also be great to hear people’s ideas for what they would like to get from future HUG meetings and what format they should take.

Finally, for anybody so inclined, the nearby Slaughtered Lamb has been suggested as a suitable establishment for post-HUG refreshments. Lets hope nobody ends up like the lamb.