CSAIL Event Calendar

Thesis Defense: Programming in an Era of Heterogeneous Parallelism

How do we maintain programmer productivity in an era where both the number and diversity of networked processing elements are exploding? I offer a partial solution in the form of WaveScript, a programming environment for embedded, distributed stream-processing that has been used in a variety of sensor network applications. Sensor networks are a prime example of heterogeneous parallelism; they are distributed collections of phones, low-power embedded devices, and PCs each of which frequently contains multiple (heterogeneous) processing cores.

WaveScript is a high-performance general purpose streaming language based on asynchronous dataflow, suited for applications from digital signal processing to financial analysis and map/reduce data processing. A WaveScript program generates a graph of communicating actors; the compiler profiles these actors on various target platforms and then solves the partitioning problem: spreading actors across heterogeneous devices as well as across homogeneous multicores. To achieve performance and platform portability, the WaveScript solution involves metaprogramming, novel type systems, optimizations, and garbage collection strategies. In this talk I will focus on metaprogramming, domain specific optimizations, and program partitioning.

See other events happening in January 2009