GRAph Parallel Actor Language — A Programming Language for Parallel Graph Algorithms

We introduce a domain-specific language, GRAph PArallel Actor Language, that enablesparallel graph algorithms to be written in a natural, high-level form. GRAPAL is based onour GraphStep compute model, which enables a wide range of parallel graph algorithmsthat are high-level, deterministic, free from race conditions, and free from deadlock. Pro-grams written in GRAPAL are easy for a compiler and runtime to map to efficient parallelfield programmable gate array (FPGA) implementations. We show that the GRAPAL com-piler can verify that the structure of operations conforms to the GraphStep model. Weallocate many small processing elements in each FPGA that take advantage of the highon-chip memory bandwidth (5x the sequential processor) and process one graph edge perclock cycle per processing element. We show how to automatically choose parameters forthe logic architecture so the high-level GRAPAL programming model is independent of thetarget FPGA architecture. We compare our GRAPAL applications mapped to a platformwith four 65 nm Virtex-5 SX95T FPGAs to sequential programs run on a single 65 nmXeon 5160. Our implementation achieves a total mean speedup of 8x with a maximumspeedup of 28x. The speedup per chip is 2x with a maximum of 7x. The ratio of energyused by our GRAPAL implementation over the sequential implementation has a mean of1/10 with a minimum of 1/80.