A Simple Synchronous Distributed-Memory Algorithm for the HPCC RandomAccess Benchmark

The RandomAccess benchmark as defined by the high performance computing challenge (HPCC) tests the speed at which a machine can update the elements of a table spread across global system memory, as measured in billions (giga) of updates per second (GUPS). The parallel implementation provided by HPCC typically performs poorly on distributed-memory machines, due to updates requiring numerous small point-to-point messages between processors. We present an alternative algorithm which treats the collection of P processors as a hypercube, aggregating data so that larger messages are sent, and routing individual datums through dimensions of the hypercube to their destination processor. The algorithm\´s computation (the GUP count) scales linearly with P while its communication overhead scales as log ₂(P), thus enabling better performance on large numbers of processors. The new algorithm achieves a GUPS rate of 19.98 on 8192 processors of Sandia\´s Red Storm machine, compared to 1.02 for the HPCC-provided algorithm on 10350 processors. We also illustrate how GUPS performance varies with the benchmark\´s specification of its "look-ahead"parameter. As expected, parallel performance degrades for small look-ahead values, and improves dramatically for large values