HPC Graph Analysis



We present a graph theory benchmark representative of computational kernels in computational biology, complex network analysis, and national security. This benchmark is based on the HPCS Scalable Synthetic Compact Applications graph analysis (SSCA#2) benchmark. SSCA#2 is characterized by integer operations, a large memory footprint, and irregular memory access patterns. It has multiple kernels accessing a single data structure representing a weighted, directed multigraph. In addition to a kernel to construct the graph from the input tuple list, there are three additional computational kernels to operate on the graph. Each of the kernels requires irregular access to the graph's data structure, and it is possible that no single data layout will be optimal for all four computational kernels.


Related Publications

The written specification is the primary reference document for the current benchmark version (SSCA#2 v2.1).

The following papers discuss algorithms and implementations of earlier versions: