The Combinatorial BLAS (CombBLAS) is an extensible distributed-memory parallel
graph library offering a small but powerful set of linear algebra primitives
specifically targeting graph analytics.

* The Combinatorial BLAS development influences the Graph BLAS standardization
  process.
* CombBLAS achieves scalability via its two dimensional distribution and
  coarse-grained parallelism.
* CombBLAS powers HipMCL, a highly-scalable parallel implementation of the
  Markov Cluster Algorithm (MCL).

WWW: https://people.eecs.berkeley.edu/~aydin/CombBLAS/html/