Reducing Latency Cost in 2D Sparse Matrix Partitioning Models

Parallel Computing - Netherlands

doi 10.1016/j.parco.2016.04.004