26-28 Jun 2019 Bordeaux (France)
Improving nested task-parallelism in the LU Factorization of H-Matrices
Rocío Carratalá-Sáez  1@  , Enrique S. Quintana-Ortí  2  
1 : Universitat Jaume I
2 : Universidad Politécnica de Valencia

Hierarchical matrices (H-matrices) lie in-between dense and sparse scenarios. Therefore, it is natural to tackle the LU factorization of H-Matrices via a task-parallel approach, which has recently reported successful results for related linear algebra problems. In this work, we will describe how to discover the data-flow parallelism intrinsic to the operation at execution time, via the analysis of data dependencies based on the memory addresses of the tasks' operands. This is especially challenging for H-matrices, as the data structures dimensions vary during the execution.

Online user: 1