I'm currently using Intel MKL library to solve a linear system.
As far as I know, Intel MKl library do not support MPI parallelization.
- Previously
- I have one big target system to calculate => thus building one big linear system to solve
- What I'm planning is
- Split the big system into pieces (to each MPI processes) => building a few small lineary systems (independent to each other) to solve
This is different from what parallel solvers would generally do (distribute 'one' big matrix to MPI processes). I will have independent a few small matrice, and will make MPI processes would solve them independently.
My question is that, is Intel MKL Pardiso solver could be utilized in this way (called simultaneously from a few MPI processes for independt problems) ?
I add a picture to describe what I'm going to do in more understandable way.

The current version of MKL provides the cluster version of Intel MKL Pardiso ( cluster_sparse_solver). You may check the link to the MKL Reference follow the link:https://software.intel.com/content/www/us/en/develop/documentation/mkl-developer-reference-c/top/sparse-solver-routines/parallel-direct-sparse-solver-for-clusters-interface.html.