topic How to export the results of PARDISO numerical decomposition in Intel® oneAPI Math Kernel Library

How to export the results of PARDISO numerical decomposition

HUANGK — Sat, 30 Apr 2022 09:45:43 GMT

Hello！

I want to use PARDISO for matrix decomposition, and I want to export the results of the decomposition, for example, I want to export the L matrix after the LLT decomposition, how do I implement this method? If PARDISO can't implement this step, is there any way to do LLT decomposition for a large sparse matrix?

Thank you for your help

Re:How to export the results of PARDISO numerical decomposition

VidyalathaB_Intel — Mon, 02 May 2022 10:37:26 GMT

Hi Huang,

Thanks for reaching out to us.

>>I want to export the L matrix after the LLT decomposition, how do I implement this method?

Could you please confirm if you are asking to know whether we could extract factors from pardiso?

If yes, this feature is already enabled for the distributed cluster sparse solver via cluster_sparse_solver_export functionality. So, as a workaround, we suggest you try it. You can use the cluster sparse solver with a single MPI process.

Reference Link: https://www.intel.com/content/www/us/en/develop/documentation/onemkl-developer-reference-c/top/sparse-solver-routines/parallel-direct-sp-solver-for-clusters-iface/cluster-sparse-solver-export.html

You can also refer to the example cl_solver_export_c.c under C:\Program Files (x86)\Intel\oneAPI\mkl\2022.0.2\examples\examples_cluster_c\c_mpi\cluster_sparse_solver\source directory.

And here is the thread with a similar discussion which you can refer to for more details.

https://community.intel.com/t5/Intel-oneAPI-Math-Kernel-Library/MKL-PARDISO-get-L-U-factors/m-p/1265720#M31046

Please let us know if it helps in answering your queries.

Regards,

Vidya.

Re:How to export the results of PARDISO numerical decomposition

VidyalathaB_Intel — Tue, 03 May 2022 10:30:53 GMT

Hi Huang,

Thanks for accepting our solution.

As the issue is resolved we are closing this thread. Please post a new question if you need any additional information from Intel as this thread will longer be monitored.

Have a Great Day!

Regards,

Vidya.