Intel® oneAPI Math Kernel Library
Ask questions and share information with other developers who use Intel® Math Kernel Library.
Announcements
This community is designed for sharing of public information. Please do not share Intel or third-party confidential information here.
6588 Discussions

accessing multiple factorizations in PARDISO

ghisistru_polimi_it
119 Views

Hello all,

is there a way to access to access multiple factorizations of sparse matrices between subsequent PARDISO calls?

More precisely, I solve the usual problem Ax=B where B is a multi-column rhs for several subsystems in different time steps of my calculation.

Since in the manual entry for the PARDISO input parameter MAXFCT it is written:
"PARDISO can process several matrices with identical matrix sparsity pattern and store the factors of these matrices at the same time. Matrices with a different sparsity structure can be kept in memory with different memory address pointers pt."
I wonder if I would be able to avoid the factorizations after the first time step.
In other words, is it possible to retrieve (from memory or whatever) the multiple factorizations without asking PARDISO to actually redo the calculation? I tried i.e. declaring the data address pointers PT as global variables, but it failed.

Thank you in advance for any advice and regards.
Aldo Ghisi

p.s. I use Intel Fortran 11.1.048 with MKL 10.33

0 Kudos
1 Reply
mecej4
Black Belt
119 Views
I don't know of any linear equation solver that does not let the user reuse the LU (LDU) factorization with different right hand sides. The quote from the manual that you gave concerns matrices A1, A2, ... which are structurally identical but have different numerical values for the non-zero coefficients. In your case A1, A2, ... are all the same.
Reply