MUMPS Main Features

Solution of large linear systems with
symmetric positive definite matrices
general symmetric matrices
general unsymmetric matrices  Version for complex arithmetic

Parallel factorization and solve phases
(uniprocessor version also available)  Iterative refinement and backward error analysis

Various matrix input formats
assembled, distributed, elemental format
 Partial factorization and Schur complement matrix (centralized or 2D blockcyclic)
 Interfaces to MUMPS: Fortran, C, Matlab and Scilab
 Several reorderings interfaced: AMD, QMAD, AMF, PORD, METIS, PARMETIS, SCOTCH, PTSCOTCH
Recent features
 Out of core numerical phases
 Reduced/condensed righthand side
 Detection of null pivots, null space basis estimate

Symmetric indefinite matrices:
preprocesssing and 2by2 pivots
 Parallel analysis and matrix scaling
 Computation of the determinant (with an option to discard factors)
 Sparse multiple righthand side, distributed solution; Exploitation of sparsity in the righthand sides
 Forward elimination during factorization
 Computation of selected entries in the inverse of a matrix
A fully asynchronous distributed solver (VAMPIR trace)
Implementation
 Distributed Multifrontal Solver (F90, MPI based) with shared memory OpenMP directives;
 Dynamic Distributed Scheduling to accomodate both numerical fillin and multiuser environment;
 Use of BLAS, BLACS, ScaLAPACK.
Partially funded by CEC ESPRIT IV long term research project
 No. 20160 (PARASOL)