, INFO(1) is 0 if the call to MUMPS was successful, negative if an error occurred (see Section 5), or positive if a warning is returned
, INFO(2) holds additional information about the error or the warning. If INFO(1)=-1, INFO(2) is the processor number (in communicator mumps_par%COMM) on which the error was detected
, INFO(3) -after analysis: Estimated real space needed on the processor for factors
, INFO(4) -after analysis: Estimated integer space needed on the processor for factors
, INFO(5) -after analysis: Estimated maximum front size on the processor
, INFO(6) -after analysis: Number of nodes in the complete tree. The same value is returned on all processors
after analysis: Minimum value of MAXIS estimated by the analysis phase to run the numerical factorization successfully ,
, INFO(8) -after analysis: Minimum value of MAXS estimated by the analysis phase to run the numerical factorization successfully
after factorization: Size of the real space used on the processor to store the LU factors ,
after factorization: Size of the integer space used on the processor to store the LU factors ,
, INFO(11) -after factorization: Order of the largest frontal matrix processed on the processor
, INFO(12) -after factorization: Number of off-diagonal pivots encountered on the processor or number of negative pivots if the SYM=1
after factorization: The number of uneliminated variables, corresponding to delayed pivots, sent to the father ,
, INFO(14) -after factorization: Number of memory compresses on the processor
, INFO(15) -after analysis: estimated total size (in millions of bytes) of all MUMPS internal data for running numerical factorization
after factorization: total size (in millions of bytes) of all MUMPS internal data used during numerical factorization. and mumps_par%INFOG : mumps_par%RINFOG is a double precision array of dimension 20. It contains the following global information on the execution of MUMPS: RINFOG(1) -after analysis: The estimated number of floating ,
, RINFOG(2) -after factorization: The total number of floating-point operations (on all processors) for the assembly process
, RINFOG(3) -after factorization: The total number of floating-point operations (on all processors) for the elimination process
, RINFOG(4) to RINFOG(11) -after solve with error analysis: Only returned on the host process if ICNTL(11) 0. See description of ICNTL
, RINFOG(12) -RINFOG(20) are not used in the current version
, mumps_par%INFOG is an integer array of dimension 40
, INFOG(1) is 0 if the call to MUMPS was successful, negative if an error occurred (see Section 5), or positive if a warning is returned
, INFOG(2) holds additional information about the error or the warning
, :2) is the same on all processors. It has the value of INFO(1:2) of the processor which returned with the most negative INFO(1) value. For example, if processor Ô returns with INFO(1)=-13, and INFO(2)=10000, then all other processors will return with INFOG(1)=-13 and INFOG(2)=10000, The difference between INFOG(1:2) and INFO(1:2) is that INFOG
, INFOG(3) -after analysis: Total estimated real workspace for factors on all processors
, INFOG(4) -after analysis: Total estimated integer workspace for factors on all processors
, INFOG(5) -after analysis: Estimated maximum front size in the complete tree
, INFOG(6) -after analysis: Number of nodes in the complete tree
, :8) : not significant
, after factorization: Total real space to store the LU factors. INFOG(10) -after factorization: Total integer space to store the LU factors
, INFOG(11) -after factorization: Order of largest frontal matrix
, INFOG(12) -after factorization: Total number of off-diagonal pivots or negative pivots if SYM=1
, INFOG(13) -after factorization: Total number of delayed pivots
, INFOG(14) -after factorization: Total number of memory compresses
, INFOG(15) -after solution: Number of steps of iterative refinement
, INFOG(16) -after analysis: Estimated size (in million of bytes) of all MUMPS internal data for running factorization: value on the most memory consuming processor
, INFOG(17) -after analysis: Estimated size (in millions of bytes) of all MUMPS internal data for running factorization: sum over all processors
, INFOG(18) -after factorization: Size in millions of bytes of all MUMPS internal data during factorization: value on the most memory consuming processor
, INFOG(19) -after factorization: Size in millions of bytes of all MUMPS internal data during factorization: sum over all processors
, INFOG(20) -after analysis: Estimated number of entries in the factors
The approximate minimum degree algorithm, 2nd SIAM Conference on Sparse Matrices, 1996. ,
Vectorization of a multiprocessor multifrontal code, Int. J. of Supercomputer Applics, vol.3, pp.41-59, 1989. ,
A fully asynchronous multifrontal solver using distributed dynamic scheduling, SIAM Journal on Matrix Analysis and Applications, vol.23, issue.1, pp.15-41, 2001. ,
URL : https://hal.archives-ouvertes.fr/hal-00808293
Multifrontal solvers within the PARASOL environment, Applied Parallel Computing, PARA'98, pp.7-11, 1998. ,
URL : https://hal.archives-ouvertes.fr/hal-00856857
Parallélisation de la factorisation LU de matrices creuses non-symétriques pour des architectures à mémoire distribuée, Calculateurs Parallèles Réseaux et Systèmes Répartis, vol.10, issue.5, pp.509-520, 1998. ,
Multifrontal parallel distributed symmetric and unsymmetric solvers, Comput. Methods Appl. Mech. Eng, vol.184, pp.501-520, 2000. ,
URL : https://hal.archives-ouvertes.fr/hal-00856651
PARASOL. An integrated programming environment for parallel sparse matrix solvers, High-Performance Computing, pp.79-90, 1999. ,
URL : https://hal.archives-ouvertes.fr/hal-00856860
Solving sparse linear systems with sparse backward error, SIAM Journal on Matrix Analysis and Applications, vol.10, pp.165-190, 1989. ,
, , 1997.
On rank-revealing factorizations, SIAM J. Matrix Anal. Appl, vol.15, pp.592-622, 1994. ,
On the automatic scaling of matrices for Gaussian elimination, J. Inst. Maths. Applics, vol.10, pp.118-124, 1972. ,
Algorithm 679. A set of Level 3 Basic Linear Algebra Subprograms, ACM Transactions on Mathematical Software, vol.16, pp.1-17, 1990. ,
Algorithm 679. A set of Level 3 Basic Linear Algebra Subprograms: model implementation and test programs, ACM Transactions on Mathematical Software, vol.16, pp.18-28, 1990. ,
Algorithm 575. Permutations for a zero-free diagonal, ACM Transactions on Mathematical Software, vol.7, pp.387-390, 1981. ,
The design and use of algorithms for permuting large entries to the diagonal of sparse matrices, SIAM Journal on Matrix Analysis and Applications, vol.20, issue.4, pp.889-901, 1999. ,
On algorithms for permuting large entries to the diagonal of a sparse matrix, SIAM Journal on Matrix Analysis and Applications, vol.22, issue.4, pp.973-996, 2001. ,
The multifrontal solution of indefinite sparse symmetric linear systems, ACM Transactions on Mathematical Software, vol.9, pp.302-325, 1983. ,
The multifrontal solution of unsymmetric sets of linear systems, SIAM Journal on Scientific and Statistical Computing, vol.5, pp.633-641, 1984. ,
METIS -A Software Package for Partitioning Unstructured Graphs, Partitioning Meshes, and Computing Fill-Reducing Orderings of Sparse Matrices -Version 4.0. University of Minnesota, 1998. ,
Towards a tighter coupling of bottom-up and top-down sparse matrix ordering methods, BIT, vol.41, issue.4, pp.800-841, 2001. ,