SuperLU Distributed 8.2.1
Distributed memory sparse direct solver
|
Performs LU factorization in 3D process grid. More...
#include "superlu_ddefs.h"
Functions | |
int_t | pdgstrf3d (superlu_dist_options_t *options, int m, int n, double anorm, trf3Dpartition_t *trf3Dpartition, SCT_t *SCT, dLUstruct_t *LUstruct, gridinfo3d_t *grid3d, SuperLUStat_t *stat, int *info) |
Performs LU factorization in 3D process grid.
Copyright (c) 2003, The Regents of the University of California, through Lawrence Berkeley National Laboratory (subject to receipt of any required approvals from U.S. Dept. of Energy)
All rights reserved.
The source code is distributed under BSD license, see the file License.txt at the top-level directory.
-- Distributed SuperLU routine (version 7.2) -- Lawrence Berkeley National Lab, Georgia Institute of Technology, Oak Ridge National Lab May 12, 2021 Last update: December 12, 2021 v7.2.0 */ /*!Purpose ======= PDGSTRF3D performs the LU factorization in parallel using 3D process grid, which is a communication-avoiding algorithm compared to the 2D algorithm. Arguments ========= options (input) superlu_dist_options_t* The structure defines the input parameters to control how the LU decomposition will be performed. The following field should be defined: o ReplaceTinyPivot (yes_no_t) Specifies whether to replace the tiny diagonals by sqrt(epsilon)*norm(A) during LU factorization. m (input) int Number of rows in the matrix. n (input) int Number of columns in the matrix. anorm (input) double The norm of the original matrix A, or the scaled A if equilibration was done. trf3Dpartition (input) trf3Dpartition* Matrix partitioning information in 3D process grid. SCT (input/output) SCT_t* Various statistics of 3D factorization. LUstruct (input/output) dLUstruct_t* The data structures to store the distributed L and U factors. The following fields should be defined: o Glu_persist (input) Glu_persist_t* Global data structure (xsup, supno) replicated on all processes, describing the supernode partition in the factored matrices L and U: xsup[s] is the leading column of the s-th supernode, supno[i] is the supernode number to which column i belongs. o Llu (input/output) dLocalLU_t* The distributed data structures to store L and U factors. See superlu_ddefs.h for the definition of 'dLocalLU_t'. grid3d (input) gridinfo3d_t* The 3D process mesh. It contains the MPI communicator, the number of process rows (NPROW), the number of process columns (NPCOL), and replication factor in Z-dimension. It is an input argument to all the 3D parallel routines. Grid3d can be initialized by subroutine SUPERLU_GRIDINIT3D. See superlu_defs.h for the definition of 'gridinfo3d_t'. stat (output) SuperLUStat_t* Record the statistics on runtime and floating-point operation count. See util.h for the definition of 'SuperLUStat_t'. info (output) int* = 0: successful exit < 0: if info = -i, the i-th argument had an illegal value > 0: if info = i, U(i,i) is exactly zero. The factorization has been completed, but the factor U is exactly singular, and division by zero will occur if it is used to solve a system of equations.
int_t pdgstrf3d | ( | superlu_dist_options_t * | options, |
int | m, | ||
int | n, | ||
double | anorm, | ||
trf3Dpartition_t * | trf3Dpartition, | ||
SCT_t * | SCT, | ||
dLUstruct_t * | LUstruct, | ||
gridinfo3d_t * | grid3d, | ||
SuperLUStat_t * | stat, | ||
int * | info | ||
) |