Skip to content
Go to file


# Copyright (c) 2017,  Lawrence Livermore National Security, LLC.
# Produced at the Lawrence Livermore National Laboratory.
# Written by Ulrike Yang ( et al. CODE-LLNL-738-322.
# This file is part of AMG.  See files COPYRIGHT and README for details.
# AMG is free software; you can redistribute it and/or modify it under the
# terms of the GNU Lesser General Public License (as published by the Free
# Software Foundation) version 2.1 dated February 1999.
# This program is distributed in the hope that it will be useful, but WITHOUT
# FITNESS FOR A PARTICULAR PURPOSE.  See the terms and conditions of the
# GNU General Public License for more details.
# You should have received a copy of the GNU Lesser General Public License
# along with this program; if not, write to the Free Software Foundation,
# Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA

General description:

AMG is a parallel algebraic multigrid solver for linear systems arising from
problems on unstructured grids.  The driver provided with AMG builds linear 
systems for various 3-dimensional problems.
AMG is written in ISO-C.  It is an SPMD code which uses MPI and OpenMP 
threading within MPI tasks. Parallelism is achieved by data decomposition. The 
driver provided with AMG achieves this decomposition by simply subdividing 
driver provided with AMG achieves this decomposition by simply subdividing 
the grid into logical P x Q x R (in 3D) chunks of equal size.
For more information, see the amg.readme file in the docs directory of the

Building the Code

AMG uses a simple Makefile system for building the code.  All compiler and
link options are set by modifying the file 'AMG/Makefile.include'

To build the code, first modify the 'Makefile.include' file appropriately, 
(it is recommended to use the options -DHYPRE_BIGINT )
then type (in the AMG directory)


Other available targets are

  make clean        (deletes .o files)
  make veryclean    (deletes .o files, libraries, and executables)

To configure the code to run with:

1 - MPI only , add '-DTIMER_USE_MPI' to the 'INCLUDE_CFLAGS' line 
    in the 'Makefile.include' file and use a valid MPI.
2 - OpenMP with MPI, add vendor dependent compilation flag for OMP
3 - to be able to solve problems that are larger than 2^31-1,
    add '-DHYPRE_BIGINT'
4 - For additional optimizations in MPI add '-DHYPRE_USING_PERSISTENT_COMM'
5 - For additional optimizations in OpenMP add '-DHYPRE_HOPSCOTCH'

Figure of Merit (FOM)

For problem 1, 
there are 2 FOMs printed out at the end of each run:
nnz_AP / setup_time
nnz_AP * #iterations / solve time
Both need to be considered.

For problem 2,
one FOM needs to be considered:
nnz_AP * (#iterations + time_steps) / time


Algebraic multigrid benchmark





No packages published