|
|
@ -1,25 +1,20 @@
|
|
|
|
This directory contains the PSBLAS library, version 3.5.
|
|
|
|
This directory contains the PSBLAS library, version 3.5.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
The architecture of the Fortran 2003 sparse BLAS is described in:
|
|
|
|
The architecture of the Fortran 2003 sparse BLAS is described in
|
|
|
|
S. Filippone, A. Buttari
|
|
|
|
S. Filippone, A. Buttari:
|
|
|
|
|
|
|
|
Object-Oriented Techniques for Sparse Matrix Computations in Fortran
|
|
|
|
Object-Oriented Techniques for Sparse Matrix Computations in Fortran
|
|
|
|
2003,
|
|
|
|
2003, ACM Trans. on Math. Software, vol. 38, No. 4, 2012.
|
|
|
|
ACM Trans. on Math. Software, vol. 38, No. 4, 2012.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
The ideas are explored further with the paper:
|
|
|
|
The ideas are explored further with the paper:
|
|
|
|
|
|
|
|
|
|
|
|
V. Cardellini, S. Filippone and D. Rouson
|
|
|
|
V. Cardellini, S. Filippone and D. Rouson
|
|
|
|
Design Patterns for sparse-matrix computations on
|
|
|
|
Design Patterns for sparse-matrix computations on hybrid CPU/GPU
|
|
|
|
hybrid CPU/GPU platforms,
|
|
|
|
platforms, Scientific Programming, 22(2014), pp.1-19.
|
|
|
|
Scientific Programming, 22(2014), pp.1-19.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Version 1.0 of the library is described in:
|
|
|
|
Version 1.0 of the library was described in:
|
|
|
|
|
|
|
|
S. Filippone, M. Colajanni
|
|
|
|
S. Filippone, M. Colajanni
|
|
|
|
PSBLAS: A library for parallel linear algebra computation on sparse matrices
|
|
|
|
PSBLAS: A library for parallel linear algebra computation on sparse
|
|
|
|
ACM Trans. on Math. Software, 26(4), Dec. 2000, pp. 527-550.
|
|
|
|
matrices, ACM Trans. on Math. Software, 26(4), Dec. 2000, pp. 527-550.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
UTILITIES
|
|
|
|
UTILITIES
|
|
|
@ -30,23 +25,20 @@ Harwell-Boeing and MatrixMarket file formats.
|
|
|
|
DOCUMENTATION
|
|
|
|
DOCUMENTATION
|
|
|
|
|
|
|
|
|
|
|
|
See docs/psblas-3.5.pdf; an HTML version of the same document is
|
|
|
|
See docs/psblas-3.5.pdf; an HTML version of the same document is
|
|
|
|
available in docs/html.
|
|
|
|
available in docs/html. Please consult the sample programs, especially
|
|
|
|
Please consult the sample programs, especially
|
|
|
|
|
|
|
|
test/pargen/psb_[sd]_pde[23]d.f90
|
|
|
|
test/pargen/psb_[sd]_pde[23]d.f90
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
OTHER SOFTWARE CREDITS
|
|
|
|
OTHER SOFTWARE CREDITS
|
|
|
|
|
|
|
|
|
|
|
|
We originally included a modified implementation of some of the
|
|
|
|
We originally included a modified implementation of some of the Sparker
|
|
|
|
Sparker (serial sparse BLAS) material; this has been completely
|
|
|
|
(serial sparse BLAS) material; this has been completely rewritten, way
|
|
|
|
rewritten, way beyond the intention(s) and responsibilities of the
|
|
|
|
beyond the intention(s) and responsibilities of the original developers.
|
|
|
|
original developers.
|
|
|
|
|
|
|
|
The main reference for the serial sparse BLAS is:
|
|
|
|
The main reference for the serial sparse BLAS is:
|
|
|
|
Duff, I., Marrone, M., Radicati, G., and Vittoli, C.
|
|
|
|
Duff, I., Marrone, M., Radicati, G., and Vittoli, C.
|
|
|
|
Level 3 basic linear algebra subprograms for sparse matrices: a user
|
|
|
|
Level 3 basic linear algebra subprograms for sparse matrices: a user
|
|
|
|
level interface
|
|
|
|
level interface, ACM Trans. Math. Softw., 23(3), 379-401, 1997.
|
|
|
|
ACM Trans. Math. Softw., 23(3), 379-401, 1997.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
INSTALLING
|
|
|
|
INSTALLING
|
|
|
@ -56,8 +48,8 @@ prerequisites (see also SERIAL below):
|
|
|
|
|
|
|
|
|
|
|
|
1. A working version of MPI
|
|
|
|
1. A working version of MPI
|
|
|
|
|
|
|
|
|
|
|
|
2. A version of the BLAS; if you don't have a specific version for
|
|
|
|
2. A version of the BLAS; if you don't have a specific version for your
|
|
|
|
your platform you may try ATLAS available from
|
|
|
|
platform you may try ATLAS available from
|
|
|
|
http://math-atlas.sourceforge.net/
|
|
|
|
http://math-atlas.sourceforge.net/
|
|
|
|
|
|
|
|
|
|
|
|
3. We have had good results with the METIS library, from
|
|
|
|
3. We have had good results with the METIS library, from
|
|
|
@ -68,49 +60,49 @@ prerequisites (see also SERIAL below):
|
|
|
|
4. If you have the AMD package of Davis, Duff and Amestoy, you can
|
|
|
|
4. If you have the AMD package of Davis, Duff and Amestoy, you can
|
|
|
|
specify --with-amd (see ./configure --help for more details).
|
|
|
|
specify --with-amd (see ./configure --help for more details).
|
|
|
|
|
|
|
|
|
|
|
|
The configure script will generate a Make.inc file suitable for
|
|
|
|
The configure script will generate a Make.inc file suitable for building
|
|
|
|
building the library.
|
|
|
|
the library. The script is capable of recognizing the needed libraries
|
|
|
|
The script is capable of recognizing the needed libraries with their
|
|
|
|
with their default names; if they are in unusual places consider adding
|
|
|
|
default names; if they are in unusual places consider adding the paths
|
|
|
|
the paths with --with-libs, or explicitly specifying the names in
|
|
|
|
with --with-libs, or explicitly specifying the names in --with-blas, etc.
|
|
|
|
--with-blas, etc. Please note that a common way for the configure script
|
|
|
|
Please note that a common way for the configure script to fail is to
|
|
|
|
to fail is to specify inconsistent MPI vs. plain compilers, either
|
|
|
|
specify inconsistent MPI vs. plain compilers, either directly or
|
|
|
|
directly or indirectly via environment variables; e.g. specifying the
|
|
|
|
indirectly via environment variables; e.g. specifying the Intel
|
|
|
|
Intel compiler with FC=ifort while at the same time having an
|
|
|
|
compiler with FC=ifort while at the same time having an MPIFC=mpif90
|
|
|
|
MPIFC=mpif90 which points to GNU Fortran. The best way to avoid this
|
|
|
|
which points to GNU Fortran.
|
|
|
|
situation is (in our opinion) to use the environment modules package
|
|
|
|
The best way to avoid this situation is (in our opinion) to use the
|
|
|
|
(see http://modules.sourceforge.net/), and load the relevant
|
|
|
|
environment modules package (see http://modules.sourceforge.net/), and
|
|
|
|
variables with (e.g.)
|
|
|
|
load the relevant variables with (e.g.)
|
|
|
|
|
|
|
|
module load gnu46 openmpi
|
|
|
|
module load gnu46 openmpi
|
|
|
|
This will delegate to the modules setup to make sure that the version
|
|
|
|
|
|
|
|
of openmpi in use is the one compiled with the gnu46 compilers.
|
|
|
|
This will delegate to the modules setup to make sure that the version of
|
|
|
|
After the configure script has completed you can always tweak the
|
|
|
|
openmpi in use is the one compiled with the gnu46 compilers. After the
|
|
|
|
Make.inc file yourself.
|
|
|
|
configure script has completed you can always tweak the Make.inc file
|
|
|
|
|
|
|
|
yourself.
|
|
|
|
|
|
|
|
|
|
|
|
After you have Make.inc fixed, run
|
|
|
|
After you have Make.inc fixed, run
|
|
|
|
make
|
|
|
|
|
|
|
|
to compile the library; go to the test directory and its
|
|
|
|
make
|
|
|
|
subdirectories to get test programs done.
|
|
|
|
|
|
|
|
If you specify --prefix=/path you can do make install and the
|
|
|
|
to compile the library; go to the test directory and its subdirectories
|
|
|
|
libraries will be installed under /path/lib, while the module files
|
|
|
|
to get test programs done. If you specify --prefix=/path you can do make
|
|
|
|
will be installed under /path/include.
|
|
|
|
install and the libraries will be installed under /path/lib, while the
|
|
|
|
|
|
|
|
module files will be installed under /path/include.
|
|
|
|
|
|
|
|
|
|
|
|
SERIAL
|
|
|
|
SERIAL
|
|
|
|
|
|
|
|
|
|
|
|
Configuring with --enable-serial will provide a fake MPI stub library
|
|
|
|
Configuring with --enable-serial will provide a fake MPI stub library
|
|
|
|
that enables running in pure serial mode; no MPI installation is
|
|
|
|
that enables running in pure serial mode; no MPI installation is needed
|
|
|
|
needed in this case (but note that the fake MPI stubs are only
|
|
|
|
in this case (but note that the fake MPI stubs are only guaranteed to
|
|
|
|
guaranteed to cover what we use internally, it's not a complete
|
|
|
|
cover what we use internally, it's not a complete replacement).
|
|
|
|
replacement).
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
LONG INTEGERS
|
|
|
|
LONG INTEGERS
|
|
|
|
|
|
|
|
|
|
|
|
We have an experimental flag --enable-long-integers that will enable having
|
|
|
|
We have an experimental flag --enable-long-integers that will enable
|
|
|
|
8-byte integer data, allowing an index space larger than 2G; some
|
|
|
|
having 8-byte integer data, allowing an index space larger than 2G; some
|
|
|
|
small cases have been tested but we do not offer full guarantee (yet).
|
|
|
|
small cases have been tested but we do not offer full guarantee (yet).
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
TODO:
|
|
|
|
TODO:
|
|
|
|
Fix all reamining bugs. Bugs? We dont' have any ! ;-)
|
|
|
|
Fix all reamining bugs. Bugs? We dont' have any ! ;-)
|
|
|
|
|
|
|
|
|
|
|
@ -119,9 +111,9 @@ The PSBLAS team.
|
|
|
|
|
|
|
|
|
|
|
|
RELATED SOFTWARE
|
|
|
|
RELATED SOFTWARE
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
If you are looking for more sophisticated preconditioners, you may be
|
|
|
|
If you are looking for more sophisticated preconditioners, you may be
|
|
|
|
interested in the package MLD2P4 from http://github.com/sfilippone/mld2p4-2
|
|
|
|
interested in the package MLD2P4 from
|
|
|
|
|
|
|
|
http://github.com/sfilippone/mld2p4-2
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Contact: https://github.com/sfilippone/psblas3
|
|
|
|
Contact: https://github.com/sfilippone/psblas3
|
|
|
|