fdurastante
/
psblas3
mirror of https://github.com/sfilippone/psblas3.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Update README.md

Browse Source 
Updated the look of the README
openacc
Fabio Durastante 2 years ago committed by GitHub
parent d7cb13a283
commit 961a315936
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 120 additions and 77 deletions
Show Stats Download Patch File Download Diff File

197
README.md
Unescape Escape View File

@ -1,37 +1,35 @@
PSBLAS library, version 3.9 # PSBLAS library, version 3.9
===========================
The architecture of the Fortran 2003 sparse BLAS is described in:
>S. Filippone, A. Buttari. Object-Oriented Techniques for Sparse Matrix
>Computations in Fortran 2003, ACM Trans. on Math. Software, vol. 38, No.
4, 2012.
The ideas are explored further with the paper: The PSBLAS library, developed with the aim to facilitate the parallelization of computationally intensive scientific applications, is designed to address parallel implementation of iterative solvers for sparse linear systems through the distributed memory paradigm. It includes routines for multiplying sparse matrices by dense matrices, solving block diagonal systems with triangular diagonal entries, preprocessing sparse matrices, and contains additional routines for dense matrix operations. The current implementation of PSBLAS addresses a distributed memory execution model operating with message passing.
>V. Cardellini, S. Filippone and D. Rouson. Design Patterns for
>sparse-matrix computations on hybrid CPU/GPU platforms, Scientific
>Programming, 22(2014), pp.1-19.
Version 1.0 of the library is described in: The PSBLAS library version 3 is implemented in the Fortran 2003 programming language, with reuse and/or adaptation of existing Fortran 77 and Fortran 95 software, plus a handful of C routines.
>S. Filippone, M. Colajanni. PSBLAS: A library for parallel linear
>algebra computation on sparse matrices, ACM Trans. on Math. Software,
>26(4), Dec. 2000, pp. 527-550.
## References
UTILITIES
---------
The `test/util` directory contains some utilities to convert to/from
Harwell-Boeing and MatrixMarket file formats.
The architecture, philosophy and implementation details of the library are contained in the following papers:
DOCUMENTATION - The architecture of the Fortran 2003 sparse BLAS is described in:
------------- >S. Filippone, A. Buttari. Object-Oriented Techniques for Sparse Matrix
See docs/psblas-3.9.pdf; an HTML version of the same document is >Computations in Fortran 2003, ACM Trans. on Math. Software, vol. 38, No.
available in docs/html. Please consult the sample programs, especially 4, 2012.
test/pargen/psb_[sd]_pde[23]d.f90
- The ideas are explored further with the paper:
>V. Cardellini, S. Filippone and D. Rouson. Design Patterns for
>sparse-matrix computations on hybrid CPU/GPU platforms, Scientific
>Programming, 22(2014), pp.1-19.
- Version 1.0 of the library is described in:
>S. Filippone, M. Colajanni. PSBLAS: A library for parallel linear
>algebra computation on sparse matrices, ACM Trans. on Math. Software,
>26(4), Dec. 2000, pp. 527-550.
PSBLAS is the backbone of the Parallel Sparse Computation Toolkit ([PSCToolkit](https://psctoolkit.github.io/)) suite of libraries. See the paper:
> D’Ambra, P., Durastante, F., & Filippone, S. (2023). Parallel Sparse Computation Toolkit. Software Impacts, 15, 100463.
### Other Software credits
OTHER SOFTWARE CREDITS
----------------------
We originally included a modified implementation of some of the Sparker We originally included a modified implementation of some of the Sparker
(serial sparse BLAS) material; this has been completely rewritten, way (serial sparse BLAS) material; this has been completely rewritten, way
beyond the intention(s) and responsibilities of the original developers. beyond the intention(s) and responsibilities of the original developers.
@ -40,27 +38,8 @@ The main reference for the serial sparse BLAS is:
>linear algebra subprograms for sparse matrices: a user level interface, >linear algebra subprograms for sparse matrices: a user level interface,
>ACM Trans. Math. Softw., 23(3), 379-401, 1997. >ACM Trans. Math. Softw., 23(3), 379-401, 1997.
CUDA and GPU support ## Installing
--------------------
This version of PSBLAS incorporates into a single package three
entities that were previouslty separated:
1. PSBLAS -- the base library
2. PSBLAS-EXT -- a library providing additional storage formats
3. SPGPU -- a package of kernels for NVIDIA GPUs originally
written by Davide Barbieri and Salvatore Filippone;
see the license file cuda/License-spgpu.md
OpenACC
-------------------
There is a highly experimental version of an OpenACC interface,
you can access it by speficifying
--enable-openacc --with-extraopenacc="-foffload=nvptx-none=-march=sm_70"
where the argument to the extraopenacc option depends on the compiler
you are using (the example shown here is relevant for the GNU
compiler).
INSTALLING
----------
To compile and run our software you will need the following To compile and run our software you will need the following
prerequisites (see also SERIAL below): prerequisites (see also SERIAL below):
@ -88,21 +67,27 @@ The configure script will generate a Make.inc file suitable for building
the library. The script is capable of recognizing the needed libraries the library. The script is capable of recognizing the needed libraries
with their default names; if they are in unusual places consider adding with their default names; if they are in unusual places consider adding
the paths with `--with-libs`, or explicitly specifying the names in the paths with `--with-libs`, or explicitly specifying the names in
`--with-blas`, etc. Please note that a common way for the configure script `--with-blas`, etc.
to fail is to specify inconsistent MPI vs. plain compilers, either
directly or indirectly via environment variables; e.g. specifying the >[!CAUTION]
Intel compiler with `FC=ifort` while at the same time having an > Please note that a common way for the configure script
`MPIFC=mpif90` which points to GNU Fortran. The best way to avoid this > to fail is to specify inconsistent MPI vs. plain compilers, either
situation is (in our opinion) to use the environment modules package > directly or indirectly via environment variables; e.g. specifying the
(see http://modules.sourceforge.net/), and load the relevant > Intel compiler with `FC=ifort` while at the same time having an
variables with (e.g.) > `MPIFC=mpif90` which points to GNU Fortran.
```
module load gnu46 openmpi >[!TIP]
``` > The best way to avoid this
This will delegate to the modules setup to make sure that the version of > situation is (in our opinion) to use the environment modules package
openmpi in use is the one compiled with the gnu46 compilers. After the > (see [http://modules.sourceforge.net/](http://modules.sourceforge.net/)), and load the relevant
configure script has completed you can always tweak the Make.inc file > variables with (e.g.)
yourself. > ```
> module load gcc/13.2.0 openmpi/4.1.6
> ```
> 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. After the
> 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
``` ```
@ -114,37 +99,97 @@ install and the libraries will be installed under `/path/lib`, while the
module files will be installed under `/path/modules`. The regular and module files will be installed under `/path/modules`. The regular and
experimental C interface header files are under `/path/include`. experimental C interface header files are under `/path/include`.
SERIAL ### CUDA and GPU support
------
This version of PSBLAS incorporates into a single package three
entities that were previouslty separated:
| Library | |
|---------|--------------------|
| PSBLAS | the base library |
| PSBLAS-EXT | a library providing additional storage formats for matrices and vectors |
| SPGPU | a package of kernels for NVIDIA GPUs originally written by Davide Barbieri and Salvatore Filippone; see the license file [cuda/License-spgpu.md](cuda/License-spgpu.md) |
### OpenACC
There is a highly experimental version of an OpenACC interface,
you can access it by speficifying
```bash
--enable-openacc --with-extraopenacc="-foffload=nvptx-none=-march=sm_70"
```
where the argument to the extraopenacc option depends on the compiler
you are using (the example shown here is relevant for the GNU
compiler).
### 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 needed that enables running in pure serial mode; no MPI installation is needed
in this case (but note that the fake MPI stubs are only guaranteed to in this case (but note that the fake MPI stubs are only guaranteed to
cover what we use internally, it's not a complete replacement). cover what we use internally, it's not a complete replacement).
INTEGER SIZES ### Integers
-------------
We have two kind of integers: IPK for local indices, and LPK for We have two kind of integers: IPK for local indices, and LPK for
global indices. They can be specified independently at configure time, global indices. They can be specified independently at configure time,
e.g. e.g.
```bash
--with-ipk=4 --with-lpk=8 --with-ipk=4 --with-lpk=8
```
which is asking for 4-bytes local indices, and 8-bytes global indices which is asking for 4-bytes local indices, and 8-bytes global indices
(this is the default). (this is the default).
TODO ## Documentation
----
Fix all reamining bugs. Bugs? We dont' have any ! ;-)
Further information on installation and configuration can be found in the documentation.
See [docs/psblas-3.9.pdf](docs/psblas-3.9.pdf); an HTML version of the same document is
available in docs/html. Please consult the sample programs, especially
- [test/pargen/psb_s_pde2d.F90](test/pargen/psb_s_pde2d.F90) [test/pargen/psb_d_pde2d.F90](test/pargen/psb_d_pde2d.F90)
- [test/pargen/psb_s_pde2d.F90](test/pargen/psb_s_pde3d.F90) [test/pargen/psb_d_pde2d.F90](test/pargen/psb_d_pde3d.F90)
which contain examples for the solution of linear systems obtained by the discretization of a generic second-order differential equation in two:
```math
- a_1 \frac{\partial^2 u}{\partial x^2}
- a_2 \frac{\partial^2 u}{\partial y^2}
+ b_1 \frac{\partial u}{\partial x}
+ b_2 \frac{\partial u}{\partial y}
+ c u = f
```
or three
```math
- a_1 \frac{\partial^2 u}{\partial x^2}
- a_2 \frac{\partial^2 u}{\partial y^2}
- a_3 \frac{\partial^2 u}{\partial z^2}
+ b_1 \frac{\partial u}{\partial x}
+ b_2 \frac{\partial u}{\partial y}
+ b_3 \frac{\partial u}{\partial z}
+ c u = f
```
dimensions on the unit square/cube with Dirichlet boundary conditions.
### Utilities
The [test/util](test/util) directory contains some utilities to convert to/from
Harwell-Boeing and MatrixMarket file formats.
## TODO and bugs
The PSBLAS team. - [ ] Improving OpenACC support
--------------- - [ ] Improving OpenMP support
Project lead: - [X] Fix all reamining bugs. Bugs? We dont' have any ! 🤓
> [!NOTE]
> To report bugs 🐛 or issues ❓ please use the [GitHub issue system](https://github.com/sfilippone/psblas3/issues).
## The PSBLAS team.
**Project lead:**
Salvatore Filippone Salvatore Filippone
Contributors (roughly reverse cronological order): **Contributors** (_roughly reverse cronological order_):
Theophane Loloum Theophane Loloum
Fabio Durastante
Dimitri Walther Dimitri Walther
Andea Di Iorio Andea Di Iorio
Stefano Petrilli Stefano Petrilli
@ -161,12 +206,10 @@ Stefano Maiolatesi
Dario Pascucci Dario Pascucci
## 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 AMG4PSBLAS from interested in the package AMG4PSBLAS from
<http://github.com/sfilippone/amg4psblas> <http://github.com/sfilippone/amg4psblas> and the whole [PSCTooolkit suite](https://psctoolkit.github.io/).
Contact: <https://github.com/sfilippone/psblas3> Contact: <https://github.com/sfilippone/psblas3>

Write Preview
Loading…
Cancel
Save