SOPT
Sparse OPTimisation
|
SOPT is an open-source C++
package available under the license below. It performs Sparse OPTimisation using state-of-the-art convex optimisation algorithms. It solves a variety of sparse regularisation problems, including the Sparsity Averaging Reweighted Analysis (SARA) algorithm.
SOPT also has several MPI wrappers that can be adapted for computational distirbution of various linear operators and convex optimisation algorithms. Wavelet Operators with SOPT also support multi-threading through OpenMP.
SOPT is written in C++
primarily but also contains partial and prototyped Matlab implementations of various algorithms.
SOPT is largely provided to support the PURIFY package, a companion open-source code to perform radio interferometric imaging, also written by the authors of SOPT. For further background please see the reference section.
This documentation outlines the necessary and optional dependencies upon which SOPT should be built, before describing installation and testing details and Matlab support. Contributors, references and license information then follows.
SOPT is mostly written in C++17
. Pre-requisites and dependencies are listed in following and minimal versions required are tested against Travis CI
meaning that they come natively with OSX and the Ubuntu Trusty release. These are also the default ones fetched by CMake
.
C++
minimal dependencies:
C++
.C++
linear algebra. Downloaded automatically if absent.C++
unit-testing framework only needed for testing. Downloaded automatically if absent.C++
micro-benchmarking framework only needed for benchmarks. Downloaded automatically if absent.Conan is a C++ package manager that helps deal with most of the C++ dependencies as well as the SOPT installation:
cppflow
package:Clone the UCL fork of cppflow and create a conan package using
``` bash git clone git@g:UCL/cppflow.git conan create ./cppflow/ ``` Note that conan requires you to specify the host (h) and the build (b) profiles on the command line ( ithu b.com-pr:h=default -pr:b=default
), unless you have defined them in your conan profile. You can set up a default profile for your system using conan profile detect
(only needs to be done once).
Once the mandatory dependencies are present, git clone
from the GitHub repository:
``` bash git clone https://github.com/astro-informatics/sopt.git ```
Then, the program can be built using conan:
``` bash cd /path/to/code mkdir build cd build conan install .. -of . –build missing conan build .. -of . ```
Things to note:
To install in directory INSTALL_FOLDER
, add the following options to the conan build command:
``` bash conan build .. -of INSTALL_FOLDER ```
conan install
using the -o
flag with a value on
or off
. Possible options are: - tests (default on) - benchmarks (default off) - examples (default on) - logging (default on) - openmp (default on) - mpi (default on) - docs (default off) - coverage (default off) - cppflow (default off)For example, to build with both MPI and OpenMP off you would use
If the dependencies are already available on your system, you can also install SOPT manually like so
On MacOS, you can also install most of the dependencies with Homebrew e.g.
Note that the ONNXruntime interface is currently only supported when compiling with Clang on MacOS, but not with g++
If you are using the g++ compiler and get an error to do with the package spdlog
, try adding the option -s compiler.libcxx=libstdc++11
to the conan build
command. This option is also necessary when building with gcc on MacOS.
You can set commonly used options, choices of compilers etc. in a conan profile. You can list profiles available on your system using conan profile list
and select the profile you want to use with conan install
with conan install .. -pr my_profile
. CMake build options can also be added to the profile under [options]
. Here is an example of a conan profile for building with a homebrew installed gcc 11 on MacOS.
To check everything went all right, run the test suite:
A separate Matlab implementation is provided with SOPT. This implementation includes some (but not all) of the optimisation algorithms implemented in the C++
code, including the SARA algorithm.
The Matlab implementation is contained in the matlab directory. This is a stand-alone implementation and does not call any of the C++
code. In future, Matlab interfaces to the C++
code may also be included in SOPT.
See matlab/README.txt
for an overview of the Matlab implementation. The stand-alone Matlab implementation is also self-documenting; corresponding documentation can be found in matlab/doc
. We thank Gilles Puy for contributing to this Matlab implementation.
Check the contributors page (github).
If you use SOPT for work that results in publication, please reference the webpage and our related academic papers:
SOPT: Sparse OPTimisation package Copyright (C) 2013-2023
This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2 of the License, or (at your option) any later version.
This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details (LICENSE.txt).
You should have received a copy of the GNU General Public License along with this program; if not, write to the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
For any questions or comments, feel free to contact Jason McEwen, or add an issue to the issue tracker.
The code is given for educational purpose. For the Matlab
version of the code see the folder matlab.