marwenkr / nasoq

NASOQ:Numerically Accurate Sparsity Oriented QP Solver

NASOQ: Numerically Accurate Sparsity Oriented QP Solver

NASOQ is a scalable and efficient Quadratic Programming solver that obtains solutions for requested accuracies. Visit our website for more details: NASOQ Website

Installation

Library requirements

MKL Pardiso or OpenBlas (BLAS), OpenMP and METIS. Cmake handles METIS. If you install OpenBlas in its default location (sudo make install), Cmake will detect it.

Building the project

Given that MKL Pardiso or OpenBlas are installed, install NASOQ using following steps:

mkdir build
cd build
cmake -DMKL_ROOT_PATH=path/to/intel  -DCMAKE_BUILD_TYPE=Release ..
make

A quick script for building and running NASOQ is provided in buildALL.sh. You need to first correct paths to libraries and then you can run it as following:

bash buildAll.sh

Upon successful build you should be able to see data/out.csv and it should be similar to data/out_correct.csv.

For installing on MAc you might need to use GCC so you need to also set the CMAKE compiler flag.

More details are provided in: https://nasoq.github.io/docs/getting-started-nasoq/

Using NASOQ as a Library

More details: https://nasoq.github.io/docs/getting-started-nasoq/

Testing a QP example

To test a QP example you may also use NASOQ-BIN which is a command line interfce for NASOQ. Some small QP problems are available in data folder. For evaluating NASOQ versus other solvers a separate repository is also provided in: https://github.com/sympiler/nasoq-benchmarks More details: https://nasoq.github.io/docs/repository/

Copyright 2020 Kazem Cheshmi