author: Alexandru Patriciu
author email: apatriciu@gmail.com

Parallel Simulator for LTI systems. Can be easily extended to LTV systems. 
Includes SciLab and Python interfaces. The library was tested on Ubuntu 10.04
with ATI and NVidia hardware.

The library implements Runge-Kutta 4th order and multi step Adams - Bashforth - Moulton 
methods.

The parallelization approach uses a scan along time and parallel matrix vector 
multiplications.

In order to build the library project you need cmake.

Build dependency
C++ boost library - the user has to have C++ boost library installed in order to 
compile the library. The user has to provide the path to the boost library in 
the cmakelists.txt file.
OpenCL - the system has to have an OpenCL library system installed as well as 
an available GPU. The library detects only the OpenCL platforms that have GPUs.

Python interface
The python extension has to be compiled from source. Go to folder 
/Cpp/PythonInterface. Edit and update the folders in ParODESetup.py to
match your system.

>>>python ParODESetup.py build

This will build a shared library ParODE.so in the folder ./build/linux... etc.
You may then install or copy the shared library in your working folder.

Scilab Interface
See the user manual.