INSTALL

We use CMake to build the library and compile a standalone test application
Check CMake for library dependencies (OpenCL, OpenGL, GLEW)

We encourage an out-of-source build, so in the base directory of the repo do:
    mkdir build
    cd build
    cmake ..
    make
    make install

You can run by executing
    ./bin/sph.x

(currently you set the path for resource files in main.cpp, this really should automatically get it using argv[0]
so for now you'll need to run from build so that relative paths are eg ./bin/cl_common)

the library will be in rtpslib/

if you have Doxygen you can generate docs:
    make doc

the docs will be in rtpslib/html



LINUX NOTES



MAC OS X NOTES



WIN32 NOTES

* from Andrew Young
Install CUDA toolkit for windows. http://www.nvidia.com/object/thankyou.html?url=/compute/cuda/3_2_prod/toolkit/cudatoolkit_3.2.16_win_64.msi
Set GPU_COMPUTING_TOOLKIT environment path to CUDA directory 
Download and install Nvidia\u2019s code samples (this comes prepackaged with glut glew and other utility libraries) http://developer.download.nvidia.com/compute/cuda/3_2_prod/sdk/gpucomputingsdk_3.2.16_win_64.exe
Download and install visual studio 10 express edition.
http://go.microsoft.com/?linkid=9709949

    create a build directory
    run cmake
        point source directory to rtps
        point build directory to your build directory
        press configure
        press generate
    
    navigate to build directory
    double-click on rtpssuite.sln
    when visual studio opens right-click on test.x in the left panel
    select build
    navigate to build directory copy rtpslib\Debug\rtps.dll to test\Debug
    double-click test\Debug\test.exe

----------------------------------------------------------------------
Possible ERRORS
- Make sure that rtpslib/cl_sph/ does not contain directories. That prevents cmake .. from running 
correctly. When in doubt, remove everything in build/ and start over.