* ABOUT
    This project is a part of the seminar work on potential of GPUs at 
    performing image processing related algorithms. Created by
    
    Srđan Rašić
    srdan.rasic@fer.hr
    Faculty of Electrical Engineering and Computing,
    University of Zagreb, Croatia
    
    April 2012
    

* REQUIREMENTS

    + Unix platfrom
    + OpenCL compatible GPU (or CPU)
    
    + gcc (>=4.1)
    + libpng (>=1.2.x) (apt-get install libpng12-dev)
    + png++ (0.2.1, but other version should work well too) (apt-get install libpng++-dev)
    + OpenCL SDK (>= 1.0, tested with v1.1)
        - to get OpenCL dev files on
            Linux:  Install proprietary drivers for NVIDIA/AMD GPUs and 'opencl-headers'
                    or install Intel OpenCL SDK for Intel GPUs and CPUs
            Mac OSX: Install XCode
            Windows: Should be same as for Linux, but to compile this code you
                     should deal with the sys/time.h and similar staff first :)


* COMPILATION AND EXECUTION
        
    1. Change to directory in which this README is placed
    2. Type 'make'
    3. Type 'make run'
    
    
* NOTES
    + Executable is placed in directory './bin' and is named 'seminar'
    + Executable MUST be executed from directory '.', or
	in other words, from directory in which this README is placed
    + To find out executable arguments, execute it without any
    
    
* NOTES ON SOURCE CODE
    + OpenCL framework subset in wrapped into a simple C++ wrapper called 'oclw'
	You can use this wrapper to do basic staff on OpenCL compatible device.
	
	To get more info about the wrapper, consult:
	./docs/html/index.html or