Cs-gO

Cs-gO is a C++ Language library that allows Fast, portable GPU programming backed by OpenGL Compute Shaders.

Installation

Linux

Linux Makefile is provided with the library
git clone https://github.com/daviddhas/CS-gO.git
A header-only library But you must link against OpenGL 4.3+
Navigate to the library directory using cd

Incase you do not have glfw or glm setup, Navigate to the vendors directory and

    git submodule update --init

Windows

Visual Studio Project Available on GitHub!
git clone https://github.com/daviddhas/CS-gO.git
A header-only library But you must link against OpenGL 4.3+
Add the project as includes

Integrating Library:

Incase you have OpenGL already installed,

Copy the includes folder in the library directory to your project working directory
Link the Cs-gO header file

If you don’t have OpenGL already installed,

Copy the includes folder, the vendor folder and src folder in the library directory to your project working directory
Link the Cs-gO header file

Examples

The various library functions, types and programs are illustrated through examples below

Average

The average function contains the CS Go DSL (domainspecific language)

    auto average(csgo::dsl::image2d<glm::vec4> in1, csgo::dsl::image2d<glm::vec4> in2)
    {
	    using namespace csgo::dsl;
	    image2d<glm::vec4> x = ( in1 + in2 ) / 2;
	    return x;
    }

The image2d type doesn’t actually contain data, just supports GLSL operations.

The image2d_io type contains an OpenGL texture, but doesn’t support GLSL operations Separates the dsl from C++. It also enables different copy constructors

    void run_average() {

    GLuint size = 32;
    csgo::program p(average, { {size, size} }, true);
    csgo::image2d_io<glm::vec4> in1(std::vector<glm::vec4>(size * size, glm::vec4(1)), size);
    csgo::image2d_io<glm::vec4> in2(std::vector<glm::vec4>(size * size, glm::vec4(1, 0, 0, 1)), size);

    std::tuple<csgo::image2d_io<glm::vec4>> results = p(in1, in2);

    auto result = std::get<0>(results).read();
    std::cout << result.size() << " " << size << std::endl;

    for (glm::vec4 v : result)
        std::cout << v.x << ", " << v.y << ", " << v.z << ", " << v.w << std::endl;

    csgo::display::image(std::get<0>(results));
    while (true);
}

The run_average function specifies specific inputs and outputs

Program

A program is constructed from a function that takes image2d’s and returns a tuple of image2d’s This contains the size of output textures. The program also determines whether or not we need to create an OpenGL context

    GLuint size = 32;
    csgo::program p(average, { {size, size} }, true);

A program is ran by passing image2d_io’s for each image2d. It returns a tuple of image2d_io’s that can be read from. The program throws a runtime error if the input or the results tuple type are the wrong sizes

    csgo::image2d_io<glm::vec4> in1(std::vector<glm::vec4>(size * size, glm::vec4(1)), size);
    csgo::image2d_io<glm::vec4> in2(std::vector<glm::vec4>(size * size, glm::vec4(1, 0, 0, 1)), size);

    std::tuple<csgo::image2d_io<glm::vec4>> results = p(in1, in2);
    csgo::display::image(std::get<0>(results));

Running the code:

In case of Linux, the code can be compiled by using the linux make file provided in the library project directory
In case of Windows, the code can be compiled by using pressing Ctrl+F7 in the visual studio window

daviddhas/CS-gO