void Renderer::render_disparity(const uint16_t* d_disp, int disp_size) {
glActiveTexture(GL_TEXTURE0);
glBindTexture(GL_TEXTURE_2D, 0);
// cuda-gl interop
cudaGraphicsResource_t cuda_gl_tex_resource;
cudaGraphicsGLRegisterImage(&cuda_gl_tex_resource, disp_texture_, GL_TEXTURE_2D, cudaGraphicsRegisterFlagsSurfaceLoadStore);
cudaGraphicsMapResources(1, &cuda_gl_tex_resource);
cudaArray_t texture_array;
cudaGraphicsSubResourceGetMappedArray(&texture_array, cuda_gl_tex_resource, 0, 0);
cudaResourceDesc desc;
desc.resType = cudaResourceTypeArray;
desc.res.array.array = texture_array;
cudaSurfaceObject_t write_surface;
cudaCreateSurfaceObject(&write_surface, &desc);
write_surface_U16_with_multiplication(write_surface, d_disp, width_, height_, 256);
cudaDestroySurfaceObject(write_surface);
cudaGraphicsUnmapResources(1, &cuda_gl_tex_resource);
cudaGraphicsUnregisterResource(cuda_gl_tex_resource);
// end cuda-gl interop
glUseProgram(program_disp_);
glBindTexture(GL_TEXTURE_2D, disp_texture_);
glEnableVertexAttribArray(0);
glEnableVertexAttribArray(1);
glBindBuffer(GL_ARRAY_BUFFER, vert_buffer_);
glVertexAttribPointer(0, 2, GL_FLOAT, GL_FALSE, sizeof(float) * 5, (void*)0);
glVertexAttribPointer(1, 2, GL_FLOAT, GL_FALSE, sizeof(float) * 5, (float*)0 + 3);
GLint loc;
loc = glGetUniformLocation(program_disp_, "tex_sampler");
if (loc != -1) {
glUniform1i(loc, 0);
}
loc = glGetUniformLocation(program_cdisp_, "inv_disp_size");
if (loc != -1) {
glUniform1i(loc, 256 / disp_size);
}
glDrawArrays(GL_TRIANGLE_STRIP, 0, 4);
glBindTexture(GL_TEXTURE_2D, 0);
}
CudaSurfaceObject(cudaResourceDesc desc)
{
CUDA_CALL(cudaCreateSurfaceObject(&m_surfaceObject, &desc));
}
int main() {
//Checks for memory leaks in debug mode
_CrtSetDbgFlag(_CRTDBG_ALLOC_MEM_DF | _CRTDBG_LEAK_CHECK_DF);
glfwInit();
glfwWindowHint(GLFW_CONTEXT_VERSION_MAJOR, 4);
glfwWindowHint(GLFW_CONTEXT_VERSION_MINOR, 4);
glfwWindowHint(GLFW_OPENGL_PROFILE, GLFW_OPENGL_CORE_PROFILE);
glfwWindowHint(GLFW_RESIZABLE, GL_FALSE);
GLFWwindow* window = glfwCreateWindow(width, height, "Hikari", nullptr, nullptr);
glfwMakeContextCurrent(window);
//Set callbacks for keyboard and mouse
glfwSetInputMode(window, GLFW_CURSOR, GLFW_CURSOR_DISABLED);
glewExperimental = GL_TRUE;
glewInit();
glGetError();
//Define the viewport dimensions
glViewport(0, 0, width, height);
//Initialize cuda->opengl context
cudaCheck(cudaGLSetGLDevice(0));
cudaGraphicsResource *resource;
//Create a texture to store ray tracing result
GLuint tex;
glActiveTexture(GL_TEXTURE0);
glGenTextures(1, &tex);
glBindTexture(GL_TEXTURE_2D, tex);
glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_S, GL_CLAMP_TO_EDGE);
glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_WRAP_T, GL_CLAMP_TO_EDGE);
glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MAG_FILTER, GL_NEAREST);
glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, GL_NEAREST);
glTexImage2D(GL_TEXTURE_2D, 0, GL_RGBA32F, width, height, 0, GL_RGBA, GL_FLOAT, NULL);
cudaCheck(cudaGraphicsGLRegisterImage(&resource, tex, GL_TEXTURE_2D, cudaGraphicsMapFlagsWriteDiscard));
glBindTexture(GL_TEXTURE_2D, 0);
Shader final = Shader("fsQuad.vert", "fsQuad.frag");
FullscreenQuad fsQuad = FullscreenQuad();
float4* buffer;
cudaCheck(cudaMalloc((void**)&buffer, width * height * sizeof(float4)));
cudaCheck(cudaMemset(buffer, 0, width * height * sizeof(float4)));
//Mesh
float3 offset = make_float3(0);
float3 scale = make_float3(15);
Mesh cBox("objs/Avent", 0, scale, offset);
offset = make_float3(0, 55, 0);
scale = make_float3(100);
Mesh light("objs/plane", (int)cBox.triangles.size(), scale, offset);
cBox.triangles.insert(cBox.triangles.end(), light.triangles.begin(), light.triangles.end());
cBox.aabbs.insert(cBox.aabbs.end(), light.aabbs.begin(), light.aabbs.end());
std::cout << "Num triangles: " << cBox.triangles.size() << std::endl;
cBox.root = AABB(fminf(cBox.root.minBounds, light.root.minBounds), fmaxf(cBox.root.maxBounds, light.root.maxBounds));
BVH bvh(cBox.aabbs, cBox.triangles, cBox.root);
Camera cam(make_float3(14, 15, 80), make_int2(width, height), 45.0f, 0.04f, 80.0f);
Camera* dCam;
cudaCheck(cudaMalloc((void**)&dCam, sizeof(Camera)));
cudaCheck(cudaMemcpy(dCam, &cam, sizeof(Camera), cudaMemcpyHostToDevice));
cudaCheck(cudaGraphicsMapResources(1, &resource, 0));
cudaArray* pixels;
cudaCheck(cudaGraphicsSubResourceGetMappedArray(&pixels, resource, 0, 0));
cudaResourceDesc viewCudaArrayResourceDesc;
viewCudaArrayResourceDesc.resType = cudaResourceTypeArray;
viewCudaArrayResourceDesc.res.array.array = pixels;
cudaSurfaceObject_t viewCudaSurfaceObject;
cudaCheck(cudaCreateSurfaceObject(&viewCudaSurfaceObject, &viewCudaArrayResourceDesc));
cudaCheck(cudaGraphicsUnmapResources(1, &resource, 0));
while (!glfwWindowShouldClose(window)) {
float currentFrame = float(glfwGetTime());
deltaTime = currentFrame - lastFrame;
lastFrame = currentFrame;
//Check and call events
glfwPollEvents();
handleInput(window, cam);
if (cam.moved) {
frameNumber = 0;
cudaCheck(cudaMemset(buffer, 0, width * height * sizeof(float4)));
}
cam.rebuildCamera();
cudaCheck(cudaMemcpy(dCam, &cam, sizeof(Camera), cudaMemcpyHostToDevice));
frameNumber++;
if (frameNumber < 20000) {
cudaCheck(cudaGraphicsMapResources(1, &resource, 0));
std::chrono::time_point<std::chrono::system_clock> start, end;
start = std::chrono::system_clock::now();
render(cam, dCam, viewCudaSurfaceObject, buffer, bvh.dTriangles, bvh.dNodes, frameNumber, cam.moved);
end = std::chrono::system_clock::now();
std::chrono::duration<double> elapsed = end - start;
std::cout << "Frame: " << frameNumber << " --- Elapsed time: " << elapsed.count() << "s\n";
cudaCheck(cudaGraphicsUnmapResources(1, &resource, 0));
}
cam.moved = false;
glUseProgram(final.program);
glActiveTexture(GL_TEXTURE0);
glBindTexture(GL_TEXTURE_2D, tex);
glClear(GL_COLOR_BUFFER_BIT);
final.setUniformi("tRender", 0);
fsQuad.render();
//std::cout << glGetError() << std::endl;
//Swap the buffers
glfwSwapBuffers(window);
glfwSetCursorPos(window, lastX, lastY);
}
