After changing the thread count in the graphical interface, it seems to have no effect on rendering time (and without modifying the thread code, the issue mentioned in the documentation does not occur). If the constructor parameters in the render_engine file are changed, without modifying the thread functions, I tried increasing the number of rasterization threads. Although the rendering time changes, no matter how many are added, the change in rendering time is the same. For example, rendering a short‑duration model like cow only reduces the time by half regardless of how many threads are added, while rendering a long‑duration model like dragon can only be reduced by about 2 seconds (which puzzles me). So how can the thread count be increased? I haven’t seen any documentation about increasing the thread count, and the documentation only mentions that the main parts that need to be modified are the three thread functions, fragmentbuff, and context.
Does anyone know the solution?
// choose render type
void RenderEngine::render(Scene& scene, RendererType type)
{
switch (type) {
case RendererType::RASTERIZER: {
rasterizer_render = std::make_unique(*this, n_threads, n_threads, n_threads);
rasterizer_render->render(scene); break;}
// case RendererType::RASTERIZER_MT: rasterizer_render->render_mt(scene); break;
case RendererType::WHITTED_STYLE: whitted_render->render(scene); break;
default: break;
}
}
Try changing the original case statements in the RenderEngine::render function of render_engine.cpp to the code above.
An extra line was added: rasterizer_render = std::make_unique<RasterizerRenderer>(*this, n_threads, n_threads, n_threads);