- Code: Select all
std::unique_lock<std::mutex> lock(lock_);
sharedTexture.reset();
backTexture.reset();
frontTexture.reset();
sharedTexture = std::make_shared<CTexture2D>(buffer, width, height, (int)gDepth, device_, context_);
backTexture = std::make_shared<CTexture2D>(width, height, (int)gDepth, device_, context_);
frontTexture = std::make_shared<CTexture2D>(width, height, (int)gDepth, device_, context_);
and when render swap texture
When I run this code, it can run on 100FPS. If I changed the texture size to 1920*1080, it just can be run on 45FPS.
My graphic card is GTX 1050Ti. I don't think it can run a very simple scene that just has one quad using to be render html page and a spin cube.
Is the performance overhead of OSR mode high? Or it is caused by thread call?