Will I get a significant performance hit if I call glGetError() per frame, even if just before the SwapBuffers call (and I'm using vsync)?
For example, if I have a persistently mapped buffer, and I've initiated a transfer, then does the glGetError() act like a barrier and would thus potentially be a significant hit?
Not really, as its something the driver just sets once and forgets. In a release build you should check everywhere during init and then at opportune places in the renderer (then disable it after first error).
In debug builds just check everywhere you think it will help, such as after rendering particles etc. as it will help you get started narrowing down bugs.
Calling glGetError() will give some performance hit. Most GL drivers nowadays are threaded - when you call glDoTheThing(), the call and its arguments are pushed to a queue which is executed by another thread. By calling glGetError(), you force a sync point in the driver so it has to wait for the GL command to finish before the error can be read.
You can make use of the debug output functions which can report errors asynchronously, which shouldn't give as much of a performance hit.
Or simply have a debug and a release version of your application.
This topic has been locked by a moderator. New replies are not allowed.
With your permission, GameDev.net uses analytics cookies to understand how people use the platform. You can accept analytics or continue with necessary cookies only. Learn more
