Believe me, or not but I got something really special :- Our game is running FASTER when started from intel GPA !!A simple sample is the time needed for a specific function, in average its 5 time faster when running trhough GPA analyser. I did some test, and this time is not taken from the DX Draw call, but from "pure C#" code !
It's like if GPA was making the code running faster ... really strange.
Do you have an idea, is GPA monitor starting the application with special parameters ?