I have a complex Visual Studio 2015 project, with a lot of C++ host code, and many .cl files.
So far, I have been developing "by instinct", using theoretical understanding of GPU architecture
to optimize and design the code. This approach has worked quite well so far, but it is time consuming.
I am ready to take things to the next level, by looking at low level machine code, occupancy, etc.
Can anyone recommend a good approach for me to start using CodeXL on my project?