Kernel Call Toolkit 4.0

Hi there,

I recently received a Tesla 1060 in addition to my GTX470 therefore I updated to Toolkit 4.0 to use Nsight properly. Now I have a couple of problems, first

I can start the debugging and run through my code but VS 2008 doesn’t allow me to set breakpoints , it always says it is not allowed at this position.

And my second problem is

I tried to compile one of my old projects (using the build rules from the 3.2 sdk) with the new API runtime from 4.0. I managed to compile it (.cu files excluded from building) but it looks like the program is not entering the kernel anymore.
So the question is are there any major changes in the kernel call or in a malloc function that could cause this ? When I switch the building rules back to v3 it works again.

One more thing a specified the tesla card as my computing device by using cudaSetDevice() .

As I am fairly new to CUDA and just thought I am on the right way I would be very grateful for any advice

cheers Markus