Emulation mode in CUDA 3.2 with VS2008

I am trying to debug into my kernel code, using the device emulation mode.

However, I set break points in my k开发者_StackOverflowernel and it doesn't break.

 MatrixMultiplication_Kernel<<<dimGrid, dimBlock>>>(Md, Nd, Pd, Width);

Can anyone assist me with this?

Device emulation was removed in the 3.1 release. You can debug on Windows using Parallel Nsight, or on Linux/Mac using cuda-gdb (included with the CUDA Toolkit).

I didn't look into it or check it, but maybe the following software is an alternative or could be supportive for debugging purposes regarding cuda:

http://code.google.com/p/gpuocelot/

The most actual paper on the front-site is titled "Ocelot: An Open Source Debugging and Compilation Framework for CUDA". So maybe this software will help you. It seems to be under heavy development.