I want to valgrind to verify that I have proper memory management in the software that I am working on that uses CUDA 6.0 and OptiX 3.6.2. I am currently developing on Ubuntu 12.04. Unfortunately, valgrind is indicating an unusual large amount of memory leaks from the OptiX API. Is this a known problem, i.e. is OptiX doing some sort of memory management that confuses valgrind? I just need to know if it’s safe for me to suppress all the errors or not.