Why not to use cutil library functions

Hi all,

Some time ago, I remember coming on the forum and being taught by others not to use the cutil functions in my code (e.g. http://forums.nvidia.com/index.php?showtopic=163224). However, I can’t seem to “re-find” the reasons why one wouldn’t want to use the cutil functions in your code. In looking at the README for the cutil functions, it mentions the following:

“These macros are compiled out in release builds and so they will not affect performance. Note that in debug mode they call cudaThreadSynchronize() to ensure that kernel execution has completed, which can affect performance.”

Is that simply it? Or are there other/better/different reasons not to use the cutil functions in your code?

Thanks,

Matt

Basically, cutil is not considered by NVIDIA to be a “public” API. It exists to remove some of the boilerplate code from the SDK examples to improve readability. The functions are not documented, and not guaranteed to be stable between releases.

Basically, cutil is not considered by NVIDIA to be a “public” API. It exists to remove some of the boilerplate code from the SDK examples to improve readability. The functions are not documented, and not guaranteed to be stable between releases.

Sounds good, thanks!

Matt

Sounds good, thanks!

Matt