running without managedCuda

[fire-emblem]

Hi~

You mentioned that this project is for C#'s managedCuda. Does it support calling CUDA code directly?

[kunzmi]

Hi,

actually no, mpp is not for managedCuda, it is more that managedCuda inspired me to start the new project. At the end (and if all goes well), mpp should be a replacement of Nvidia's npp library: with a C-API same as NPP and probably the C++/classed based API as it is right now.
Not sure what you mean by "calling CUDA code directly": you mean that mpp calls CUDA code or that your code together with mpp interacts with the CUDA api?

[fire-emblem]

Great, what I am looking for is exactly an open-source project that offers the same functionality and usage as NPP, uses a C-API style, and is developed based on CUDA code.

According to your description, the mpp project you developed is exactly what I want.
Thank you for your work.
I am now trying to compile and use it.