C# and cudaMemcpyToSymbol

[MartyMcFly]

Hello,

I’m trying to implement my own CUDA.NET implementation in order to simplify my work with C# and CUDA at all.
I therefore created my own memory class that manages allocating and disposing the memory via the garbage collector
and I don’t have to handle the data management anymore.

In my implementations I have a few problems using constants. I have an array at the host and want to copy the data from the
host array data to the device constant data. Since there is no constant keyword I have no “symbol”. Can you show me the interface
of how to implement the cudaMemcpyToSymbol and how to access the constant in a CUDA kernel then?

Thank you for your help
Martin

[MisterAnderson42]

From the CUDA reference manual:

… symbol can either be a

variable that resides in global or constant memory space, or it can be a character string, naming a vari-

able that resides in global or constant memory space

The version that takes the symbol must be compiled with nvcc. The version that takes the character string can be called from anywhere, I think. I haven’t actually tried it.

[MartyMcFly]

Hello MisterAnderson42,

thank you very much for your inspiration. You have brought me on the right track.

Thanks

Martin

From the CUDA reference manual:

The version that takes the symbol must be compiled with nvcc. The version that takes the character string can be called from anywhere, I think. I haven’t actually tried it.