This project is read-only.
2
Vote

Add cl_float to avoid alignment problems

description

I have been writing a GPU application to do fast matrix multiplications that can be run from C# using your library. I have run into the problem that the conversion from C# floats to OpenCL floats does not work 100% of the time. I believe this may be solved by including a cl_float type in the API that is equivalent to the scalar data type which is currently not included.

Thanks.

comments