Hello,
i compiled fortrancl with ifort and the msvs c compiler. It works quite well, but after i tried some examples i noticed that i cannot use the method clSetKernelArgLocal which makes the implementation of a efficient matrix vector multiplication next to impossible.
I checked the source, and the interface is included in cl_kernel.f90 but not in the final library ( i checked using nm ).
I am really dependent on this functionality and new to fortran and opencl in general. Any help would be greatly appreciated.
-Sascha Walther