Hello,
I have strange problem and can´t imagine that I´m the only one having that problem. The NVidia Sdk don´t report an error on clEnqueueNDRangeKernel when I forget to set a kernel parameter. So if have a kernel like this
__kernel void test(__global struct* in_test, __global float* in_test2, float in_test3)
{
}
and only set the first two parameters with clSetKernelArg I DO NOT get an error. The function by the way don´t work then. Is it a known bug?!??! It´s working with the Intel SDK.
Ok… could someone please confirm or stating the opposite?!
Yeah, that also happened to me. It doesn’t cause any error, but it just return the wrong output. I’m using OpenCL 1.0 in CUDA 3.2.