I have a large code base running OpenCl on AMD GPUs and I am trying to run it on nVidia but I get an error of -1000 for the call to clEnqueWriteImage. All other functions related to opening of the device, context, and command queue all succeed (I can queue the device, version etc…). The OpenCl context is interop’d with OpenGl and the cl_mem item in use is built from an OpenGl texture. All of this works on AMD GPUs just not with the GTX 980Ti (only one tested so far) I have here. I am not even sure what the error code means as -1000 is not an OpenCl error. Early on in development I remember seeing -1001 as well but cannot remember what it meant (or how we got around it).
I can skip pass the upload error but them my call to enqueue the kernel fails with the same error code -1000.
Call is setup to similar to this:
unsigned char *buffer;
…
cl_int status;
cl::size_t<3> origin;
cl::size_t<3> region;
origin[0]=0;
origin[1]=0;
origin[2]=0;
region[0]=m_width;
region[1]=m_height;
region[2]=1;
status=clEnqueueWriteImage(m_commandQueue, m_image, CL_FALSE, origin, region, 0, 0, buffer, 0, NULL, NULL);