clEnqueueNDRangeKernel is returning an invalid error code with a value of -1000. Calling clEnqueueNDRangeKernel on AMD and Intel hardware/drivers works just fine. I only get this error code of -1000 on NVIDIA. I’m using a Quadro K5200 with driver R347.52. Does anybody know what -1000 is suppose to stand for or what could cause it?
Funny thing is… this is not a CUDA API but an OpenCL API ! :)
Funny thing is…you must not realize NVIDIA does not have a section dedicated exclusively to the OpenCL API. An additional funny this is…you must not realize that NVIDIA’s OpenCL 1.1 implementation uses CUDA. So many funny things…
-1000
CL_INVALID_GL_SHAREGROUP_REFERENCE_KHR
clGetGLContextInfoKHR, clCreateContext
CL and GL not on the same device (only when using a GPU).
[url]http://streamcomputing.eu/blog/2013-04-28/opencl-1-2-error-codes/[/url]
a lot on google pertaining to opencl error code -1000
Funny thing is little_jimmy found documentation for wrong API’s… Even more funny is I cannot find information for -1000 and this particular API as well… even more funny thing is Carnaval now happing and I must go eat… and even more funny… why use a clone-api if you can use the real thing ! ;) =D At least CUDA apperently better documented ! Or funny thing is… your error return code is somehow wrong/invalid ? :) :P* :)