Error CL_INVALID_WORK_GROUP_SIZE when calling clEnqueueNDRangeKernel for kernel 0.

[Fuzzyma]

I get the above error when specifying a worksize > 2. So basically only 1 and 2 are working.

I have an Intel HD Graphics 4600 which I use to mine. I am on win7 and compiled the current dev branch.
Can I somehow circumvent the CL_INVALID_WORK_GROUP_SIZE error so I can increase the worksize?

//EDIT: It seems like the xmr-stak-amd is using my nvidea card instead of the intel HD. What to do against taht?

[psychocrypt]

If the nvidia card is used please set the correct index or platform. Run clinfo to found the right indices.

[Fuzzyma]

Where can I find clinfo? I tried to change the index to 1 - actually first thing i tried. No device found in this case:

"gpu_threads_conf" : [
{ "index" : 1, "intensity" : 100, "worksize" : 2, "affine_to_cpu" : false },
],

[2017-09-13 09:52:52] : Selected OpenCL device index 1 doesn't exist

When changing the platform index to 1 I get

[2017-09-13 10:00:58] : Compiling code and initializing GPUs. This will take a while...
[2017-09-13 10:00:59] : Device 0 work size 2 / 512.
[2017-09-13 10:00:59] : Error CL_INVALID_DEVICE when calling clCreateCommandQueueWithProperties.
Press any key to exit.