Is there any opencv-ocl capable dev board on the shelf?

edit

hello every one, I would like to run opencv application with opencl speed-up on arm embedded deg-board. I have tried odroid-xu but the work-group size does not match what opencv requires(larger than 256) and it always assert fail when comparing the sum of local thread and the max work group size provided by GPU.

One thing I can start with is to change the local thread parameter in the function I used and before that I would like to know and leverage the experience of opencv with ocl module on arm dev-board.

Is there any one successfully done what I would like to do before or any suggestion for me?

thank you