Honored Contributor
Honored Contributor@nadram and forum admins: Have you received any example of multiple compute units? I am interested the most in how to allocate input data in global memory for individual compute units. I did some basic kernel simulations in Modelsim and it looks like all compute units try to read the same base address on global DDR bus. This would imply that they can only work on the same input data. Of course my simulation setup can be wrong. It would be good to have such example because there are quite a few questions regarding use of num_compute_units seem have be not been answered here for a long time.
