I’ve noticed that the OpenCL 1.1 spec has an upper bound on work_dim of CL_DEVICE_MAX_WORK_ITEM_DIMENSIONS rather than 3 (where CL_DEVICE_MAX_WORK_ITEM_DIMENSIONS >= 3). Will this be an implementation-dependent ceiling? What sort of lead time might there be before 1.1 implementations arrive that will allow a >3 ceiling? (I’m looking to start writing some 4D code and am wondering whether it would make sense to wait, or just implement it as a 1D problem)
“4D code” - what a times… 8)
I doubt dim > 3 will be there soon. Look at CUDA - 3 years on the market and still dim <= 3.