If n is constant, it should be a wiring operation and it shouldn't require multiple itterations so it should be allowed in a SCTL.
Use the Option link next to your message and Report to Moderator to request that it be moved.
I'd also suggest that you add some more detail to your post. Perhaps even an image as to what you are talking about. It would go a long way in clarifying what you are talking about and in selling your idea to entice people to kudo it.
That's good feedback RavensFan. To clarify, currently, if you use rotate 1D array in a single cycle timed loop, you get an error like this:
I don't think there's any fundamental reason why that shouldn't be allowed. It was probably disallowed incase someone wired a control into the n terminal at the top.
It is probably because rotating an array is computationally expensive, and might not be guarenteed to execute in the given time.
Mythilt, in an FPGA (which is what this idea is actually for) a rotate with a constant for the number of elements to rotate should be nothing more than a routing issue. No computations needed. Admittedly, it probably is because of the non-constant rotations that it is not allowed inside of a SCTL.
Interesting, I did not know that FPGAs handled arrays that much differently. Have to do some reading on that. Makes sense due to the limitations of their embedded nature.