lax.dynamic_slice with (constant) stride

[PhilipVinc]

Is there a way to implement some sort of dynamic_slice with a stride? apparently it's possible to do that only with non -dynamic start indices...

[hawkinsp]

My suggestion is to compute slice(dynamic-slice(...), ..., stride). i.e., take a second strided slice on the output of the dynamic-slice. It's quite possible the compiler will fuse the two.