I’m curious as to what is the best way and what happens when nesting reduce_sum calls.
I have N time series that each have M_n entries and I want to call reduce_sum over N and then inside each time series as the computations are very long.
- Is this smart?
- what actually happens?
- is there some special way to set up the grain size to control this?
Currently I have just been calling reduce_sum manually on each of the M time series, but I have read in other posts that this is not the best way to go.