Vectorisation (rewriting the code to avoid loops) in Python can make the code 10-100x faster. The profiler identifies get_lcoe as the most costly function for ftt-p, and it should not be too difficult to vectorise.
I think if this one comes down from 20 s to <1 s, this should be the moment we can switch from no_it to 20 without it impeding development too much.
Vectorisation (rewriting the code to avoid loops) in Python can make the code 10-100x faster. The profiler identifies get_lcoe as the most costly function for ftt-p, and it should not be too difficult to vectorise.
I think if this one comes down from 20 s to <1 s, this should be the moment we can switch from no_it to 20 without it impeding development too much.