Created
January 8, 2019 10:29
-
-
Save min2bro/f710819b4248d83e9b625c4f4166f1d0 to your computer and use it in GitHub Desktop.
Coldsp33d
commented
Jan 8, 2019
- Start by defining a problem and outline these methods in context of the problem. Showing a snippet of the data (5-10 rows) so readers can get a better understanding also helps.
- Do not suggest apply. It is not faster than a loop — it is actually slower, and consumes more memory.
- ne.evaluate is faster because there are no overheads wrt indexing alignment, NaNs, and mixed dtypes. Thought this should be made clear to readers.
- See if you can hide the perfplot output (for example, using https://stackoverflow.com/questions/6735917/redirecting-stdout-to-nothing-in-python)
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment