You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
At some point (if feasible), it might be a good idea to measure how much overhead NumPy and PyData/Sparse incur when functions are invoked. That way, we can verify that any observed difference between NumPy and TACO is not just due to the overhead of interpreting Python functions or something trivial like that. This might require modifying NumPy's and PyData/Sparse's source code to insert some timers inside the functions we are benchmarking.