Question

I have a script where I am computing:

def sumsquared(arr):
    sum = 0
    idx = 0
    len = arr.__len__()
    while idx < (len - 1):
            sum = sum + (arr[idx] * arr[idx]) + (arr[idx+1] * arr[idx+1])
            idx = idx + 2

    return sum

The above function is called in a loop which populates two lists and calls this function twice: First time with a list of len ~ 1024 items and second time with len ~ 44100 items. The loop itself can run anywhere from 100 to 100000 times depending upon the input.

For a small sized input, cProfile based profiling informs me:

ncalls  tottime  percall  cumtime  percall  filename:lineno(function)
---------------------------------------------------------------------
 2560   12.065   0.005    12.065    0.005    beat.py:8(sumsquared)

which is about 95% of the total running time for the script. Is there some way I can get some speed up on the function?

Answer 1

This is the fastest I can find

from itertools import imap
from operator import mul
def sumsquared(arr):
    return sum(imap(mul, arr, arr))