Small clean-ups for the random module

[rhettinger]

Timings:

$ python3.9 -m timeit -r 11 -s 'from math import floor' -s '_int = int' -s 'x=3.14' '_int(x)'
5000000 loops, best of 11: 68 nsec per loop
$ python3.9 -m timeit -r 11 -s 'from math import floor' -s '_int = int' -s 'x=3.14' 'floor(x)'
10000000 loops, best of 11: 36.7 nsec per loop

[rhettinger]

The default argument _int = int is no longer of much value. The global variable lookup is now slightly cheaper than the cost of the extra default argument for the first lookup.

Baseline timings:

$ pytime -s 'from random import randrange' 'randrange(1000)'
500000 loops, best of 11: 559 nsec per loop
$ pytime -s 'from random import randrange' 'randrange(1000, 2000)'
500000 loops, best of 11: 744 nsec per loop
$ pytime -s 'from random import randrange' 'randrange(1000, 2000, 100)'
200000 loops, best of 11: 1.04 usec per loop

New timings:

$ pytime -s 'from random import randrange' 'randrange(1000)'
500000 loops, best of 11: 556 nsec per loop
$ pytime -s 'from random import randrange' 'randrange(1000, 2000)'
500000 loops, best of 11: 747 nsec per loop
$ pytime -s 'from random import randrange' 'randrange(1000, 2000, 100)'
200000 loops, best of 11: 1.05 usec per loop

[tim-one]

Looks fine to me. Not sure what's up with int v floor - the latter is much faster (over factor of 3) for me under a current master build (Win64) than under the released 3.8.1.

[miss-islington]

Thanks @rhettinger for the PR 🌮🎉.. I'm working now to backport this PR to: 3.9.
🐍🍒⛏🤖

[bedevere-bot]

GH-21065 is a backport of this pull request to the 3.9 branch.