The proposed PR makes multi-argument functions in the math module atan2(), copysign(), remainder() and hypot() to use the fast call convention and inline arguments tuple unpacking.

Results:

$ ./python -m timeit -s "from math import atan2" "atan2(1.0, 1.0)"
Unpatched:  5000000 loops, best of 5: 79.5 nsec per loop
Patched:    5000000 loops, best of 5: 66.1 nsec per loop

$ ./python -m timeit -s "from math import copysign" "copysign(1.0, 1.0)"
Unpatched:  5000000 loops, best of 5: 90.3 nsec per loop
Patched:    10000000 loops, best of 5: 35.9 nsec per loop

$ ./python -m timeit -s "from math import remainder" "remainder(1.0, 1.0)"
Unpatched:  5000000 loops, best of 5: 69.5 nsec per loop
Patched:    5000000 loops, best of 5: 44.5 nsec per loop

$ ./python -m timeit -s "from math import hypot" "hypot(1.0, 1.0)"
Unpatched:  5000000 loops, best of 5: 63.6 nsec per loop
Patched:    5000000 loops, best of 5: 47.4 nsec per loop

New changeset d0d3e99120b19a4b800f0f381b2807c93aeecf0e by Serhiy Storchaka in branch 'master':
bpo-35719: Optimize multi-argument math functions. (GH-11527)
https://github.com/python/cpython/commit/d0d3e99120b19a4b800f0f381b2807c93aeecf0e