Raymond Hettinger <raymond.hettin...@gmail.com> added the comment:
The code for listiter_next() and listreviter_next() is almost the same. The main difference that the code for reversed saves the index to Py_ssize_t variable. Maybe that causes a 32-bit to 64-bit conversion and back. The change was made on 30 March 2016 in fbb1c5ee068d209e33f6e15ecb4821d5d8b107fa Another possible cause is that this is just a random build outcome due to PGO or incidental branch mis-prediction from aliasing (as described in https://stackoverflow.com/a/17906589/1001643 ). On the python.org mac 64-bit build, I'm not seeing any difference, so this is unique to the Windows 32-bit build. $ python3.7 -m timeit -r11 -s "a = list(range(250)) * 4000" "list(iter(a))" 100 loops, best of 11: 3.08 msec per loop $ python3.7 -m timeit -r11 -s "a = list(range(250)) * 4000" "list(reversed(a))" 100 loops, best of 11: 3.04 msec per loop $ python3.8 -m timeit -r11 -s "a = list(range(250)) * 4000" "list(iter(a))" 100 loops, best of 11: 3.07 msec per loop $ python3.8 -m timeit -r11 -s "a = list(range(250)) * 4000" "list(reversed(a))" 100 loops, best of 11: 3.01 msec per loop ---------- nosy: +rhettinger _______________________________________ Python tracker <rep...@bugs.python.org> <https://bugs.python.org/issue39521> _______________________________________ _______________________________________________ Python-bugs-list mailing list Unsubscribe: https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
