Dennis Sweeney <sweeney.dennis...@gmail.com> added the comment:
Benchmarks are definitively better with the most recent change (in which the
major data-copying loop is between parts of the same buffer):
-------------- MSVC ------------
Slower (3):
- (None,) * 2: 54.0 ns +- 0.7 ns -> 63.4 ns +- 2.5 ns: 1.17x slower
- [None] * 2: 61.6 ns +- 3.3 ns -> 69.0 ns +- 1.4 ns: 1.12x slower
- [None] * 10: 72.3 ns +- 2.2 ns -> 73.9 ns +- 2.3 ns: 1.02x slower
Faster (33):
- [None] * 10000: 25.1 us +- 0.2 us -> 12.8 us +- 0.2 us: 1.96x faster
- (None,) * 10000: 26.7 us +- 0.2 us -> 15.2 us +- 0.2 us: 1.76x faster
- list(range(1000)) * 100: 248 us +- 95 us -> 146 us +- 50 us: 1.70x faster
- list(range(1000)) * 10: 16.8 us +- 0.2 us -> 10.4 us +- 0.1 us: 1.61x faster
- list(range(100)) * 100: 13.8 us +- 0.3 us -> 8.60 us +- 0.09 us: 1.60x faster
- list(range(10)) * 100: 1.41 us +- 0.02 us -> 924 ns +- 15 ns: 1.53x faster
- list(range(100)) * 10: 1.47 us +- 0.10 us -> 981 ns +- 14 ns: 1.50x faster
- tuple(range(10)) * 10000: 255 us +- 91 us -> 171 us +- 51 us: 1.49x faster
- [None] * 100: 355 ns +- 11 ns -> 245 ns +- 7 ns: 1.45x faster
- (None,) * 100: 357 ns +- 4 ns -> 261 ns +- 3 ns: 1.37x faster
- tuple(range(1000)) * 10: 17.8 us +- 0.3 us -> 13.3 us +- 0.1 us: 1.34x faster
- list(range(10)) * 10000: 205 us +- 49 us -> 157 us +- 95 us: 1.31x faster
- list(range(1000)) * 2: 3.32 us +- 0.07 us -> 2.63 us +- 0.03 us: 1.26x faster
- tuple(range(100)) * 100: 15.9 us +- 0.2 us -> 13.0 us +- 0.1 us: 1.22x faster
- list(range(1000)) * 10000: 33.9 ms +- 0.9 ms -> 28.2 ms +- 1.6 ms: 1.20x
faster
- ["Python", "Perl"] * 10000: 31.4 us +- 4.3 us -> 26.2 us +- 0.7 us: 1.20x
faster
- tuple(range(1000)) * 10000: 33.8 ms +- 0.8 ms -> 28.3 ms +- 1.3 ms: 1.19x
faster
- tuple(range(100)) * 10: 1.68 us +- 0.02 us -> 1.42 us +- 0.01 us: 1.18x faster
- ["Python", "Perl"] * 100: 442 ns +- 67 ns -> 376 ns +- 6 ns: 1.18x faster
- tuple(range(10)) * 100: 1.63 us +- 0.02 us -> 1.40 us +- 0.02 us: 1.17x faster
- tuple(range(1000)) * 2: 3.61 us +- 0.06 us -> 3.11 us +- 0.06 us: 1.16x faster
- list(range(10)) * 10: 238 ns +- 8 ns -> 206 ns +- 6 ns: 1.16x faster
- list(range(100)) * 10000: 3.15 ms +- 0.13 ms -> 2.73 ms +- 0.15 ms: 1.15x
faster
- list(range(100)) * 2: 370 ns +- 12 ns -> 323 ns +- 6 ns: 1.14x faster
- tuple(range(100)) * 10000: 3.24 ms +- 0.13 ms -> 2.85 ms +- 0.13 ms: 1.14x
faster
- ("Python", "Perl") * 10000: 31.1 us +- 0.3 us -> 27.4 us +- 0.3 us: 1.14x
faster
- ("Python", "Perl") * 100: 438 ns +- 7 ns -> 389 ns +- 9 ns: 1.13x faster
- ["Python", "Perl"] * 10: 86.1 ns +- 4.2 ns -> 77.3 ns +- 2.4 ns: 1.11x faster
- list(range(10)) * 2: 81.5 ns +- 2.8 ns -> 77.0 ns +- 2.1 ns: 1.06x faster
- ("Python", "Perl") * 10: 92.6 ns +- 7.8 ns -> 88.4 ns +- 4.8 ns: 1.05x faster
- ["Python", "Perl"] * 2: 63.9 ns +- 2.6 ns -> 61.3 ns +- 1.7 ns: 1.04x faster
- tuple(range(100)) * 2: 414 ns +- 19 ns -> 404 ns +- 6 ns: 1.02x faster
- tuple(range(10)) * 10: 258 ns +- 7 ns -> 253 ns +- 8 ns: 1.02x faster
Benchmark hidden because not significant (4): ("Python", "Perl") * 2,
tuple(range(10)) * 2, (None,) * 10, tuple(range(1000)) * 100
Geometric mean: 1.21x faster
-------- GCC PGO on WSL --------
Slower (4):
- [None] * 10: 58.1 ns +- 0.7 ns -> 59.4 ns +- 0.9 ns: 1.02x slower
- ("Python", "Perl") * 2: 48.9 ns +- 0.6 ns -> 49.9 ns +- 0.6 ns: 1.02x slower
- [None] * 1000: 1.56 us +- 0.02 us -> 1.59 us +- 0.02 us: 1.02x slower
- [None] * 2: 51.9 ns +- 0.5 ns -> 52.4 ns +- 0.8 ns: 1.01x slower
Faster (44):
- list(range(100)) * 10000: 1.98 ms +- 0.33 ms -> 976 us +- 43 us: 2.03x faster
- list(range(1000)) * 100: 168 us +- 2 us -> 84.8 us +- 1.0 us: 1.98x faster
- list(range(1000)) * 10: 15.9 us +- 0.2 us -> 8.47 us +- 0.20 us: 1.88x faster
- list(range(100)) * 100: 13.1 us +- 0.2 us -> 7.02 us +- 0.12 us: 1.87x faster
- list(range(1000)) * 1000: 1.80 ms +- 0.05 ms -> 986 us +- 12 us: 1.83x faster
- list(range(100)) * 1000: 140 us +- 2 us -> 79.1 us +- 0.9 us: 1.77x faster
- list(range(10)) * 10000: 136 us +- 2 us -> 77.5 us +- 1.0 us: 1.76x faster
- list(range(1000)) * 10000: 26.0 ms +- 0.6 ms -> 15.2 ms +- 0.3 ms: 1.71x
faster
- list(range(10)) * 1000: 12.2 us +- 0.2 us -> 7.18 us +- 0.08 us: 1.70x faster
- tuple(range(1000)) * 100: 167 us +- 2 us -> 106 us +- 1 us: 1.58x faster
- tuple(range(1000)) * 1000: 1.78 ms +- 0.03 ms -> 1.13 ms +- 0.01 ms: 1.58x
faster
- tuple(range(10)) * 10000: 148 us +- 7 us -> 95.5 us +- 1.1 us: 1.55x faster
- list(range(10)) * 100: 1.19 us +- 0.10 us -> 795 ns +- 20 ns: 1.49x faster
- tuple(range(1000)) * 10: 15.9 us +- 0.4 us -> 10.6 us +- 0.2 us: 1.49x faster
- list(range(100)) * 10: 1.17 us +- 0.03 us -> 797 ns +- 17 ns: 1.47x faster
- (None,) * 10000: 27.5 us +- 0.3 us -> 18.9 us +- 0.1 us: 1.46x faster
- tuple(range(100)) * 1000: 138 us +- 2 us -> 95.4 us +- 0.7 us: 1.45x faster
- tuple(range(100)) * 10000: 1.52 ms +- 0.02 ms -> 1.05 ms +- 0.01 ms: 1.45x
faster
- tuple(range(10)) * 1000: 12.9 us +- 0.2 us -> 9.28 us +- 0.18 us: 1.39x faster
- list(range(1000)) * 2: 3.05 us +- 0.07 us -> 2.22 us +- 0.04 us: 1.37x faster
- tuple(range(100)) * 100: 13.0 us +- 0.2 us -> 9.53 us +- 0.28 us: 1.37x faster
- tuple(range(1000)) * 10000: 23.2 ms +- 0.4 ms -> 17.0 ms +- 0.3 ms: 1.36x
faster
- (None,) * 1000: 2.66 us +- 0.02 us -> 1.97 us +- 0.02 us: 1.35x faster
- ("Python", "Perl") * 10000: 33.8 us +- 0.8 us -> 25.4 us +- 0.2 us: 1.33x
faster
- tuple(range(10)) * 10: 236 ns +- 17 ns -> 178 ns +- 1 ns: 1.33x faster
- ["Python", "Perl"] * 1000: 3.41 us +- 0.03 us -> 2.67 us +- 0.03 us: 1.28x
faster
- ["Python", "Perl"] * 10000: 33.8 us +- 0.3 us -> 26.4 us +- 0.2 us: 1.28x
faster
- list(range(10)) * 10: 172 ns +- 2 ns -> 135 ns +- 2 ns: 1.27x faster
- ("Python", "Perl") * 1000: 3.18 us +- 0.03 us -> 2.56 us +- 0.03 us: 1.24x
faster
- ("Python", "Perl") * 100: 428 ns +- 9 ns -> 345 ns +- 4 ns: 1.24x faster
- (None,) * 100: 306 ns +- 4 ns -> 248 ns +- 2 ns: 1.23x faster
- ["Python", "Perl"] * 100: 440 ns +- 8 ns -> 360 ns +- 14 ns: 1.22x faster
- tuple(range(1000)) * 2: 3.03 us +- 0.06 us -> 2.49 us +- 0.03 us: 1.22x faster
- tuple(range(10)) * 100: 1.25 us +- 0.08 us -> 1.04 us +- 0.03 us: 1.20x faster
- tuple(range(100)) * 10: 1.22 us +- 0.02 us -> 1.05 us +- 0.01 us: 1.17x faster
- list(range(100)) * 2: 320 ns +- 19 ns -> 275 ns +- 10 ns: 1.16x faster
- ["Python", "Perl"] * 10: 85.0 ns +- 3.1 ns -> 73.7 ns +- 1.3 ns: 1.15x faster
- (None,) * 10: 63.4 ns +- 1.7 ns -> 55.4 ns +- 0.6 ns: 1.14x faster
- tuple(range(100)) * 2: 345 ns +- 7 ns -> 318 ns +- 13 ns: 1.09x faster
- ("Python", "Perl") * 10: 88.6 ns +- 2.7 ns -> 83.3 ns +- 1.9 ns: 1.06x faster
- list(range(10)) * 2: 73.5 ns +- 1.2 ns -> 70.7 ns +- 1.9 ns: 1.04x faster
- tuple(range(10)) * 2: 85.5 ns +- 3.9 ns -> 82.7 ns +- 1.0 ns: 1.03x faster
- ["Python", "Perl"] * 2: 56.1 ns +- 0.6 ns -> 54.5 ns +- 0.6 ns: 1.03x faster
- (None,) * 2: 46.7 ns +- 0.5 ns -> 45.6 ns +- 0.9 ns: 1.02x faster
Benchmark hidden because not significant (2): [None] * 100, [None] * 10000
Geometric mean: 1.32x faster
----------
_______________________________________
Python tracker <rep...@bugs.python.org>
<https://bugs.python.org/issue46235>
_______________________________________
_______________________________________________
Python-bugs-list mailing list
Unsubscribe:
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
