perf: specialize cmp, const_eq, const_is_zero for small sizes
#561Comparing
DaniPopes:dani/cmp-specialization
(b89a765
) withmain
(c147f59
)+13%
- Improvement
- 4
- Untouched
- 382
Benchmarks
386 total
root/127/128/128
benches/bench.rs::benches::benches::group+13%
9.7 ms8.6 ms
to/f64/4096
benches/bench.rs::benches::benches::group+11%
29.4 µs26.4 µs
to/f32/4096
benches/bench.rs::benches::benches::group+11%
29.2 µs26.3 µs
most_significant_bits/4096/4096
benches/bench.rs::benches::benches::group+11%
28.9 µs26.1 µs
root/2/128/128
benches/bench.rs::benches::benches::group+5%
4.7 ms4.5 ms
parse/hex/4096/zero
benches/bench.rs::benches::benches::group+5%
62.5 µs59.6 µs
parse/binary/4096/zero
benches/bench.rs::benches::benches::group+5%
62.5 µs59.6 µs
parse/decimal/4096/zero
benches/bench.rs::benches::benches::group+5%
62.5 µs59.6 µs
parse/octal/4096/zero
benches/bench.rs::benches::benches::group+5%
62.5 µs59.6 µs
neg/4096
benches/bench.rs::benches::benches::group+4%
72.9 µs70 µs
root/3/128/128
benches/bench.rs::benches::benches::group+4%
4.3 ms4.2 ms
root/5/128/128
benches/bench.rs::benches::benches::group+4%
4.8 ms4.7 ms
inv_mod/128
benches/bench.rs::benches::benches::group+3%
6.3 ms6.1 ms
sub/4096
benches/bench.rs::benches::benches::group+3%
96.1 µs93.1 µs
add/4096
benches/bench.rs::benches::benches::group+3%
96.1 µs93.2 µs
cmp/64
benches/bench.rs::benches::benches::group+3%
108.2 µs105.4 µs
gt/128
benches/bench.rs::benches::benches::group+2%
189.4 µs185.2 µs
ge/128
benches/bench.rs::benches::benches::group+2%
189.4 µs185.3 µs
le/128
benches/bench.rs::benches::benches::group+2%
189.4 µs185.3 µs
lt/128
benches/bench.rs::benches::benches::group+2%
189.4 µs185.3 µs
log/128
benches/bench.rs::benches::benches::group+2%
7.4 ms7.2 ms
log/64
benches/bench.rs::benches::benches::group+2%
1.5 ms1.4 ms
reduce_mod/128
benches/bench.rs::benches::benches::group+1%
578.3 µs572.7 µs
pow_mod/128
benches/bench.rs::benches::benches::group+1%
2.4 ms2.4 ms
is_zero/4096
benches/bench.rs::benches::benches::group+1%
4.1 µs4.1 µs
Commits
Click on a commit to change the comparison rangeBase
main
c147f59
+12.68%
perf: specialize Ord::cmp for small sizes
b89a765
23 hours ago
by DaniPopes