search

llogiq / newlinebench

MIT License
22 stars 2 forks source link

MacBook Pro (Early 2015), 2.7 GHz Intel Core i5 #2

Closed regexident closed 7 years ago

regexident commented 8 years ago 
test test_fast_newlines          ... bench:       2,351 ns/iter (+/- 673)
test test_fast_nonewlines        ... bench:       2,326 ns/iter (+/- 613)
test test_fast_somenewlines      ... bench:          21 ns/iter (+/- 2)
test test_faster_newlines        ... bench:       3,411 ns/iter (+/- 2,687)
test test_faster_nonewlines      ... bench:       2,992 ns/iter (+/- 428)
test test_faster_somenewlines    ... bench:          26 ns/iter (+/- 5)
test test_fastest_newlines       ... bench:       1,351 ns/iter (+/- 437)
test test_fastest_nonewlines     ... bench:       1,309 ns/iter (+/- 197)
test test_fastest_somenewlines   ... bench:          16 ns/iter (+/- 2)
test test_screaming_newlines     ... bench:       1,982 ns/iter (+/- 328)
test test_screaming_nonewlines   ... bench:       1,924 ns/iter (+/- 526)
test test_screaming_somenewlines ... bench:          11 ns/iter (+/- 2)
test test_slow_newlines          ... bench:       5,896 ns/iter (+/- 784)
test test_slow_nonewlines        ... bench:       3,675 ns/iter (+/- 609)
test test_slow_somenewlines      ... bench:          31 ns/iter (+/- 4)
llogiq commented 8 years ago

Very interesting – it appears on newer hardware, POPCNT really shines, especially when amortized.