4 ms·
> We usually think of I/O as expensive, but I/O isn’t the bottleneck here...The tokenization and hash table operations are the bottleneck by a long shot Intere
by ijidak 4y ago
> We usually think of I/O as expensive, but I/O isn’t the bottleneck here...The tokenization and hash table operations are the bottleneck by a long shot
Interesting. This is something I'll have to keep in mind.
Best practice is always been to consider I/O the slowest.
But, it's true... times have changed.
Maybe we shouldn't make this assumption anymore.
- bjourne 4y agoHe didn't clear the disk cache when profiling so of course the IO overhead was negligible. Had he flushed the cache between runs the numbers would have been very different.
- igouy 4y agoMeasurements of the similar k-nucleotide programs show that effect — N=250,000 compared to N=2,500,000 https://benchmarksgame-team.pages.debian.net/benchmarksgame/measurements/java.html https://benchmarksgame-team.pages.debian.net/benchmarksgame/...
- FridgeSeal 4y agoHere’s a couple of articles on the performance of IO and slowness of other stuff that I find particularly interesting: https://itnext.io/modern-storage-is-plenty-fast-it-is-the-apis-that-are-bad-6a68319fbc1a https://itnext.io/modern-storage-is-plenty-fast-it-is-the-ap... https://gregoryszorc.com/blog/2021/04/06/surprisingly-slow/ https://gregoryszorc.com/blog/2021/04/06/surprisingly-slow/