2 ms·
>However, the constant coefficient hidden by the Big O notation is so large that these algorithms are only worthwhile for matrices that are too large to handle
by ChrisRackauckas 9y ago
>However, the constant coefficient hidden by the Big O notation is so large that these algorithms are only worthwhile for matrices that are too large to handle on present-day computers.
No wonder I never heard about these. I updated it for correctness (and fixed the link, ugh) but yeah, these are details I'd hope to not dwell on in the post since for all intents and purposes you'd still never want to do the worst fallback unless you have to.
- tmyklebu 9y agoYep. (Largely) irrelevant in practice. But the gap between what we can do and what lower bounds we can prove is very large indeed for many matrix operations.