MIT's Introduction to Algorithms, Lectures 22 and 23: Cache Oblivious Algorithms

You're viewing a comment by Joseph Garvin and its responses.

Joseph Garvin Permalink
July 14, 2009, 04:05

@Wolter: I can think of two reasons.

1. The immediately practical reason: You'd need to recompile the code to achieve best performance (treating the cache size as a constant and not something that varies at run time). You could do this though and there might even exist a sufficiently advanced JIT somewhere.

2. The real reason: Because you can write a cache oblivious algorithm and have it perform best all the time without having to worry about that. The cache-aware-with-adjusting actually requires more complexity. The cache oblivious way "just works."

Reply to this comment

Reply To This Comment

(why do I need your e-mail?)

(Your twitter name, if you have one. (I'm @pkrumins, btw.))

Comment Help

Type first 3 letters of your name: (just to make sure you're a human)

Please preview the comment before submitting to make sure it's OK.