Why does loop tiling decrease running time of a function?

Asked Jan 24 '22 at 10:02

Active Jan 24 '22 at 10:20

Viewed 43 times

I am working with a function that transposes a matrix. My question is why does loop tiling decrease the running time especially of that function?

I understand why it works when, for example, multiplying matrices as the elements are being reused, but why is tiling faster when we still only access each element once in the transpose function?

edited Jan 24 '22 at 10:20

Jabberwocky

48,281
17
65
115

asked Jan 24 '22 at 10:02

AlphaList

1

Caching perhaps? – Some programmer dude Jan 24 '22 at 10:06
1

@lucidbrot yes, thank you! Haven't seen that post – AlphaList Jan 24 '22 at 10:14
Maybe you should [edit] and post both versions of your code. Also look at the generated assembly code which might give you a hint. – Jabberwocky Jan 24 '22 at 10:19

Why does loop tiling decrease running time of a function?

0 Answers0