When processing images, why is it recommended to first iterate over Y, and then X seconds?

Question

I read in several places that when processing pixels in an image, it is recommended to iterate over Y, and then X pixels, because it is more memory efficient. Why is this so?

Answer 1

A1 A2 A3 A4 
B1 B2 B3 B4 
C1 C2 C3 C4 
D1 D2 D3 D4 

Assuming that your image (and some coordinates to it indicates), it is stored in memory in a line, as:

A1 A2 A3 A4 B1 B2 B3 B4 C1 C2 C3 C4 D1 D2 D3 D4.

Let me simulate these parameters:

For X, and then through Y:

A1 B1 C1 D1 A2 B2 C2 D2 A3 B3 C3 D3 A4 B4 C4 D4 

Now check the line how messy the access will be (like random reading, right?)

Not, for Y, and then through X

A1 A2 A3 A4 B1 B2 B3 B4 C1 C2 C3 C4 D1 D2 D3 D4

Cm? Read it straight ahead!

That's why.

Answer 2

, .
, . , , .

, , , CPU . 8 16 cache line.
, , " ", , .

, CPU , .

, , , . , "" .

cache lines wikipedia.

-
, "" , .
- , , .