I’m just curious, are BufferedImages faster if they’re in powers of two? (ie 16x16 or 64x64) And does it make them any faster is they’re of equal width and length?
It may help a little, depending on the platform, but I wouln’t count on it.
But, in case an image gets accelerated (so a cached version is created in video ram), some drivers allow only power of two dimensions for offscreen surfaces (i.e, textures), so we have to round the size of the surface to be allocated to the next power of two. Which is, of course, waste of precious video memory - for example, we’ll have to create 128x128 offscreen surface for a 65x65 image).
Faster? No I don’t think so.
But as the previous post points out…
More efficient use of resources? Definitely.