On Thu, 28 Mar 2002, Dexter wrote: > Using this method, I got a factor of 25 speed improvement on the previous > system, which gave time to do processing to the byte being transferred, > say to change color depth etc.
Supplemental, based on ZN's comments: I said byte, but yes, what he says is absolutely correct. It's quicker to fetch the largest block you can and then move it out. Assuming the two screens are in different pages, it also saves page switching time to switch pages as infrequently as possible. In my case, it was on an ARM processor, so I would fill R3 to R11 with 8 32-bit long word values, then write them out. Are we having fun yet? Dave
