| ??? 11/03/06 16:49 Read: times |
#127332 - it's not quite that simple ... Responding to: ???'s previous message |
Erik Malund said:
What is the basis for your comment about the 85 MIPS rate?
experience, my stuff run in that neighborhood Can you direct me to any documentation that supports the sort of function I want to create? assuming you refer to the "small precisely timed loop": stated already http://www.8052.com/forum/read.phtml?id=127313 chapter 16 in the datasheet. Playing with the cache is "real fun" but once you get the drift of it, it's quite easy. one more comment: if the time is 'irrelevant' but the precision of the time is 'relevant' the simple solution is to make the linker save the particular routine at the same place every time. The only way the timing of a routine is 'unpredictable' is when the linker keep relocating it. ANY routine in ANY location will (save for interrupts, of course) ALWAYS execute in the same time till it is changed or relocated. Erik When I said "small," I meant less than 512 bytes. It's not that the time doesn't matter. What I have to do is write the loop so that, no matter what the input data is, which will force branching to different loop segments, the loop will always take the same amount of time, and remain within a very tight time budget. I haven't yet written the loop, but I've written similar loops for other MCU's, so I have some idea of what's involved. As you may recall, I don't use HLL's for MCU's. Maybe some day... RE |
