Glenn wrote:
> If someone can explain what's happening at the tail end of the scanline for
> why I don't seem to have as many cycles left as it seems, I'd appreciate it.
I'll try :)
1. You have an error in your cycle counting, the second write to PF2 needs 3 cylces, so this operation ends at cycle 48 (not 47).
2. You have to count the additional code arround your PF-writes too. The maximum cylce count for this is 27 (19 before, 8 after the PF-writes).
So, you need a total of 48+27 = 75 cylces. That's why any additional nop will break that kernel.
But there might be (as always :) some optimizations possible to reduce those 27 additional cylcles.
Examples
- You don't have to decrease scanLine every loop, it would be enough to decrease only when X reaches 0. This requires just a different scanLine starting value.
- You could use Y to save more cycles.
- It might be also possible to avoid WSYNC by making the cycle count constant. But that will make the kernel very fragile, so I would do that near the end of the optimizing process.
- Put the 3 nops directly behind WSYNC and join those free cycles with the other free cycles before WSYN to gain more flexiblity here.
- ...
Have fun!
Thomas
_______________________________________________________
Thomas Jentzsch | *** Every bit is sacred ! ***
tjentzsch at web dot de |
_______________________________________________________________________
1.000.000 DM gewinnen - kostenlos tippen - http://millionenklick.web.de
IhrName@xxxxxx, 8MB Speicher, Verschluesselung - http://freemail.web.de
-
Archives (includes files) at http://www.biglist.com/lists/stella/archives/
Unsub & more at http://www.biglist.com/lists/stella/
