> The register file might be the toughest part. Each bit will take a
> Macrocell, so just a 4x24 consumes 75% of a 128 m-cell device. :-( A 24
> bit adder isn't exactly skinny either. I suppose there's tricks that can
> be played here though... You could keep the vectorlist in a large (32Kx8
> or 128Kx8 SRAM) and use a few bytes high in RAM to store the position
> counters, then just load the one you need at any given time. If you made
> an assumption that any delta value must be less than 1, you could
> accumulate delta's with a 12 bit adder and carry 1 with a separate 12bit
> add for the whole portion. Clock speed is on our side, so doing a bunch of
> operations isn't too much of a problem...
If the ram isn't fast enough, you can leave the last value read (a delta)
sitting on the data lines for the duration of the draw :-) I suppose we
can simulate the process by writing an equivalent line drawer for an
emulator - then we could test the required number of bits. I think 10.10
fixed point or *maybe* 10.8 would be enough. But I still want curves ;-)
-- ___ __ _ _ _ | \ / \ | | | || | phkahler@oakland.edu Engineer/Programmer | _/| || || |_| || |__ " What makes someone care so much? |_| |_||_| \___/ |____) for things another man can just ignore. " -S.H.Received on Wed Dec 17 06:43:43 1997
This archive was generated by hypermail 2.1.8 : Thu Jul 31 2003 - 23:01:05 EDT