Really it was related to my first question - I was considering the
plausability of moving my data table to be positioned around address
zero when I wrote the post. And obviously a table that reaches 2kb in
either direction from address 0 covers address 0x0038.

Tests last night with an older, slower version of multiply, suggest
that I'm now using around 4,000 cycles to transform a point in world
space (9 multiplies, 6 adds) and another 5,000ish to project it to the
screen (2 divides, 2 adds, some hard-coded-with-shifts multiplying by
128 and 96). Moving the multiply table so that it is centred around 0
should cut about 1/4 from the transformation time, so 1/9 from the
time I spend on points. Which I think is worth it. But obviously I
need to do something about my divides.

On Wed, May 21, 2008 at 2:29 PM, Colin Piggot <[EMAIL PROTECTED]> wrote:
> Adrian Brown wrote:
>> Ok, i haven't read all the posts on this, but why not stick the code in
>> LMPR and use IM1 - saves having the table of vectors.
>
> That's my favoured approach too.
>
>
>> ** UIP Sam Port 4100+ lines of z80 and climbing
>
> Woo! It's so tantilisingly close!
>
>
> David Brant wrote:
>> But then again is there any hardware for the SAM that uses them?
>
> As far as I'm aware there's nothing.
>
> Colin
> =====
> Quazar : Hardware, Software, Spares and Repairs for the SAM Coupe
> 1995-2008 - Celebrating 14 Years of developing for the SAM Coupe
> Website: http://www.samcoupe.com/
>

Reply via email to