Date   : Sun, 12 Feb 2006 16:32:16 -0000
From   : "Mike" <profpep@...>
Subject: Re: MOS 3.20 v 3.50

> I did some speed tests a while back which involved MOS 3.50 amongst
> other things:
>
> http://www.beebmaster.co.uk/BeebHelp/6502Test.html
>
I did a lot of fractal work with BBC's and Masters. The trick with the
second processor is interleaving:  the math is done on the second processor,
and the graphics on the host. By a 'compute and plot' interleave the host
will be drawing whilst the parasite is doing the next calculation. I expect
the recoding speed was on the fill and shape algorithms, not the basic plot
stuff.

Running interpreted BASIC on both, a master Turbo would quite
comprehensively hammer a 16MHz Compaq 286.  Once I used compiled Pascal on
the Compaq, it won. I never had a comparable pair of compilers to make
furthre sensible tests. The Z80 BASIC is not slow, and didn't suffer from
page zero space, which cold nobble a recursive algorithm on the 6502. BAS32
on the 32016 seemed just about as fast as the Turbo, though I never timed
it. The advantage was the amount of space....

||\/||ike