[Coco] megaread

[Arthur Flexser]

I agree.  Every time the read head "just misses" a sector, and has to wait
for the disk to do a rotation to get to that sector, it costs two-tenths of
a second, which can add up quickly.  So seemingly small software
optimizations can make a big difference.
The test program would be a better speed indicator if written in ML.  Under
BASIC, it would be an improvement if it poked in the DSKCON track and sector
numbers ($00EB-00EC) and then directly EXECed DSKCON, rather than using
DSKI$.

Art
On Sun, May 23, 2010 at 3:48 PM, Darren A <mechacoco at gmail.com> wrote:


> > CLEAR 500
> > LOOP
> > C=C+1
> > FOR T = 0 TO 34
> > FOR S = 1 TO 18
> > DSKI$ 0,T,S,A$,B$
> > ?T;S,C
> > NEXT S,T
> > repeat
> >
>
> ---
>
> Floppy drive performance will fluctuate greatly depending on the
> particular usage, and how well it is tailored to the sector interleave
> (Skip Factor).  Compare the speed of the BASIC program above with the
> speed of the DSKINI verify pass (the second pass after the head moves
> back to track 0).  It does exactly the same thing (read all 630
> sectors sequentially) but is significantly faster.
>
> For many typical tasks DriveWire can out-perform a real floppy, but
> for operations that do a sequential scan of the disk using optimized
> software, a real disk will likely do better.
>