Re: Western design chip

[David.Empson@bbs.actrix.gen.nz]

In article <at76094@pro-gumbo.cts.com> tgeer@pro-gumbo.cts.com (System Administrator) writes:
> In <fmlin.3i11@terapin.com> fmlin@terapin.com (Frank M. Lin) writes:
> 
> >If you're lucky, you'll be able to go 15mhz like me.  :)  But I had to get 15ns
> >SRAMs, and increase my voltage to 5.6v to be here.  I'm at about 4x the speed
> >of a stock gs.
> 
> Pardon my nitpicking but if you're running at 15Mhz then:
> 
>      15Mhz / 2.8Mhz = 5.36 times the *rated* speed on a stock gs
> or
>      15Mhz / 2.6Mhz = 5.77 times the *effective* speed of a stock gs
> 
> a good bit more than 4 times the speed.

Not necessarily.  The speeds you give are the ideal speed increases,
assuming the accelerator can achieve 100% acceleration.  However, any
I/O access may cause the accelerator to 'go slow' for a short period,
accesses to non-cached memory locations will cause a normal access
cycle, and all writes (on the ZIP, at least) are performed at normal
speed.  [By 'normal', I mean 'unaccelerated', not necessarily '1 MHz'.]

Thus for typical operations, I would expect that a IIgs accelerated to
15 MHz would run about three to four times faster than a stock IIgs.
Very intensive graphics (lots of 1 MHz writes) would cut it down quite
a bit.
-- 
David Empson

Internet: David.Empson@bbs.actrix.gen.nz    EMPSON_D@kosmos.wcc.govt.nz
Snail mail: P.O. Box 27-103, Wellington, New Zealand