[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Questions on Accelerators and RamWorks III

On 11/2/2010 5:58 PM, mdj wrote:

On Nov 3, 5:26 am, dog_...@macgui.com (D Finnigan) wrote:

The last piece of desired equipment for my "dream" IIe system is an
accelerator card. Before I pry open my wallet, I'd like to have some
questions satisfied.

First, I've done some research on various accelerators. From what I've read,
not all of them "accelerate" the 64 KB AUX RAM area on a IIe, while others
do. Ideally, since I'd be using a RamWorks III card, I'd like to know if
there are any accelerators which can accelerate all banks of RAM.


For Ramworks compatibility, you need a Transwarp or newer accelerator.
None of them can accelerate more than 1mb of Ramworks memory. The Zip
uses a fairly standard tagged RAM cache. The Transwarp uses one of its
64k banks for caching the Ramworks, and I *think* it simply
invalidates the cache when a bank switch occurs. Simple enough to
determine with a test program if you care enough.


The Zip Chip caches up to 2MB of RamWorks-style memory.


The Transwarp II is the clear pick of the bunch, but they're almost
pure unobtainium.

The best accelerator you're likely to *find* is a Zip 8000. However,
Zip Chips are notoriously unreliable and you'll find that people are
loathe to part with them.


From my experience, only the first run of Zip Chips were unreliable--
the ones in which the tops of the two *packaged* SRAM chip packages are
visible on the underside of the Zip Chip seem reliable.


Michael and I did some limited objective testing of a Zip vs a
Transwarp a few years back (thread should be easy to find). The Zip is
a *clear* winner.  Subjectively I'd say there's not that much
difference, so unless you're running a huge assembly job (which may
well be IO bound anyway) either one provides a substantial improvement
over stock.

The Transwarp has some advantages over the Zip: being card based it's
more reliable, and relatively easy to repair. It doesn't exhibit that
annoying pause on ctrl-reset that the Zip Chips have. Lastly you can
install a 65816 part in place of its stock R65C02 and run Merlin-16
which is both more feature rich and considerably faster. Merlin-8
makes extensive use of SWEET16 internally, so the switch to 16 bit has
a large performance improvement. The 16 bit full screen editor is also
much nicer than the 8 bit version.


In the last 15 years, I haven't had *any* problems with my Zip Chips.
Perhaps the biggest problem for the later run chips is mistreatment
resulting in broken pins.  Easy solution:  don't remove it!  If you
don't want it to be active, hold down Escape during reset.


Disadvantages are: current draw (make sure your power supply is
fresh), you lose a slot (although slot 3 isn't useful for much else),
and of course, performance.


Right--it provides a *lot* of heat...


Furthermore, I'd like to know if there are any incompatibilities with
accelerators and mass storage systems such as the FocusDrive. I suspect that
since the Focus does not use IDE, there are none. But I'd like to confirm.


I think you mean DMA... Seriously, DMA on an 8 bit Apple II isn't
really much of a win, since ProDOS 8 forces all disk access to be "one
block at a time". The accelerator will more than compensate you, and
from experience I can tell you that a Focus on a Transwarped or Zipped
IIe is *wicked* fast.


Agreed--skip DMA and make your life simpler.  Of course, that rules out
the original SoftCard-style Z80 cards (no big loss ;-).


Lastly, would an accelerator affect the timing of a clock card, such as the
Timemaster II H.O.? Again, I suspect not, but I'd like to ask anyway.


It won't. BTW, the Timemaster II has a 1024Hz interrupt generator, so
combined with an accelerator you can do millisecond resolution
instrumentation.


All accelerators offer slot-by-slot control of acceleration to deal with
cards that use CPU loops for timing (the Disk ][ Controller card is, of
course, the king of CPU timing ;-).  In addition, the Zip Chip offers
controls of speaker acceleration and joystick acceleration (both usually
run at 1MHz.).


In closing, if anyone has suggestions for what I should get, please tell.
I'll be posting a WTB advertisement in the csa2.marketplace newsgroup
shortly.


I suspect it'll be more a case of taking what you can get, than being
picky. Since virtually all 8 bit software was designed to run well on
a stock machine, any accelerator gives a substantial subjective
improvement over stock speed. If you happen to get one of the faster
ones, good for you, but if someone offers you a Transwarp I'd
recommend picking it up.


It's true, they don't come around that often.  I'd get a Zip Chip over
a TransWarp, but that's a matter of taste and experience.  ;-)

-michael

NadaNet 3.1 for Apple II parallel computing!
Home page:  http://home.comcast.net/~mjmahon/

"The wastebasket is our most important design
tool--and it's seriously underused."