[no subject]

This is a venture with Linus Torvalds (of Linux fame) and I know IBM
are manufacturing the chips.  The current chip is Crusoe which is a
very low power x86 compatible.  The idea is that with very few
transistors and software instruction set means very low power
consumption and a battery life of days rather than hours.

Anyway, take a look at www.transmeta.com and discuss!

What about a PC that can switch to native GS mode at a high clock
speed with one or two 8-bit Apple II slots and all the bits of a GS
built in.  Perhaps the latest Ensoniq chipset that is backward
compatible to the GS one, Second-sight or a newer variant, SCSI etc
built in?  I'm sure the hardest part would be getting the custom
chipsets for ie SWIM controllers.  Perhaps we could forgo the
SuperDrive as many people can use ZIP drives as an alternative.

If we were being realistic this is what I would like to see in a new
GS...

Primarily a new modern architecture and new OS.  Perhaps based on
GS/OS, perhaps not.
Capable of natively running GS/OS software or seamless emulation.
SuperSerial port or GS-serial port (Appletalk)
ADB & USB
IDE & SCSI connector (SCSI II enough?  Or U2W?)
SmartPort

Hmmm... Would we really need 8-bit slots if this lot was built-in?

Perhaps yet another way to approach this is a peripheral card for PC
or Mac which has the missing ports on it.  Then an emulator could be
written to use that card.  Whats the chances of a PCI peripheral card
sporting ADB, Smartport, SuperSerial and perhaps an Ensoniq?

Just my GBP 0.02  ;-)
--
John





Subject: Re: All this talk of a new GS...
From: <A HREF="mailto:mystix1210@my-deja.com";>mystix1210@my-deja.com</A>
Date: Mon, Jan 24, 2000 11:02 PM
Message-id: <86jhqp$qr4$1@nnrp1.deja.com>

In article <388cbf13.11507897@news.demon.co.uk>,
  lefebvre@klatch.demon.co.uk (John & Anne Le Febvre) wrote:

> Hmmm... Would we really need 8-bit slots if this lot was built-in?
>
> Perhaps yet another way to approach this is a peripheral card for PC
> or Mac which has the missing ports on it.  Then an emulator could be
> written to use that card.  Whats the chances of a PCI peripheral card
> sporting ADB, Smartport, SuperSerial and perhaps an Ensoniq?

Well I'd really want to bring over the VOC, being as a reasonable
priced genlock for a PC or Mac is out of my price range.




For more droolings over the possible cores for transmeta's chip, follow
comp.sys.c64. =)

Especially when run at a (very) modest 100mhz.  That would take the IIgs to
previously unheard-of speeds =)

> built in?  I'm sure the hardest part would be getting the custom
> chipsets for ie SWIM controllers.  Perhaps we could forgo the

I didnt think apple was too kind about letting us know about the SIWM chips
(it's not swim; it's siwm - super integrated woz machine)..  I'd love to
know some about the super ones. =)

> SuperDrive as many people can use ZIP drives as an alternative.
> 
> If we were being realistic this is what I would like to see in a new
> GS...
> 
> Primarily a new modern architecture and new OS.  Perhaps based on
> GS/OS, perhaps not.

If you have a look at <A HREF="http://www.clips64.de";>http://www.clips64.de</A>
, you'll see what the c-64 and
it's 20mhz 16-meg '816 accelerator is in for.  All the author has so far is
screenshots for public view, but I've had opportunity to play with the
betas.  It's a VERY impressive OS that I'd love to see on more than just the
c-64.  It'd feel quite welcome on a //gs, considering its advanced video
system.  My programming group has been asked to port our module player to it
once complete, so I imagine CLiPS will have some good support.  Some of us
are even talking about tv-in-a-window (most likely in black/white but still
a very cool concept given the machine's age).  See below =)

> Capable of natively running GS/OS software or seamless emulation.
> SuperSerial port or GS-serial port (Appletalk)
> ADB & USB

Oh, for an ADB port on my c64.. I love these mac keyboards =) For USB, I'm
shopping about for host controllers to be used on a stock c64.  If someone
hardware-knowledged in the apple2 realm would be interested in aiding a
cross-port of the hardware so that you could have USB on your IIgs's and
IIe's,  Let me know via email.  The primary interest for USB currently is
the Happauge WinTV USB device that allows (currently) windows systems to get
tv on their desktop via USB.  We intend on using such a device to do small
video captures, possibly put video on a CLiPS desktop, and other cool fun
things.

> IDE & SCSI connector (SCSI II enough?  Or U2W?)

SCSI-II is quite fast. =) If you could use FAST scsi, you could reach
20meg/sec.  Quite enough for a //gs environment, where 8 megs ram is the
max. (am I wrong here?)  I've decided my project, at 1meg/sec, is speedy
enough for our 15.5meg max super64's, so I sit next to a prototype adaptor
equipped with the same chip that powers the vintage macs' scsi system.

> SmartPort
> Hmmm... Would we really need 8-bit slots if this lot was built-in?

Of course! You never know what one might find that would be fun to have in
the ol' apple =)

> Perhaps yet another way to approach this is a peripheral card for PC
> or Mac which has the missing ports on it.  Then an emulator could be
> written to use that card.  Whats the chances of a PCI peripheral card
> sporting ADB, Smartport, SuperSerial and perhaps an Ensoniq?

A-la the Siamese systems Amiga card for PC's.  I hear it's quick, but I also
hear of scads of compatibility problems.  However, it's more likely to see
one of these than a PC that's designed to run a vintage computer's software. 
Dont let me discourage you though!! It's possible, and I'd love to see it.

> Just my GBP 0.02  ;-)

Plus another 4 cents or so =)

> --
> John
> 

-jonas






> > IDE & SCSI connector (SCSI II enough?  Or U2W?)
> 
> SCSI-II is quite fast. =) If you could use FAST scsi, you could reach
> 20meg/sec.

That would be Fast/Wide, or Ultra/Narrow.

(Standard is 5 MB/sec, Fast is 10 MB/sec, assuming narrow, i.e. 8 bits.)


				>>>Right. my error. =)<<<


> Quite enough for a //gs environment, where 8 megs ram is the max. (am I
> wrong here?)

8 MB is the practical limit for the current motherboard design.  In
theory, a new design might be able to use as much as 14 MB of RAM by
extending the fast RAM through banks $80 to $DF.  This would require
changes to the Memory Manager toolset (and would cause compatibility
problems with the TransWarp GS for certain, and probably also the Zip
GS; who cares if we're talking about a faster processor anyway!).

-- 
David Empson
dempson@actrix.gen.nz
Snail mail: P O Box 27-103, Wellington, New Zealand




This is already being done in Bernie, the IIgs emulator. They haven't
been able to access bank $DF - this seems to be in use by the system
in some way or another - but it's possible to use all the other
banks, from 00 to DE, and the software I've tried all work without
showing any signs of incompatibility. So no problem there.
If you want to know what it takes to have the Memory Manager be
14MB-aware, contact Andre Horstmann.

> --
> David Empson
> dempson@actrix.gen.nz
> Snail mail: P O Box 27-103, Wellington, New Zealand


-- 
Groetjes, Pim

To send me e-mail, insert a t right before the dot.
Which reminds me, I'd better get on with... IT!





> If you want to know what it takes to have the Memory Manager be
> 14MB-aware, contact Andre Horstmann.

It isn't exactly a major patch.  I think that the only change needed is
in MMBootInit, where it does its RAM test: it currently starts at bank
$7F and works backwards.  (There is probably a related piece of code
that may assume that all banks $80-$DF are to be reserved as a single
block.)

-- 
David Empson




Could such a patch be made to the existing GS/OS v6.0.1 to impliment 
the use of 14 MB's on a REAL IIgs (either ROM 3 or ROM 1 or both) ?

Bernie is indeed 14MB-aware.  And, in my use with Gus - it goes even futher
with being 16MB-aware.  I'm wondering on this (if Dave Lyons and the other
authors of Gus are reading this post - PLEASE REPLY)  perhaps
any (and/or all) of you could reply to this post with the method used to allow
this with Gus and tell us if such actually could be applied with a patch to a
REAL IIgs - to allow such with it.  Looking forward to any reply, info or
suggestions anybody cares to offer!

Cheers,
Tom





As a suggestion to avoid compatibility issues. How about implementing
a virtual memory system for memory over the 8 meg limit. This way, it
could take advantage of speedy ram, or the slower disk space if ram is
unavailable, or a combination of both. IMHO, this would be better than
some systems implementing >8m and others who follow the original spec's
as well as the original hardware which is limited to 8 megs.

Phoenyx






Oooooo. =) I'd love to see something like this.  I agree with the
accelerator compatiblity problems, but since there'll be no need in a
(dreaming here) 200mhz '816 environment, it doesnt matter.

Let's not drop out on this one, it's a quite feasible idea.  The only real
challenge is figuring out how to implement the custom grpahics hardware
employed by the IIgs as well as any copyrights that Wdesignc (the '816
creator) might hold.  They _do_ have a core one could license for PGA/CPLD
units.  If you dont mind tossing bux in for the core and putting it on a
highspeed PLD of sorts, it's not unrealistic to guess at 100mhz+ operation.
<grin>..

For more info on our beloved '816 CPU, check out <A
HREF="http://www.wdesignc.com";>http://www.wdesignc.com</A>

-Jonas






What about an embedded processor like a 386 or 486 running an optimized
IIGS core.  Virtual memory would be easier considering the 386 and up
has it built into the chip.  Then you would have a super fast processor
using cheap hardware.  You could have an effective 4 gigabyte address
space!  Well, something like that anyway.





With the speed of the newer cpu's one could implement a true
multi-tasking GS. Such a system could have it's own memory space as
well as a shared heap.

Phoenyx





If you look closely, you'll see that a K6II-400 is cheaper than you'll find
most embedded 386s and 486s these days.  PII-400's sell for $100 at a
certain irritating auction site <grin> quite often.  Cyrix cpu's, with their
quite superior integer math machine, are almost a dime a dozen =)

-Jonas







I was talking about building a non-IBM compatible system that connects
directly to the IIgs CPU socket.  I was not talking about the embedded
systems that can run Win95 and can support PC-104 cards.  I was thinking
of building a system that runs a core with none of the crap that is in
the PC now.  And yes, if this project were to mature to bigger faster
processors such as the 400Mhz machines that would be great.  I guess
when people think of "embedded" they think of an embedded PC.  I want
this thing to fit in the original IIgs case, but not be a full PC.





++++++++++++++
On Thu, 3 Feb 2000, Phoenyx wrote:

> Date: Thu, 03 Feb 2000 20:38:12 -0600
> From: Phoenyx <quazar@dcnet2000.com>
> Newsgroups: comp.sys.apple2
> Subject: Re: All this talk of a new GS...
> 
> I think at this point, adhering to as many standards as possible is a
> must. Part of the problems the early Apple2 faced
> was because there were no standards which left developers in
> the position of making their own. In the floppy driven world of the A2,
> this wasn't too bad, but when the system became able to
> access hard drives, it became more of an issue.

That's something I hadn't thought about, even though I should be shot for
failing.

> There are still other issues to deal with. Perhaps the biggest
> one is the GS rom. Apple may ignore the emulators, but if a
> hardware project were developed, they may decide it's time to
> examine the copyright issues. Since Apple doesn't care about
> any growth on the II series, it might be possible for someone
> to aquire the right to use the ROM image. I doubt it would be
> feasible to try and write clean roms, but I am not extremely
> familiar with the GS rom anyway. Then again, it may be necessary to
> create a compatible rom just to make the hardware
> work.

I'm unaware of the size/complexity of the GS rom as well. :/  Someone clue
us in?

> Some decisions are fairly simple. For instance, I think the
> mini tower case is probably the best. The current ones are
> a reasonable size and offer a decent power supply. Of course,
> it may force the motherboard design to resemble the PC due to
> the slot locations. At least this design offers the end user
> the option of selecting which case they want and if this style
> were followed they could even use the desktop style case.

woah, woah! =) I'm not talking about some peecee looking piece of crap..er..
equipment..  What I"m thinking about is something you drop into slot 3 or 4
and wire to the '816 socket, just like the TWGS.  Of course, it would
provide a full complement of ram, per se 16 megs (8 megs to emulate IIGS
ram, and 8 more for whatever: ramdisks, workspace for the emulator), and do
all of its I/O via the IIgs.

> Also, it might be better to design this as a kit type of system
> where the basic electronics are pre-assembled, motherboard and
> such, but it would still require the end user to finish the
> product. Then as a service, whoever distributes it could offer
> to pre-assemble everything. This may also help in the rom issue
> as you aren't actually offering a new computer, but a project
> which the user must supply a rom image themselves. In this case
> though a method of loading the image will be needed.

Hm.  That's definately a consideration.  I dont swear by not building a
standard-form motherboard, but in the moment of considering the feasiblility
of writing clean-room GS roms, you failed to consider the feasibility of
re-structuring the IIgs's hardware.  I find the first to be far more likely
to occur than the second.  I'd love love love to see something like this
come out, but not only for the apple2 world.  We love our apple2's (I miss
my laser128 like the desert misses the rain), but other poor saps need some
help too (ie the Atari scene, commodore, etc).  But, since we're on an
apple2 group, let's focus on the most important stuff: the apple2 =)


> Well, just a little more to think about.

You've got me thinking, friend.  Dont doubt it for a minute.

-
Jonas






I agree, this needs to be a drop in for the original case.  

Lets look at this on the cheap side of hardware development as well as
the cheap side for hardware.  I am sure there are chips that give you
quite a bit of flexibility in supporting new processors, but the cost to
develop this may be very large.  I was hoping to design this baby using
old motherboard parts laying around the house.  I know some things will
need to be new, but I have got at least ten motherboards for 386 and 486
systems.  When you start talking about supporting SDRAM and hard drives
the cost for development goes way up.  If this thing could start with
logic chips, cpus, and mayber gals for glue then the cost is much
lower.  

Remember this is a hobby and not everyone has access to storage scopes
and other necessary troubleshooting hardware.

On the software side I just thought it might be useful to provide a
monitor program for troubleshooting.  Another possibility would be a
terminal that could either use a dedicated port or one of the IIgses
ports.  Maybe it could even download new cores for development.  Again,
it could be used to support other systems like the Atari and the
Commodore by making different cores!

There is one thing that needs to happen before more things are added. 
Someone needs to get something working before to much discussion of how
hard it will be.  From what I understand of open source projects you
have to have a working program before a project will really take off. 
Anyway, it is a thought.





 I know some things will
> need to be new, but I have got at least ten motherboards for 386 and 486
> systems.  When you start talking about supporting SDRAM and hard drives
> the cost for development goes way up.  If this thing could start with
> logic chips, cpus, and mayber gals for glue then the cost is much
> lower.  

Right.  Using some complex logic array to interface between the 65816 socket
and the host CPU would be the best solution.  Perhaps something we can
program from a measly parallel card (there are a few like that, they use two
lines, one for bidirectional data, the other for unidirecitonal clock).  All
you'd have to do is manually twiddle the port lines.

While the 386 may be lying around in your place, I dont have any (donated
them all to people wanting to learn basic linux skills).  However, if you
_do_ get it working with a '386, it shouldnt be too hard to re-adapt it for
other x86 processors, since most share the same style of bus.  In the long
run, it will be cheaper to grab an old socket7 cpu and an mvpIII chipset,
and go off with it.  My K6/233mhz _and_ my mvpIII chipset motherboard costed
$110 together, and I'd honestly pay that to see my IIgs scream like win98 on
a K7/500 =)

> Remember this is a hobby and not everyone has access to storage scopes
> and other necessary troubleshooting hardware.
> 
> On the software side I just thought it might be useful to provide a
> monitor program for troubleshooting.  Another possibility would be a
> terminal that could either use a dedicated port or one of the IIgses
> ports.  Maybe it could even download new cores for development.  Again,
> it could be used to support other systems like the Atari and the
> Commodore by making different cores!

You know, that's a rather smart idea.  you'd have to remember, though, taht
the atari has a modest 6502 powering it, and the c-64 has a 6510 (=6502 with
internal 5-bit data port and DMA support on-chip).  The superCPU accelerator
also uses a 44pin PLCC form for the 65816.  While I'm not against doing
these, you'd have to provide adaptors. =)

> There is one thing that needs to happen before more things are added. 
> Someone needs to get something working before to much discussion of how
> hard it will be.  From what I understand of open source projects you
> have to have a working program before a project will really take off. 
> Anyway, it is a thought.

Well, if we'd all like to begin such a project, let's get started. =)

-Jonas





Somewhere I must have lost the train of thought on this
subject. I was thinking a complete motherboard replacement.
Mainly because most PC parts are a dime a dozen. If you need
older 386 or 486 parts, you can find them at most any PC
repair shop. A few friends I know used to have a lot.

One benefit to a new motherboard design would the ability to
use all the low cost PC hardware on the market. For instance,
you can get a case for $20 including a 200 watt power supply.

It seems to me though, if you use the raw board, most of what
you want to do can be done in software. With the proper coding,
the PC board is very adaptable. In this case, it might be
possible to use a flash rom type of system.

I remember a II clone I worked on a few years ago. It had a
6800 cpu (I think) and looked nothing like the II on the inside
but it performed pretty well.

So this brings up another question, how close to authentic do
you want this to be?

Phoenyx






I guess I am a IIgs purist.  I would still like it to look like a IIgs. 
However, if we build up a hardware and code base who is to say we do not
have ten different models to choose from?  I was not planning on
manufacturing anything more than ROM code and schematics for people to
build thier own.  If we have 10 different schematics then there will be
more people able to build them.





That's an interesting idea, which follows my thoughts about a
kit. I not real familiar with building boards, though I have
repaired a few.

Phoenyx






> subject. I was thinking a complete motherboard replacement.
> Mainly because most PC parts are a dime a dozen. If you need
> older 386 or 486 parts, you can find them at most any PC
> repair shop. A few friends I know used to have a lot.

Yeah.  Cases are cheap and all, but it's just the cost of designing a whole
new board

> One benefit to a new motherboard design would the ability to
> use all the low cost PC hardware on the market. For instance,
> you can get a case for $20 including a 200 watt power supply.

Yep yep! especially when you look at all these el-cheapo network cards with
open-source drivers hanging about. =)

> It seems to me though, if you use the raw board, most of what
> you want to do can be done in software. With the proper coding,
> the PC board is very adaptable. In this case, it might be
> possible to use a flash rom type of system.
> 
> I remember a II clone I worked on a few years ago. It had a
> 6800 cpu (I think) and looked nothing like the II on the inside
> but it performed pretty well.

I dont believe the 6800 can run 6502 code.  They ARE closely related (same
creator) though =)

> So this brings up another question, how close to authentic do
> you want this to be?

There are a few classes of people in every community.  I've learned that
especially well in the Commodore scene (my most active spot), where there
are purists that believe anything beyond joysticks, one disk drive, and a
composite monitor is unneeded and removes the commodore-ism from it.  Of
course, there are the rest, who are open-minded and would love to see the
machine grow and evolve.  I gather you to be of the second group, which is
great.  

I dont know where the apple2 community stands on modification and
advancement, but given there is a standardized, modular extensible operating
system already in place, I imagine that it's just fine.. clue me in if I"m
wrong <:-) 

-Jonas





On Sat, 5 Feb 2000, Phoenyx wrote:

> Date: Sat, 05 Feb 2000 01:20:16 -0600
> From: Phoenyx <quazar@dcnet2000.com>
> Newsgroups: comp.sys.apple2
> Subject: Re: All this talk of a new GS...
> 
> That's an interesting idea, which follows my thoughts about a
> kit. I not real familiar with building boards, though I have
> repaired a few.
> 
> Phoenyx
> 
I've found a nice on-line service, <A
HREF="http://www.expresspcb.com/";>http://www.expresspcb.com/</A> 

These guys will give you a free CAD to help you lay out a PCB, and can even
give you good estimates on what it would cost to build the board.  I've used
it to design a few projects, and have seen most two-board prototype orders
around $90.  There is a (low) $70 setup fee, and some additional charges per
square inch, as well as a per-board setup.

They offer (at larger initial setup cost) fancies like the green coating and
lettering layers on top, but since they're only needed for production, it
wont be needed for those who dont need the fancy. =)

Have a look.  The software is for win95, and works ok here.  I has a few
quirks, which yelling easily solves (grin).


-Jonas









Does anyone know how the Transwarp accelerators work?
I get the idea that there is a microkernel that runs and does fetching
of instructions into the cache then also has to provide a system clock
to the IIgs and other housekeeping signals.  Am I off the mark here?  If
it is easier to build a 65816 accelerator with the existing 20Mhz chips
then this might be the way to go.  Any thoughts?




It might be more feasible to do this.  However, these 20mhz chips dont
exist. =)  My 20mhz (from factory) accelerator for my C-64 uses a 14mhz
WD65c816.  It's overclocked from the start.

I'm unaware of the transwarp's exact methodology, but I believe you're
close.

-Jonas




Overclocked only means if you it can run at a faster speed than it is
rated.  Intel has on occasion taken faster chips and marked them as
slower to meet volumes for lower end chips.  The 14Mhz chips may just be
the cream of the crop or it may be standard for them to be able to run
at 20Mhz.  If it is the latter, I would say the manufacturer is being
very conservative on the chip ratings.




Most OEM's are conservative on their ratings, especially Intel.
It helps to build a good reputation. A while back, I remember
reading a spec from Intel which stated they test for the most
reliable speed and then reduce by 10%. At least I think that
was the figure they used. I don't think they are doing this any
more except maybe with the Celeron. All the P2 cpu's I have
worked with won't tolerate over clocking. The Celeron though
will, but if it's highly over clocked the system will report an
incorrect speed. For example, I over clocked a Celeron 300 to
the max the board would allow. The bios claimed 450 mhz, but
Linux reported a more accurate speed of 380. This was just a
test though, to prove a point to another fellow who swore his
system was running at 450. I normally discourage over clocking
a system unless there is no other way to speed things up. Some
times, this 'enhancement' can lead to serious and hard to find
system problems.

I have found that most of the older cpu's were more tolerant of
over clocking. I expect that this is mainly due to the extra
heat and possibly timing. As example of a heat issue. My AMD
500 system was a little flaky when I first put it together. It
would give random failures and core dumps in Linux. When the
cpu fan gave out, it trashed the data on both hd's. Well, I
replaced the fan with a larger cooling system and have had no
trouble with this at all.

Every time I work on a PC, I realize why I love the Apple 2!

Phoenyx






Hi Jonas,

I'm all for the system growing, but I would prefer it to keep
the basic flavor using the GS Toolbox and GSOS. Unfortunately,
these are not my specialties so at best I can offer ideas and
hopefully motivation.

Phoenyx





Of course!  I'd love to keep the GSness (Macness? <duck>), because that's
what makes the GS the GS.  Since the OS is there and it's modular, you only
have to boot from a device supported by the ROMs =)

-Jonas






On Sun, 30 Jan 2000, Frank Carney wrote:

> Subject: Re: All this talk of a new GS...
> 
> Jonas Bevren wrote:
> > 
> > If you look closely, you'll see that a K6II-400 is cheaper than you'll find
> > most embedded 386s and 486s these days.  PII-400's sell for $100 at a
> > certain irritating auction site <grin> quite often.  Cyrix cpu's, with
their
> > quite superior integer math machine, are almost a dime a dozen =)
> 
> I was talking about building a non-IBM compatible system that connects
> directly to the IIgs CPU socket.  I was not talking about the embedded
> systems that can run Win95 and can support PC-104 cards.  I was thinking
> of building a system that runs a core with none of the crap that is in
> the PC now.  And yes, if this project were to mature to bigger faster
> processors such as the 400Mhz machines that would be great.  I guess
> when people think of "embedded" they think of an embedded PC.  I want
> this thing to fit in the original IIgs case, but not be a full PC.
> 

Ah, but you misunderstand me.  I talk of the chip itself.  You can get even
modest (but relatively fast) 233mhz pentium-class CPUs for pennies these
days.  If run at lower frequencies, say 100mhz, they can likely be operated
without fan, and possibly even without heatsink (the mediagx-233 cpu can run
at 166mhz without any ventilation or heatsink in fact from what I hear).

Right. I agree that using a full peecee wouldnt suit the bill.  I'd never
wish taht on an apple2 lover (as I"m a commodore lover and dislike any
similar fate for my beloved =).  No winCE core, no win95, linux, or
anything.  Just a raw 100% machine language opcode interpreter running
rampant on the CPU socket to make GS/OS look like win3.11 on a PIII <grin>.

Hmz.. perhaps even operating like the commodore's superCPU, which houses,
powers, and accesses only local ram, and mirrors only what's needed
(selected by the user) to ram to be sure the video display gets updated at
regular interfals.  The only other timing constraints are the I/O devices,
which cause a 20-cycle delay (thus increasing access time for the needed 1mhz
clock).  Put in a small core for running the emulator, which should be small
enough to fit as much as possible in the little on-chip CPU cache, and run
all GS/OS apps from the accelerator's internal RAM.  This leaves the CPU
free of delays and worries, and only needs to slow down for slow devices (ie
the DOC chip, writing to the video display's I/O registers, etc).  In fact,
toss on an OS friendly localbus harddisk and your necktie would be flying
behind you.

Hm.  Maybe I should be in marketing. (duck, cringe)

-Jonas





Exactly, I was thinking of starting small (386 small) and then moving
up.  Who knows eventually it could support socket 7 or even super socket
7 chips and really scream.  I would not be worried about a heat sink and
fan if it got to that point.  Though with the faster chips I would see a
lot of hardware interfacing the lower voltages of the chips to a the 5
volt logic of the IIgs.  This is why I was thinking of starting small. 
Even a lot of the 486s were still at 5 volts up to the 66 I believe.





486's go to 80mhz at 5v.  Many of the AMD 486's, including my 133mhz (which
I have been running at 160mhz for 4 years now), are 5v-tolerant, meaning
they can be hooked directly to a 5v input.

Let's consider something (you will likely balk at it at the start, but be
patient).  I know of an individual that is trying to get PCI, AGP, and USB
into the Amiga series computers from Commodore.  Under my advice, he decided
to use a PC chipset, specifically teh Apollo MVPIII.  He'll be making glue
logic to make the expansion slot on the amiga appear to be the host
processor. =)

We could do quite the reverse on the faster-IIGS project.  In fact, I'd
recommend using a chipset between the host CPU and the apple, since a couple
chips will provide full function including SDRAM control, PCI expansion
(fancy this), etc etc.  The IIgs's internal memory, or what of it needs
really be used, can be mapped through a PCI "slot" and mapped into high
memory, allowing a GS/OS clean-room core to run in native mode on the host
CPU.  GS/OS apps, of course, will operate on an emulated 65816 core, but
calls to the tool box will escape the emulator and run the toolbox call in
native mode, rendering a lightning-fast display, disk, and other operating
system features.

It would also become quite possible to address the full 14M ram mentioned
earlier, and toolbox permitting, a 32-bit address could be specified by
'savvy' programs to enable huge amounts of memory (the "ACE" operating
system for the commodore 64 will support a 32-bit 'far memory' address for
data workspace.. quite handy).

While this sounds like installing a whole PC inside the IIGS, you wont have
to deal with the large box, noisy fan, etc.  

Another alternative is the MediaGX I mentioned earlier.  In two chips, the
processor and a companion chip specifically designed for it, there is 16bit
aduio in/out, SVGA video, 2 IDE controllers, a PCI interface, an ISA bridge,
joystick input, two serial ports, a parallel port, ps/2 mouse, and keyboard. 
While we wouldnt necessarily use all of these ports, it's quite simple to
add a glue interface between a PCI 'slot' on the mediagx cpu (the companion
isnt needed for PCI, ram, video or video) for direct access to the IIGS's
internal devices.  I'm not sure of pricing or availability for these CPUs,
but I do know they're sufficiently fast.  I run linux on one, and it's quite
a breeze to operate.  The chip appears to be a BGA device (= no legs.. just
solder pads, you literally bake it onto the PCB), so actually building a
board for it can be a challenge.  More later. =) comments?

-Jonas






Well, I suppose the next question is, what will it take to
make the board and who is capable of doing so?

Phoenyx





Er, uh... um, uh... =)

Sometimes I really wish I had cash for my hardware projects (I'm quite lucky
to be able to afford the ones I do).  I'm no expert on PC hardware tho.

Hooking it all up shouldnt be that large a deal, asice from the difference
in bus widths (which can be compensated for in software).  Hm.  I'm guessing
we'd go in through the 65816's port, since most other accelerators do that
anyhow <grin>.  From there, it's a matter of building a 'proper'
pci->65816-socket interface.  It would be rather nice to be able to have it
friendly to the PCI standard, even though this _is_ an embedded project.

It's not difficult to map a 16M address space through a PCI slot either,
especially when you are running a custom embedded system.  If we bypass all
the standards I'd like to uphold, we could easily just decode the top 8 bits
of the 32-bit address bus on the core CPU, and use that to enable/disable
a set of transcievers on the 65816's socket.  Ya know, if you dont mind
sacrificing some speed, you could have the firmware just copy the GS's rom
data into its local ram and run everything behind an emulated 65816.  I'm
pondering how difficult it would be to emulate the CPU, and I dont think it
would be hard at all, especially considering the (relatively) large number
of registers in the beast.  A few could be dedicated for the 816s own
registers, and the remaining could be used for the emulation engine to
load and decode opcodes. 

Easiest emulation IMHO would be two brute-force interpreters, one for native
mode, and another for emulation mode.  I chose two because each interpreter
cant be all that large, and it makes it significantly faster if the code
doesnt have to check whether its in emulation mode or not on every opcode.

Simple example for brute force interpretation (in a friendly language):

X/A are 16bit of course
coreloop lda [IP] ; fetch opcode from memory
         asl      ; multiply by two for vectored jump
         tax     
         jmp [dispatch,x]
         bcc coreloop
; switch modes here

anyhow, you get the idea =)  256 routines in memory for 256 entries. 
Returning a set carry means its time to switch to the other mode.

Ok, enough babble.  Let's get back to hardware. =)

-Jonas
 
         


I think at this point, adhering to as many standards as possible is a
must. Part of the problems the early Apple2 faced
was because there were no standards which left developers in
the position of making their own. In the floppy driven world of the A2,
this wasn't too bad, but when the system became able to
access hard drives, it became more of an issue.

There are still other issues to deal with. Perhaps the biggest
one is the GS rom. Apple may ignore the emulators, but if a
hardware project were developed, they may decide it's time to
examine the copyright issues. Since Apple doesn't care about
any growth on the II series, it might be possible for someone
to aquire the right to use the ROM image. I doubt it would be
feasible to try and write clean roms, but I am not extremely
familiar with the GS rom anyway. Then again, it may be necessary to
create a compatible rom just to make the hardware
work.

Some decisions are fairly simple. For instance, I think the
mini tower case is probably the best. The current ones are
a reasonable size and offer a decent power supply. Of course,
it may force the motherboard design to resemble the PC due to
the slot locations. At least this design offers the end user
the option of selecting which case they want and if this style
were followed they could even use the desktop style case.

Also, it might be better to design this as a kit type of system
where the basic electronics are pre-assembled, motherboard and
such, but it would still require the end user to finish the
product. Then as a service, whoever distributes it could offer
to pre-assemble everything. This may also help in the rom issue
as you aren't actually offering a new computer, but a project
which the user must supply a rom image themselves. In this case
though a method of loading the image will be needed.

Well, just a little more to think about.

Phoenyx