Overview Features Coding ApolloOS Performance Forum Downloads Products Order Contact

Welcome to the Apollo Forum

This forum is for people interested in the APOLLO CPU.
Please read the forum usage manual.
Please visit our Apollo-Discord Server for support.



All TopicsNewsPerformanceGamesDemosApolloVampireAROSWorkbenchATARIReleases
Information about the Apollo CPU and FPU.

Firebird Replacing 68010 On Non Amiga HW

Michael Wagner

Posts 1
08 Jun 2023 23:34


Can I use Firebird to replace a 68010 on a non-Amiga Mainboard?
Of course I will have to write the software to use it myself.
But my question is: is there anything about the Firebird that is strictly Amiga specific and prevents its use on non-Amiga boards,
like a VME-Board?


Gunnar von Boehn
(Apollo Team Member)
Posts 6193
09 Jun 2023 07:44


Michael Wagner wrote:

  Can I use Firebird to replace a 68010 on a non-Amiga Mainboard?
  Of course I will have to write the software to use it myself.
  But my question is: is there anything about the Firebird that is strictly Amiga specific and prevents its use on non-Amiga boards,
  like a VME-Board?
 

 
As you might know, Commodore Amiga did plan on developing a new much improved chipset for Amiga, having these features:
  - improved Sprites
  - 32bit Copper
  - improved Blitter
  - improved Amiga 16bit Audio channels
  - true color modes
  - 3D acceleration
 
 
The Firebird finally brings to you what Commodore always wanted Amiga to have. It allows you to upgrade your A500 to the new Super-AGA chipset that Commodore planned to developed but unfortunately failed to bring out.
 
Yes the Firebird is designed for the Amiga specifically.
 
The Firebird packs into one card the Amiga chipset enhancements and  the world most advanced 68K CPU, the 68080 - and includes ApolloRom a new Amiga Kickstart. You can not easily use it outside Amiga.




Antony Coello

Posts 153
11 Jun 2023 12:08


It depends on what signals your target system uses from the CPU.

i.e. Theres been some attempts to use the V2 boards with an Atari ST system, but some of the full 68000 implementation is missing.

Specifically, TRAPs are not handled (amongst other things).


Gunnar von Boehn
(Apollo Team Member)
Posts 6193
11 Jun 2023 15:43


Antony Coello wrote:

    It depends on what signals your target system uses from the CPU.
   
    i.e. Theres been some attempts to use the V2 boards with an Atari ST system, but some of the full 68000 implementation is missing.
   
    Specifically, TRAPs are not handled (amongst other things).
   

   
Let me correct what you say:  TRAPS are fully 100% working on Apollo 68080.
But maybe you mean something different and maybe can explain what you mean?
 

But the point is the Vampire cards are "not naked" CPUs
The card include a lot of Amiga "chipset" and they come with a lot of Amiga "knowledge".
The cards know which areas in Amiga are ROMs and overmap their own Rom there.
The cards know which areas in Amiga are IO-areas, and handle it automatically correctly (never cache this areas, never access them out of order).
And the cards know which areas are pure memory and know which tunings they are allowed to enable for these memory regions.
 
This means the cards are designed with Amiga in mind.



Antony Coello

Posts 153
11 Jun 2023 22:34


Gunnar von Boehn wrote:

 
Antony Coello wrote:

      It depends on what signals your target system uses from the CPU.
     
      i.e. Theres been some attempts to use the V2 boards with an Atari ST system, but some of the full 68000 implementation is missing.
     
      Specifically, TRAPs are not handled (amongst other things).
     

     
  Let me correct what you say:  TRAPS are fully 100% working on Apollo 68080.
  But maybe you mean something different and maybe can explain what you mean?
   
 
  But the point is the Vampire cards are "not naked" CPUs
  The card include a lot of Amiga "chipset" and they come with a lot of Amiga "knowledge".
  The cards know which areas in Amiga are ROMs and overmap their own Rom there.
  The cards know which areas in Amiga are IO-areas, and handle it automatically correctly (never cache this areas, never access them out of order).
  And the cards know which areas are pure memory and know which tunings they are allowed to enable for these memory regions.
   
  This means the cards are designed with Amiga in mind.
 
 

 
  Sorry if I am mistaken. I believe it was at least some sort of exeption handling that was one of the reasons for not being able to work on Atari. Perhaps you can elaborate on what the core does NOT do which a 68000 does and therefore prevents a V2 card working in an Atari ST?


Gunnar von Boehn
(Apollo Team Member)
Posts 6193
12 Jun 2023 12:43


Antony Coello wrote:

    ... and therefore prevents a V2 card working in an Atari ST?
   

     

As I tried to explain: 
The Vampire cards are "not just naked" CPUs.
The card in fact include some Amiga "chipset" and they come with a lot of Amiga "knowledge".
   

For example:

The cards know which areas in Amiga are ROMs and will overmap their own Rom there.

   
The cards know which areas in Amiga are IO-areas, and handle it automatically correctly (never cache this areas, never access them out of order).
   
And the cards know which areas are pure memory and know which tunings they are allowed to enable for these memory regions.  (allowed to prefetch, allowed to access out of order)
 
 
The Vampire card will do some system checking on poweron, including memory tests on Amiga memory, and also do some Amiga chipset test, and even check the checksum of the Amiga Rom before continue.
   
If you think about this then this should easy to see that you can not just plug the card into another Computer and assume it will run.
As the cards are designed for a Amiga memory layout and will actively check this Amiga and "complain" if they not find an Amiga.

Of course in theory this "Amigaawareness" could be disabled
and it would needed to build a non Amiga-aware card to run in Atari.


Gunnar von Boehn
(Apollo Team Member)
Posts 6193
16 Jun 2023 17:48


Gunnar von Boehn wrote:

Antony Coello wrote:

    Specifically, TRAPs are not handled (amongst other things).
   

   
  Let me correct what you say:  TRAPS are fully 100% working on Apollo 68080.

Let me give you more information here.
ATARI OS and ATARI programs use TRAP instruction a lot.
Its common coding practice for Programs to start with TRAP #1 and end with TRAP #1 instructions.

This all works just perfectly on Vampires.
As you might know - you can run ATARI OS / and EMUTOS and can run ATARI program perfectly on Vampire.

Maybe you even know that a number of high profile Atari coders use the V4 as the ATARI development of the choice.

Some of Arne new games like "APOLLO INVADER" have even ATARI versions - and be played on ATARI OS - on the Vampire.

I think this clearly shows that there is absolutely no instruction problem, right?


Antony Coello

Posts 153
18 Jun 2023 17:30


Ive seen this great initiative. A lot of work has gone into this.

However, this is a form of hybrid Atari emulation on an Amiga platform.

Ive always wondered how much work it would be to strip all the Amiga specific bootstrap/startup mapping from the Vampire card and provide bus error support (that the Amiga doesnt use but the Atari needs) to have a more basic 68080 CPU that can slot straight in to Atari/Apple/Sun/etc machines.

Even having it hardwired as a superfast 68000 (or 010/020/030/040/060) for full plug n play compatibility would save a majority (if not all, depending on if theres any badly coded instruction timing dependant code) OS patching on the target systems. You could then have 6 types of card to cover all the 68k generations.

Im a little surprised this hasnt been looked into more, as it would increase target audience/sales somewhat I imagine.

The original poster was asking about being able to plug the card into a different 68xxx architecture/system after all.


Antony Coello

Posts 153
18 Jun 2023 17:31


Gunnar von Boehn wrote:

  Of course in theory this "Amigaawareness" could be disabled
  and it would needed to build a non Amiga-aware card to run in Atari.

This is what Im talking about! :)


Gunnar von Boehn
(Apollo Team Member)
Posts 6193
19 Jun 2023 10:44


Antony Coello wrote:

Im a little surprised this hasnt been looked into more, as it would increase target audience/sales somewhat I imagine.
 
The original poster was asking about being able to plug the card into a different 68xxx architecture/system after all.

As you know our goal is to revive Amiga.
You for sure understand that this is a big job.
We produce several different cards as accelerators for old Amigas - each model designed to fit in a typical Amiga model.
Not all Amigas are build the same - and a single card can never fit in all models.

Regarding "increase" of sales. We have over 10,000 happy Amiga customers.
What do you think about what volume do we talk here?




Antony Coello

Posts 153
19 Jun 2023 17:49


Gunnar I fully appreciate that you have your hands full and I would not want you to take resources from what you are already doing (i.e. perfecting the Vampire cards on the Amiga) to then thin them out trying to cater for every 68k based machine in existence.

Its more of a whats theoretically and currently possible type of topic up till now.

As for increase in sales. Well thats up to you. :) Ive no idea when Amiga based sales will start to plateau/reach saturation level. Are you happy to leave things there when that happens, or is there another future target that could be a possibility (like discussed in my post above)?


Gunnar von Boehn
(Apollo Team Member)
Posts 6193
20 Jun 2023 07:55


Antony Coello wrote:

or is there another future target that could be a possibility (like discussed in my post above)?

Anything is possible.

Today already some of best ATARI developers are active V4SA users.
As far as I know they are very happy with the features the V4SA does offer them.




Olivier Landemarre

Posts 147
22 Jun 2023 20:27


Antony Coello wrote:

Gunnar I fully appreciate that you have your hands full and I would not want you to take resources from what you are already doing (i.e. perfecting the Vampire cards on the Amiga) to then thin them out trying to cater for every 68k based machine in existence.
 
  Its more of a whats theoretically and currently possible type of topic up till now.
 
  As for increase in sales. Well thats up to you. :) Ive no idea when Amiga based sales will start to plateau/reach saturation level. Are you happy to leave things there when that happens, or is there another future target that could be a possibility (like discussed in my post above)?

I think Atari market is small today

I don't think there is a lot of ST user that wan't upgrade there machine like it is done on Amiga, and Falcon more adapted for this has already CT60, TT could be perhaps better target but there is low user on this computer so huge for near nobody.
Just my point of view.




Olivier Landemarre

Posts 147
22 Jun 2023 20:30


Gunnar von Boehn wrote:

Antony Coello wrote:

  or is there another future target that could be a possibility (like discussed in my post above)?
 

 
  Anything is possible.
 
  Today already some of best ATARI developers are active V4SA users.
  As far as I know they are very happy with the features the V4SA does offer them.
 
 
 

Yes I confirm for me nice small computer for Atari clone, of course it could be a bit more compatible could be nice for have more large diffusion, but with time perhaps.
The most exciting computer I have seen for long time



Brett Eden

Posts 15
09 Feb 2024 11:02


Olivier Landemarre wrote:

Gunnar von Boehn wrote:

 
Antony Coello wrote:

  or is there another future target that could be a possibility (like discussed in my post above)?
 

 
  Anything is possible.
 
  Today already some of best ATARI developers are active V4SA users.
  As far as I know they are very happy with the features the V4SA does offer them.
 
 
 
 

 
  Yes I confirm for me nice small computer for Atari clone, of course it could be a bit more compatible could be nice for have more large diffusion, but with time perhaps.
  The most exciting computer I have seen for long time
 

This is absolutely true -- the Atari ST community is very small and the attempts to keep the platform alive are nowhere near the efforts seen with the Amiga.


Brett Eden

Posts 15
09 Feb 2024 11:12


Antony Coello wrote:

Ive seen this great initiative. A lot of work has gone into this.
 
  However, this is a form of hybrid Atari emulation on an Amiga platform.
 
  Ive always wondered how much work it would be to strip all the Amiga specific bootstrap/startup mapping from the Vampire card and provide bus error support (that the Amiga doesnt use but the Atari needs) to have a more basic 68080 CPU that can slot straight in to Atari/Apple/Sun/etc machines.
 
  Even having it hardwired as a superfast 68000 (or 010/020/030/040/060) for full plug n play compatibility would save a majority (if not all, depending on if theres any badly coded instruction timing dependant code) OS patching on the target systems. You could then have 6 types of card to cover all the 68k generations.
 
  Im a little surprised this hasnt been looked into more, as it would increase target audience/sales somewhat I imagine.
 
  The original poster was asking about being able to plug the card into a different 68xxx architecture/system after all.

Hi Antony,

Using a trimmed-down V4 to upgrade a 68K Mac is an intriguing idea, however I think you'd find that emulating one on a fast x86 CPU (using ShapeShifter/SheepShaver etc) would be both faster and much cheaper. 

Of course, you'd miss out on the feeling of the Apple hardware, which is why people continue to upgrade/maintain their Amiga instead of just running WinUAE, which is many, many times faster than an Apollo V4 when run on a present-day Intel i9.



Antony Coello

Posts 153
09 Feb 2024 15:33


FWIW Brett, I believe every individual person who wants to use retro machines has their own 'psychological line'. That point where they will not cross; a step too far if you will, where once crossed, the whole thing ceases to become retro computer HW and crosses into emulation. (Others still, however dont care about any of the original 'real' HW and just emulate).
 
  My own interpretation of this is using my A500+ board (housed in a Checkmate 1500 case) running a Vampire V500 card, so I still have all the legacy electronics. In my case however, I do still have some old HW pheripherals which cant be attached to a V4SA.
 
  I also believe an AC68080 is a 100% valid CPU core and the Vampire system as a whole a very well thought out enhanced/augmented solution for an Amiga upgrade  path.
 
  BTW I also like the C64DTV for several of the same reasons! ;)


Brett Eden

Posts 15
10 Feb 2024 06:10


Antony Coello wrote:

  FWIW Brett, I believe every individual person who wants to use retro machines has their own 'psychological line'. That point where they will not cross; a step too far if you will, where once crossed, the whole thing ceases to become retro computer HW and crosses into emulation. (Others still, however dont care about any of the original 'real' HW and just emulate).
   
    My own interpretation of this is using my A500+ board (housed in a Checkmate 1500 case) running a Vampire V500 card, so I still have all the legacy electronics. In my case however, I do still have some old HW pheripherals which cant be attached to a V4SA.
   
    I also believe an AC68080 is a 100% valid CPU core and the Vampire system as a whole a very well thought out enhanced/augmented solution for an Amiga upgrade  path.
   
    BTW I also like the C64DTV for several of the same reasons! ;)
 

 
  Hi Antony,
 
  Oh yes, I agree completely about that proverbial psychological line that everyone has when it comes to true retro vs. emulation.
 
  Personally, I am something of a Commodore purist, and initially
  wrestled with the idea of going down the Vampire path with my A600 and A1200.  I was worried that the FPGA nature of Vampire would take away much of the "Commodoreness'.  I know now this is not the case, because running AmigaOS 3.2.2 on Vampire is just a very fast Amiga, with nice extras you couldn't have otherwise (RTG, etc).  Even though it is slower than a uh, you know, that Raspberry Pi thing...in raw MIPS, it is more compatible and doesn't feel like a hack, which PiThingy does.
 
  The idea of running AROS didn't appeal to me, as much as it is a very attractive-looking OS.  Sadly, one cannot
  even determine where AROS is in its journey to achieving AmigaOS 3.1 parity -- their roadmap/status page is very outdated. 
 
  Now you have Hyperion AmigaOS 3.2.2, which I feel is moving in the direction Commodore would have taken AmigaOS 3.1 if given the chance.  The question remains though, how does this affect AROS?  I believe Apollo is making the right choice by providing ApolloOS and Coffin, because of its open-source nature.  I guess ultimately it matters not since you can run AmigaOS and ApolloOS side by side.

The sad thing about Hyperion's hold on AmigaOS is that it can't benefit from the Apollo MMX instructions at the OS level...if they supported them, that very old AmigaOS 3.x codebase could get a real boost.
 
  I commend Apollo for their work on the 68080 and the improvements they have made to the GCC compiler.  This is necessary to keep the platform alive.

posts 18