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.

Is the 68080 Soft Core Available On Its Own?page  1 2 3 

Gunnar von Boehn
(Apollo Team Member)
Posts 6197
12 Apr 2020 09:06


gsteemso del Canuckistan wrote:

    - If a bare '080 were readily available for purchase, people would purchase it.
   

 
   
Modern chips need different voltage than old chips.
This means the same way you can NOT just plug in a new INTEL-CPU into an old 486 mainboard.
you could not just plug in a new 68080 into an old 68K mainboard.
You will need a new mainboard too.
 
The Vampire includes all this what you need.
If you want to "tinker" an 68080 to an old System then the simplest way forward is to make an adapter to use a Vampire card.

If I understand you right then what you ask for is a simple to use 68080 CPU and this is already available - its the Vampire.



Vojin Vidanovic
(Needs Verification)
Posts 1916/ 1
12 Apr 2020 12:21


BigGun is right guy to answer. Like you, I see it as "Catch 22" situation

- Team is small and cannot do faster then they do, nevermind our desires our even cash flux. Human limits, plus overall situations with components, prices, capitalism instability as usual.

- Team would like to license tech to "a big partner" but that is not what was 2017 idea - small fee licensing to any similar FPGA at half speed that could be unlocked to full, or completely free design that is paid by end user to fully unlock (I dont get which one of two was the overall idea). So seems original idea is kind of dead. Vampire v1 is fully open by Majsta but that is not AC68080 core but somewhat improved TG68, if I get that right. And that is current really "open sourced" limit

- There are no physical CPUs. Since FPGA tech needs a lot of tech knowledge and constant improvements to make it operate really well, I understand desire not to just license technology, but a real partner that would "handle all problems and constatly work on core" is needed. This is way closer to what I define as "strategic partner". However, there is none on horizon. Simply, our tech is not (yet) globally attractive, even it comes close (or better) to early ARM CPUs that used to drive e.g. Finnish Nokia great phones and pre-Android Symbian generation and early handheld palm devices.

Team does this (support and development) to a certain degree, and pro partner or any kind of strenghtening here is vital. But needs "as good people" as already in team. And such people would require high salaries, which team and current "Vamp economy" cannot provide. Not everyone is Amiga enthusiast (or retro Atari-Mac).

- Mass Production would be great and killing both price, raising availability, finally supplying dealers and making it closer to a real product. Team tries its best with Igor knocking himself down and V4sa being produced best it could be done in small scale economy. But again, here we ask "for bigger brother".

I dont know how much are these two points, but they are really for some bigger company with a vision. We dont have that in Amiga land, we see only LCCs revamped to steal some glory (CUSA, Phase 5, AEON to some extent) and cashflow.

Hope it will be one day, but I dont expect it quick. If I had few million $$$ making physical ITX-ATX Vamp board with high clocked 080-ammx, PCI slots, memory slots and warp3d compatibile 3D chip onboard (be it old 3D Virge, Parmedia, CirrusLogic or old Radeons) would be best thing after "Boxer failure" and certainly better investment then Varisys made boards. But its a long road to go, and these are mature, but "baby steps".

For development, providing non-ASM compilers and other forms of coders support and documentation seems to be biggest obstacle to having more Vamp related software. As well as same difficulty to get cards to develop and test on. However, this can still be done, and for coders I hope will get some "attention".

Recent Aminet answer of only "20 plus pieces of software" which are Vamp related or supporting shows it best. In several years of existence period.

Seems coding for 040-060 and classic Amiga chipset (well documented so far) is best we can expect - like e.g. Arti does most of his ports for overall m68k market, but its high hardware demands for Amigas, makes it fly on Vamp.

So in one sentence - "being fastest Amiga" is currently way more selling point then introduced new instructions, AMMX, SAGA no matter how great they are. I mean, we have 16-bit sound on v4SA and dont have AHI-MHI driver yet, as plastic example.


Gsteemso Del Canuckistan

Posts 8
14 Apr 2020 22:36


... Wow, I am _incompetent_ at being clear, concise & brief.

I reiterate:

••• What steps are needed to make a bare "68080" processor available to the public? •••

By "bare" I mean _just the processor_.

A bare CPU is not an "upgrade". It is not directly useable by an end user. At all.

A bare CPU is a _component_, like a capacitor or an SRAM chip.

To buy a bare 68040 on eBay is -- and I just checked this a couple of minutes ago -- roughly $40-45 American.

To buy a bare 68060 on eBay -- I have likewise just now checked -- roughly $85-145 in U.S. money.

Neither one of those can be "plugged into an existing computer" in the sense you keep discussing.

For the third time: THAT'S NOT WHAT I'M SEEKING.

The posted prices on the Vampire line range between €300-590, not including shipping and who-even-knows-what taxes / customs fees / etc. That works out to roughly $330-650 in U.S. funds.

I suspect the market would willingly pay about... I don't know, maybe €180-US$200? ...for a bare '080. That's roughly 40% of what a full Vampire unit costs.

Please imagine, if you will, that I had the only supply of MC68040RC40 parts in the world -- but I refused to discuss selling one to anybody unless they also purchased an entire product built around it, at roughly 2-3 times the cost of the bare microchip, and the product I was forcing all of _my_ customers to buy was not terribly relevant to any of _their_ customers. If you were attempting to arrange such a purchase, would you be irritated? I suspect you would.

I am completely confident that everyone here is acting with honest good will. However, it looks like I have failed to convey that Gunnar and I are working towards completely different goals.

With fading optimism for having conveyed my meaning rather than merely my words,

Gordon S.


Gunnar von Boehn
(Apollo Team Member)
Posts 6197
14 Apr 2020 23:08


gsteemso del Canuckistan wrote:

By "bare" I mean _just the processor_.

     
Yes selling  the processor would be nice.
We do not have a CPU chip fab so this is not that easy.
And a naked Processor would also be totally unusable to people, as they also need a PCB to hold it.
   
Selling the CPU as FPGA module will need quite a lot PCB and Chips around it to make "useable".
The Vampire comes very close to a good CPU module that people can actually use.
 
If you want to put the 68080 CPU into an old system then you will need to put some "glue" and "support" logic around anyway - and this is what the vampire provides. I think that you can not reasonable build your own card in small number cheaper than buying a Vampire.
 
If you want to mass produce some card for some purpose, then you should directly talk to us.


Alan Haynes

Posts 140
15 Apr 2020 12:42


To make a bare 68080 CPU you would require the following:
Pay the Team enough to make it worth their while.
The completion of testing and upgrading the Vampire Core over the next few years. This will cost  you a lot of money but if you have it then no problem.
Once the core is finalised in the FPGA you will then need to find a Chip Manufacturer. I believe there are several around the world.
You then negotiate with the chip manufacturer the number of chips you require and the form factor of the ASIC. Remember the form factor is important as it determines what device it can be plugged into. Pay the money and remember the more chips you want the cheaper it will be. But if you only want one chip it will probably cost you several hundred thousand dollars.
Wait a few weeks and the chip manufacturer should deliver the chip or chips to you and you are free to do what you want with it or them.
Also remember that for every chip produced a royalty is paid to Gunnar, Igor and the rest of the team as they own the patent and rights to the 68080 core. Perhaps you can buy them out which would be really good for them so they can retire and put their feet up.
Wait a minute your real name is not Elon Musk is it?

If you have unlimited funds at your disposal then I would say you could bring this plan to fruition in about 3 years give or take 6 months.
There you have what is needed to get a 68080 Chip. A bare processor, a CPU.

Cheers and I look forward to the announcement from Gunnar, Igor and the team about what most of us have been hoping for for 25 years. A real saviour for the Amiga. Thank you.


Ian Parsons

Posts 230
15 Apr 2020 17:49


Do bare CPUs make much sense or would it be better to go for an SoC?


Gunnar von Boehn
(Apollo Team Member)
Posts 6197
15 Apr 2020 19:12


Ian Parsons wrote:

  Do bare CPUs make much sense or would it be better to go for an SoC?
 

You are absolutely right.
A SOC is a good decision:
CPU with memory controller, with NIC, with USB, with SAGA all in one chip.


Alan Haynes

Posts 140
16 Apr 2020 03:36


Ian Parsons wrote:

Do bare CPUs make much sense or would it be better to go for an SoC?

Ian, No and Gunnar is right but this nice wealthy person only wants the bare CPU.


Mercury Thirteen

Posts 9
20 Apr 2020 23:18


gsteemso del Canuckistan wrote:

... Wow, I am _incompetent_ at being clear, concise & brief.
 
  I reiterate:
 
  ••• What steps are needed to make a bare "68080" processor available to the public? •••
 
  By "bare" I mean _just the processor_.
 
  A bare CPU is not an "upgrade". It is not directly useable by an end user. At all.
 
  A bare CPU is a _component_, like a capacitor or an SRAM chip.
 
  To buy a bare 68040 on eBay is -- and I just checked this a couple of minutes ago -- roughly $40-45 American.
 
  To buy a bare 68060 on eBay -- I have likewise just now checked -- roughly $85-145 in U.S. money.
 
  Neither one of those can be "plugged into an existing computer" in the sense you keep discussing.
 
  For the third time: THAT'S NOT WHAT I'M SEEKING.
 
  The posted prices on the Vampire line range between €300-590, not including shipping and who-even-knows-what taxes / customs fees / etc. That works out to roughly $330-650 in U.S. funds.
 
  I suspect the market would willingly pay about... I don't know, maybe €180-US$200? ...for a bare '080. That's roughly 40% of what a full Vampire unit costs.
 
  Please imagine, if you will, that I had the only supply of MC68040RC40 parts in the world -- but I refused to discuss selling one to anybody unless they also purchased an entire product built around it, at roughly 2-3 times the cost of the bare microchip, and the product I was forcing all of _my_ customers to buy was not terribly relevant to any of _their_ customers. If you were attempting to arrange such a purchase, would you be irritated? I suspect you would.
 
  I am completely confident that everyone here is acting with honest good will. However, it looks like I have failed to convey that Gunnar and I are working towards completely different goals.
 
  With fading optimism for having conveyed my meaning rather than merely my words,
 
  Gordon S.

I hear you loud and clear here; I, too, inquired about this possibility in the past in the "Any Use for 68080 Outside the Retro Community?" thread.
 
  The Apollo Core 68080 would make an _excellent_ base for a homebrew computer or any number of other projects. To deliver such a thing, the Apollo team would have to simply make their FPGA of choice available for purchase, preloaded with the Apollo Core. That way one could buy a pre-configured FPGA-based processor from them the way one would order an existing 68K CPU from eBay or other online retailer today. This would not involve the end buyer gaining any kind of access to the internal code of the processor, and would also not need to include any kind of controllers - memory, video or otherwise. This "stand-alone" version would basically only need a data bus and an address bus along with whatever few control signals would be necessary, similar to an 68000. All the magic that happens inside, stays inside, and the CPU would simply output addresses to the address bus, leaving it up to the buyer to fashion their own memory controllers or what-have-you which would be selected/designed specific for their needs and leave no additional burden on the Apollo team.
 
  I must say here that it is not my intention to fault or blame the team for not being open to this idea. This core is their invention and they have every right to market it as they see fit.
 
  However, for my own purposes, I do hope that someday an Apollo processor will become available for sale without all the Amiga-specific pieces.
 
  Overall, though, the team has done an awesome job breathing new like into the 68K line, and that alone is exciting to see. :)


Alan Haynes

Posts 140
21 Apr 2020 04:35


If you just want the bare cpu just buy the Altera V5 online. It is an FPGA and you can buy them from whomever is selling them. I think Intel own them so just look up their website.

Once you do that you can do whatever you want with it.

But if you want the code that has been programmed into the FPGA then you will have to negotiate with the boys at Team Vampire because that is theirs and I don't think they want to give it away for free just yet.


Gsteemso Del Canuckistan

Posts 8
21 Apr 2020 16:13


Mercury Thirteen is exactly correct.

Alan Haynes, while "correct" in the technical sense, is also missing the point with such astounding precision that I honestly can't tell if he's trolling me.

I am aware that, if I knew which precise calibre of Cyclone V the current Vampire was optimized for, I could stick one on a little BGA-to-PGA carrier board to make an object physically indistinguishable from the "standalone 68080-alike" I've been discussing, needing only a simplified Apollo Core loaded into it -- "simplified" in this case meaning it wouldn't need any of the Vampire-specific stuff like the SAGA sub-core.

Yes, that would require discussion with Gunnar and the other Apollo developers. WHAT IN THE BLOODY SMEGGING HELL DO YOU THINK I AM HERE TO TALK ABOUT???

I don't want or expect anyone to "give stuff away free". Alan Haynes, you cannot possibly have gotten this far into the thread but simultaneously missed the part where I estimated that people would willingly pay well over 100 dollars/Euro/whatever for a bare '080. Who on Earth would be taking that money if not THE PEOPLE WHO MAKE THE APOLLO CORE?

I don't understand your thought process here, and the only hypotheses I've been able to think of involve you having STARTED by making some very insulting assumptions about both my intent and my character.

I would like to believe that there is some explanation for your words that _doesn't_ involve such insults. I am far too angry to usefully continue this conversation until someone clarifies that.


Gunnar von Boehn
(Apollo Team Member)
Posts 6197
22 Apr 2020 06:52


gsteemso del Canuckistan wrote:

if I knew which precise calibre of Cyclone V the current Vampire was optimized for, I could stick one on a little BGA-to-PGA carrier board

 
We use a Cyclone-V A5.
Please ping me when you have done the board.



Mercury Thirteen

Posts 9
23 Apr 2020 00:34


One other random thought to consider... :)

While some FPGAs - Microsemi's ProASIC 3, for example - do have onboard memory, most others, like the Cyclone series, need additional memory chips separate from the FPGA itself to contain the configuration information used by the chip at boot time, meaning any "adapter" board to make a stand-alone Apollo 68080 a reality would also need space for a Flash RAM chip or two.


Alan Haynes

Posts 140
23 Apr 2020 12:52


gsteemso del Canuckistan wrote:

Mercury Thirteen is exactly correct.
 
  Alan Haynes, while "correct" in the technical sense, is also missing the point with such astounding precision that I honestly can't tell if he's trolling me.
 
  I am aware that, if I knew which precise calibre of Cyclone V the current Vampire was optimized for, I could stick one on a little BGA-to-PGA carrier board to make an object physically indistinguishable from the "standalone 68080-alike" I've been discussing, needing only a simplified Apollo Core loaded into it -- "simplified" in this case meaning it wouldn't need any of the Vampire-specific stuff like the SAGA sub-core.
 
  Yes, that would require discussion with Gunnar and the other Apollo developers. WHAT IN THE BLOODY SMEGGING HELL DO YOU THINK I AM HERE TO TALK ABOUT???
 
  I don't want or expect anyone to "give stuff away free". Alan Haynes, you cannot possibly have gotten this far into the thread but simultaneously missed the part where I estimated that people would willingly pay well over 100 dollars/Euro/whatever for a bare '080. Who on Earth would be taking that money if not THE PEOPLE WHO MAKE THE APOLLO CORE?
 
  I don't understand your thought process here, and the only hypotheses I've been able to think of involve you having STARTED by making some very insulting assumptions about both my intent and my character.
 
  I would like to believe that there is some explanation for your words that _doesn't_ involve such insults. I am far too angry to usefully continue this conversation until someone clarifies that.

No trolling here. I have been following the Vampire since the early Natami days. I also own an A500 V2 and am looking forward to getting a V4 once this COVID-19 is behind us.
I have seen all the hard work behind this Amiga Miracle and I am sorry if you feel like you aren't getting through to anyone. It is in part because to someone as un-educated as I am it appears as if you are trying to get something for almost nothing.  As Gunnar said let him know when you have finished the board BUT YOU DON'T HAVE TO YELL.
Maybe it is the Isolation frustration. Everyone is experiencing it even here in Australia. Hang in there.
Cheers from Oz.



Peter Heginbotham

Posts 214
23 Apr 2020 14:55


Gunnar von Boehn wrote:

gsteemso del Canuckistan wrote:

  if I knew which precise calibre of Cyclone V the current Vampire was optimized for, I could stick one on a little BGA-to-PGA carrier board
 

 
  We use a Cyclone-V A5.
  Please ping me when you have done the board.
 

I'm guessing it would have something like the Quad_doubler 

EXTERNAL LINK


Nonesuch Everwas

Posts 4
23 Apr 2020 19:45


Ah, and now things are a little clearer.
 
  For the record, I did briefly correspond with Gunnar by email, once he was pointed out to me as a person of relevance. When he helped straighten out my account here, I brought the conversation back to this thread because I felt it was of general interest to a large audience.
 
  From the outset, I have been present here to carry out the following steps:
 
  1. If a bare "68080" is available:
  1. a. Buy one.
  1. b. Tell everyone I can contact how to also buy one.
  1. c. (Algorithm terminates.)
  2. Otherwise, remove the cause of bare-'080 unavailability:
  2. a. If the problem involves intellectual property rights:
  2. a. 1. Lawyers exist. Apply money (proceed to step 2c).
  2. b. If the problem involves an insane person who will not entertain discussion:
  2. b. 1. Arranging psychiatric assistance for a stranger on the other side of the world is not practical. Error: the goal is not realistically attainable. Algorithm termnates.
  2. b. If the problem involves resource shortfalls:
  2. c. 1. Organize sufficient new resources:
  2. c. 1. a. Establish an appropriate group effort. (NOTE 1)
  2. c. 1. b. Repeat:
  2. c. 1. b. 1. Publicize and manage the group effort.
  2. c. 1. b. Until sufficient resources have been arranged.
  2. c. 2. Deliver the gathered resources. (NOTE 2)
  2. c. 3. (Algorithm terminates.)
  3. Algorithm terminates.
 
  (NOTE 1) An "appropriate group effort" could be anything from passing an envelope around at one of our monthly Seattle Retro-Computing Society meetings (after the pandemic finally ends), to a crowdfunding campaign, to setting up a small single-purpose corporation and/or a lawyer. It would depend on the exact nature of the problems.
 
  (NOTE 2) As far as I'm aware, this part would involve some kind of contract.
 
  I have not hitherto bothered to make any of this explicit, for two reasons: firstly, it's more detail than I would expect anyone who is not directly affected to find worth knowing, and secondly, because there's very little point in trying to enter detailed negotiations with people who refuse to admit that they do in fact comprehend your request.
 
  I would like to draw your collective attention to lines 1a, 1b, 2c, and 2c2 of my thought process, above. All of those steps implicitly assume that if a "68080" comes from the Apollo team, money goes to the Apollo team in exchange. This is what honest people do when they want to enjoy the product of several years' effort by someone else. To blindly assume any other intent -- without even asking! -- is, yes, incredibly insulting.
 
  I'm not quite sure _why_ Alan Haynes felt justified in defaulting to such a disgusting set of assumptions about someone he has never met, but as he admits that his only involvement with the Apollo team is to have watched it for years and bought a Vampire, I will be immensely relieved to henceforth ignore him.
 
  Gunnar states above that he will discuss this topic after all, once I bring him a usable board. Progress!
 
  I can't actually think of anything else that I might need to ask about in this thread until the pseudo-CPU reaches a field-testable state, but I will check in here every now and again in case someone has a suggestion. There are after all a lot of people here with more knowledge and/or experience in this realm than I have!
 
  I expect all my near-term correspondence will be via individual emails, as I nail down the design details. I will pop back in here as and when significant progress is made.
 
  With enthusiasm,
  Gordon Steemson
 
  P.S. For whatever reason, the forum is not notifying me of replies to this thread. If I don't reply to something, I probably just haven't seen it yet.


Ali B.

Posts 24
23 Apr 2020 23:04


Nonesuch,
I am sorry for being rude, but I am just suffering from curiosity:
What is your experience or qualification in / for kicking off such an effort and what would be your share regarding the group’s work?
I can see you are motivated, you passed the resilience test of this forum, but what else?
Regards, stay healthy
Ali B.


Mercury Thirteen

Posts 9
24 Apr 2020 00:22


Gunnar von Boehn wrote:

 
gsteemso del Canuckistan wrote:

    if I knew which precise calibre of Cyclone V the current Vampire was optimized for, I could stick one on a little BGA-to-PGA carrier board
 

   
  We use a Cyclone-V A5.
  Please ping me when you have done the board.
 
 

 
  More specifically, according to EXTERNAL LINK it's a Cyclone V 5CEFA5, so EXTERNAL LINK and EXTERNAL LINK may help in the creation of an adapter board.
 


Stefan "Bebbo" Franke

Posts 139
24 Apr 2020 17:10


Just stop fiddling around, request a price for 100k+ licenses and I bet you'll be offered one.


Manfred Bergmann

Posts 226
24 Apr 2020 19:41


Stefan "Bebbo" Franke wrote:

Just stop fiddling around, request a price for 100k+ licenses and I bet you'll be offered one.

Yeah. This is not a discussion for a forum.
Just talk to the developers and figure something out.

posts 46page  1 2 3