@kolla

Quote:

kolla wrote: @tangoone What, did I make you vomit? You do not see how a feature rich implementation of 68030 in FPGA can be usefull also for Amiga users? Btw, that site also sells SCSI based SD card adapters, which is pretty cool for Amiga computers too.

Just that this project is a Atari st FPGA.

But why is this possible for atari st but no amiga fpga project gets completed.

Last edited by tangoone on 24-Apr-2015 at 07:55 PM.

@tangoone

The hardware on the ST was a lot simpler than the Amiga's custom chips, to lower costs and be able to compete on price. Jack Tramiel was a firm believer in low prices by having simple and inexpensive hardware and compensating for its weaknesses with clever but still inexpensive software, whereas Jay Miner was a gifted hardware designer. Each man's philosophy is reflected in the machines they made. The simpler hardware is just easier to re-create.

edit: replaced "cheap" with "inexpensive" - the ST was the latter, not the former.

Last edited by CodeSmith on 24-Apr-2015 at 08:54 PM.

@tangoone

Quote:

tangoone wrote: @kolla Quote: kolla wrote: @tangoone What, did I make you vomit? You do not see how a feature rich implementation of 68030 in FPGA can be usefull also for Amiga users? Btw, that site also sells SCSI based SD card adapters, which is pretty cool for Amiga computers too. Just that this project is a Atari st FPGA. But why is this possible for atari st but no amiga fpga project gets completed.

There are plenty of finished Amiga FPGA projects, Minimig in various incarnations runs on many boards you can buy... heck, it should even be possible to run on this "Atari ST/Falcon" board, just like it runs on the MiST. In fact, any finished Atari on FPGA project, most likely can also be Amiga projects just by bringing the Minimig chipset implementation to it.

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

What has been plaguing many Amiga FPGA projects, is what known as "feature crawl", instead of agreeing on a set target for a project, many developers change target as they go, and add complexity and features over compatibility and robustness. This is what killed Natami, and from the looks of things, Apollo core is also doing this, one week it is announced that 3 instructions per cycle is agreed upon:

http://www.apollo-core.com/knowledge.php?b=1¬e=2827

but just a few weeks later, it "look - 6 instructions per cycle!"

http://www.apollo-core.com/knowledge.php?b=4¬e=2950

This is just an example, the Apollo project is progressing, but it really is also all over the place and who knows what the end result will be. And that is just messing with the 68k CPU core. In the meantime, the Atari peeps have apparently reimplemented a full 68030 with MMU, which I find very interesting, I wonder if they have 68882 and 56002 ready anytime soon as well, it would make quite a cool machine to play with.

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

@kolla

3 pipelines + 3-way opcode fusion = 6 opcodes per cycle, maximum. These are 2 related but different features. Opcode fusion combines multiple opcodes into one. It combines two stages of optimization.

Quote:

tangoone wrote: this is a amiga forum not atari ST

Amigas used 68030 too. A full-featured 68030 in FPGA is not an Atari-specific thing. It's a 68030... And if license-compatible, Amiga FPGA kits could make direct use of it, extracting that portion of Suska out of the Atari-ST/Falcon bits. Besides, what if other bits of Suska could be interesting as well? Do we have any SCSI in an Amiga FPGA yet? Why not use theirs and make a driver? Then we have SCSI without having to code the Verilog/VHDL part, only the OS driver part (scsi.device), which I would expect to be needed either way?

If MikeJ's processor is license-compatible, why wouldn't Suska people or anyone else, consider using it if they have not completed their own equivilent already? Again, a processor is a processor... Other things use TG68, it is not only for Amiga...

_________________
All glory to the Hypnotoad!

@kolla

Quote:

kolla wrote: What has been plaguing many Amiga FPGA projects, is what known as "feature crawl", instead of agreeing on a set target for a project, many developers change target as they go, and add complexity and features over compatibility and robustness. This is what killed Natami, and from the looks of things, Apollo core is also doing this, one week it is announced that 3 instructions per cycle is agreed upon: http://www.apollo-core.com/knowledge.php?b=1¬e=2827 but just a few weeks later, it "look - 6 instructions per cycle!" http://www.apollo-core.com/knowledge.php?b=4¬e=2950 This is just an example, the Apollo project is progressing, but it really is also all over the place and who knows what the end result will be. And that is just messing with the 68k CPU core. In the meantime, the Atari peeps have apparently reimplemented a full 68030 with MMU, which I find very interesting, I wonder if they have 68882 and 56002 ready anytime soon as well, it would make quite a cool machine to play with.

I think that the real problem of such Amiga FPGA projects is the lack of releases. The developers wants to add features, features, and then features, and never release something.

In the end, it can happen that due to some problems the project is not released / abandoned, and all the good ideas go to the trash.

I know it very well, unfortunately.

So, the "rule" should be: release often, release early. This can:
- create a market (yes, boards can be sold even for unfinished projects);
- people can enjoy them;
- people can tests the product and provide feedback;
- the community can start supporting the product.

Whereas playing with the specs and continuously raising the bar and limits is just a pure mind exercise...

@cdimauro

So the question is this " how can we get new amiga hardware". ?

I dont see anyhting wrong with that they have tried to improve on the amiga hardware. The problems is what happend to project like natami, where one team member owns the code and he takes the project and goes underground.

And thats where it is now.

So how do we get a "natami" type project started that can produce a amiga clone.

Is there anything this community can do. ?

@tangoone

There are successful projects like Minimig. Too big ambitions are always problem.

@pavlor

Quote:

pavlor wrote: @tangoone There are successful projects like Minimig. Too big ambitions are always problem.

Minimig is not a dedicated amiga but you can run many different computers on this. What we want is
dedicated amiga, so when I turn it on it is a amiga nothing else.

@tangoone

Quote:

What we want is dedicated amiga, so when I turn it on it is a amiga nothing else.

Minimig was designed primary as Amiga replacement - even its name shows this legacy.

If you want working new "Amiga" board, you may buy AmigaOne, but that certainly is not something you are looking for...


What features you need for such new "Amiga"?
68030 or 68060 CPU performance?
ECS or even AGA?
RTG?

The more features you need, the more work must be done. Failure of projects like Natami or Tina shows, how hard is to get all, what you want.

@pavlor

no AmigaOne is not something I am looking for I want a updated amiga classic.

Want a amiga where you have possible of connect a lan/ethernet, usb, hardrive or CF. External keyboard
24bit graphics.

So we can have a updated amiga to todays tech.

@tangoone

Quote:

Want a amiga where you have possible of connect a lan/ethernet, usb, hardrive or CF. External keyboard

Quote:

24bit graphics.

So basicaly 68k AmigaOne with full AGA chipset. Too ambitious I fear.
Designing motherboard around SoC with common PC parts (GFX, sound) is much easier than designing custom hardware fully compatible with 20+ years old hardware.

@tangoone

Quote:

tangoone wrote: @cdimauro So the question is this " how can we get new amiga hardware". ? I dont see anyhting wrong with that they have tried to improve on the amiga hardware. The problems is what happend to project like natami, where one team member owns the code and he takes the project and goes underground. And thats where it is now. So how do we get a "natami" type project started that can produce a amiga clone. Is there anything this community can do. ?

Too much effort and a few resources. I don't think so.

@pavlor

Quote:

pavlor wrote: @tangoone Quote: Want a amiga where you have possible of connect a lan/ethernet, usb, hardrive or CF. External keyboard Quote: 24bit graphics. So basicaly 68k AmigaOne with full AGA chipset. Too ambitious I fear.

24-bit graphics doesn't necessarily mean "AGA chipset". Take a look at the RTG cards, which were already available at the Amiga time.
Quote:

Designing motherboard around SoC with common PC parts (GFX, sound) is much easier than designing custom hardware fully compatible with 20+ years old hardware.

The problem is creating the enhanced chipset, not the motherboard.

It also depends on how do you want to extend the original chipset. What I saw with such projects is that hardware engineers want to patch the existing chipset(s) specs, adding new features using the same "infrastructure". For example, they duplicate the chipset registers bank to "gain" double audio channels, bitplanes, and sprites, just because this way they can recycle their existing code base for the original chipset.

I don't see a future for such solutions: no new clean design (with backward compatibility). Just hacks.

Anyway, the post-Amiga scene is only hobby. And those kind of choices can be seen as are the result of people playing with their toys. Just for fun.

@cdimauro

Quote:

24-bit graphics doesn't necessarily mean "AGA chipset". Take a look at the RTG cards, which were already available at the Amiga time.

That is what I had in mind - AGA+RTG.

Quote:

The problem is creating the enhanced chipset, not the motherboard.

No chipset, no motherboard.

Quote:

I don't see a future for such solutions: no new clean design (with backward compatibility). Just hacks.

Isn´t that intention behind AmigaOne project?

Quote:

And those kind of choices can be seen as are the result of people playing with their toys. Just for fun

Exactly.

@pavlor: no, there's nothing related to AmigaOnes.

@cdimauro

Quote:

no, there's nothing related to AmigaOnes.

No Amiga legacy components, based fully on common parts, maintaining backward compatibility - isn´t this what you described?

@cdimauro

Quote:

cdimauro wrote: @tangoone Quote: tangoone wrote: @cdimauro So the question is this " how can we get new amiga hardware". ? I dont see anyhting wrong with that they have tried to improve on the amiga hardware. The problems is what happend to project like natami, where one team member owns the code and he takes the project and goes underground. And thats where it is now. So how do we get a "natami" type project started that can produce a amiga clone. Is there anything this community can do. ? Too much effort and a few resources. I don't think so.

What is needed to complete it when u say "resources". ?

Funding = cash

or

knowledge

?

@pavlor

Quote:

pavlor wrote: @cdimauro Quote: no, there's nothing related to AmigaOnes. No Amiga legacy components, based fully on common parts, maintaining backward compatibility - isn´t this what you described?

Absolutely no. For such projects, emulation of the old chipset is the bare minimum required. No discussion here, and nothing which is related to the AmigaOne.

The "dispute" is about how to evolve/expand/improve such hardware to introduce more modern (and desirable) features. That's "all".