A New Amiga Architecture - 3AC

[Gorf]

Originally Posted by EAUniW

Why a 100 million? Can't we just take the existing AGA chips and use them as the basis for upgraded internals? As in soup up those chip designs?

because every ASIC with that complexity costs around a million - just for the production setup (without development). Just one prototype. And you probably need more then one ...

Even if you had access to the AGA internals and where allowed to modify them - but they are still copyright protected ...
Then you could maybe save some development time ... but even that I doubt, since technology changed a lot, a complete new reimplementation would probably cost the same...)

And you don't want just one ASIC but many chips ....

The "ZX Spectrum Next" campaign collected almost a million - but still they can't afford a singe ASIC and use a FPGA...

[EAUniW]

..

[Rotareneg]

Originally Posted by EAUniW

Why a 100 million? Can't we just take the existing AGA chips and use them as the basis for upgraded internals? As in soup up those chip designs?

Ah, I see a slight flaw in your idea already...

[Mathesar]

Originally Posted by Gorf

since AHI uses 14bit modes to play back 16bit samples/audio (I have seen no other use case yet) it makes sense to have 16bit channels.
"Pamela" on the Vampire does that (mixed at 24bit internally)



we have this already in different FPGA solutions ... but modern TVs/monitors don't need it anyways: they are perfectly flicker free at 25Hz



use case?
(you would need no additional DMA slots: if Paula could handle 32bit accesses the DMA timing can stay the same ... but who wants to handle floppies? And why the hell would you need HD floppies )



So? MISTer or Minimig do that already

Not all monitors handle interlaced modes correctly. Also, not all monitors can sync to 15khz. At least not analog. Maybe a next gen Amiga would be better of with an HDMI output.

As for HD floppies, I don't need them but when you're at it then why not support it as well? And you can also do it via 32bit accesses. Whatever is convenient.

My point is, keep everything transparant for the current software so that means no messing around with registers and such. I'm pretty sure that the minimig (at least the original cores) don't do that as well. And yet it supports a scan doubler, faster floppy interface, transparant SD card harddrive support and faster blitter and chipram access.

Now take this to the max and you have a really fast bit very compatible Amiga 2400 or something

[EAUniW]

..

[Gorf]

Originally Posted by EAUniW

Huh, really? Ok, FGPA it is then. Can I have discrete FGPAs for each chip though or is it an all-in-one deal?

you can have of course many ... but even 5 smaller ones are more expensive than just 1 bigger containing everything...

And then you are back to projects like Minimig or Vampire....

That said-
Even if it is totally useless and makes no sense and is probably a stupid idea:

I actually would very much like a new board for my A3000 with individual and exchangeable FPGA (and/or microcontroller) replacements for the custom chips ... just for the geek effect ...

[EAUniW]

..

[Mathesar]

Originally Posted by EAUniW

Based on that, what would your specifications for the Amiga 2400 (All-in-one) and 4800 (Desktop) be? So I can get an idea of where to pitch it in terms of wanted power versus actual implementation power.

My dream Amiga would be neither an all-in-one nor a desktop computer....
It is that I don't have the time but otherwise I would build an Amiga mobile phone. Like this: https://www.concept-phones.com/tag/c...iga-500-phone/
Note that I am not joking here! It's doable with todays tech like fpga's 3D printing and oem lte/gsm modules. Just imagine....

[EAUniW]

..

[Gorf]

Originally Posted by EAUniW

It's a good idea but I want this to be the first new all-in-one Amiga since the 1200, internals housed in a keyboard case with side-floppy. I don't mind it being smaller though while maintaining the keypad. I'm thinking 3/4 the width and height of the 1200.

Also, what's the chances of using common PC USB floppy drives as an Amiga DD and HD floppy drive? Or are the mechanisms too different? I was thinking the drives could be bought, stripped down and housed internally.

I guess real floppy drives are a pain in the ass to get right ... and they are rather big and clunky ...
I would not mind an external connector for people the really need/want that medium - but you could have such a slim and nice keyboard computer without!
(I think the ZX Next is close to perfect in this regard ...)

[NorthWay]

Originally Posted by EAUniW

Can I have discrete FGPAs for each chip though

Sure, but why? Jay Miner&Co would have gone for a single chip if it was possible (in as much as it virtually _is_ a single chip, it's just split between 3 ICs).
(You might want to study some pics of the original IBM Power CPU...)

[EAUniW]

..

[Gorf]

Originally Posted by Mathesar

My dream Amiga would be neither an all-in-one nor a desktop computer....
It is that I don't have the time but otherwise I would build an Amiga mobile phone. Like this: https://www.concept-phones.com/tag/c...iga-500-phone/
Note that I am not joking here! It's doable with todays tech like fpga's 3D printing and oem lte/gsm modules. Just imagine....

"Shut up and take my money!"

(Even if it it would be just a Android/Linux/Sailfish underneath I would buy it!)

[EAUniW]

..

[Mathesar]

Originally Posted by Gorf

"Shut up and take my money!"

(Even if it it would be just a Android/Linux/Sailfish underneath I would buy it!)

It would run workbench of course! On a tg68 core! Ok, so no whatsapp, youtube or other apps then but we could have a phone app, texting app, play dynamite online, run ibrowse and a mail client. Play lemmings or KK's dread Doom clone, show of the latest demo's like eon and way too rude, listen to mod files, etc, etc.
It really is feasible!!! Now the only thing we need is a touch interface / mouse emulation....

[EAUniW]

..

[NorthWay]

I believe this is known as retcon'ing, a very popular sport in the Amiga communities.
My own take on it is to speculate what could have been achieved if the design was targeted at what they could expect to be possible in 1985 (I know, they had a vague hope of a 1983 release IIRC).

(Just for fun, here is the Power1 CPU (yes, just a cpu - number of ICs be damned) https://upload.wikimedia.org/wikiped.../Power-cpu.jpg)

[Gorf]

Originally Posted by EAUniW

A rough and ready approximation of how the 2400 should look.

I would prefer a keyboard like this Amiga-prototype from 87 has:
https://cdn.mos.cms.futurecdn.net/0f...19-1024-80.jpg

It looks almost modern, doesn’t it?

[EAUniW]

..

[EAUniW]

..