@Nameless

It really makes good sense . Pixie (or a board
similar to it) provides plenty of Pins for output
that could be used to emulate real Amiga ports
and it is small enough to fit in an Amiga 500 or
1200 case probably ...

@blizz1220

No biggie about the identity confusion.

One thing about the pixi that may not be ideal is how the fellow who made it set it up. I read somewhere that he left out a chip or something related to memory retention. So that the FPGA contents would need to be reloaded with each startup. Or something like that... I'm a FPGA newbie, so don't know the tech details. I think it's a cheap alteration to make it work properly though, so I guess if anything came of this, it wouldn't really add much to the overall price.

As for what to do, a kickstarter could make sense, if there was an actual manufacturer who wanted to handle the FPGA stuff. A kickstarter for a JIT ARM could work too, of course, but would need the coder to do his own kickstart campaign.

It's the sort of thing I'd think someone related to the Amiga could do ... Jens with his Clone A thingy... CUSA (if they didn't go down the wrong path) ... Amiga... Hyperion + partners. I'm just not sure we can do it, without at least a coder interested.

One alternative is to use interpreted ARM, rather than JIT. All it may need then is an altered UAE, which has to be a lot faster to make than the JIT. It should still be able to do 1200 speeds, I'd hope, even interpreted.

@Nameless

Well , to keep it simple you could start a kickstarter
that would require a purchase of just Pi and Pixie ...

Description should spell something like "a project
that will need us to hire FPGA coder to write hardware
language code to hardware emulate essential parts
of AGA Chipset for the Pixie card and hire competent
coder to create small lightweight linux distro and modify
source code for the UAE Amiga emulator and write
a JIT for it (when you pay for it it's not a problem to
find a guy to do that) " ...

I would look into starting Kickstarter myself but it has
to be done by US or UK citizen.I will go more deeply into
the technicalities of this in next few days ...

And I wouldn't expect much help from any of the people
or organizations you mentioned above

End result would be something that has real AGA chipset
and CPU that's faster than 060 so :

A1200 060 (100+ Mhz) 128+ Mb P96 LAN and USB adapter
all for the price less than 150 Eur

Last edited by blizz1220 on 29-Jul-2013 at 11:13 PM.

@Tom01

Quote:

Simlpy run UAE on a fast cpu. Setup a minimal Linux system. No FPGA neded.

because emulating chipset takes a major share of host cpu time and fpgas are exactly suited for such tasks afaik, but otherwise i agree with you.

@wawa

Because software emulating a chipset will
never give you 100 % identical result to using
the real hardware ...

Using FPGA to hardware emulate is the same
thing as taking real chips out of your 1200/4000
and attaching them to RPi ...

@blizz1220

Yeah, I wouldn't expect any help from those groups either. But logically it would have made sense for them to do something like this. It still beats me as to why Natami or none of the FPGA groups ever set up an ASIC kickstarter. To me, that is the real way to get Amiga in tons of people's hands. An 'Amiga on a chip' device could in theory be sold for $25-$35 or so, and probably sell quite well.

As for starting a kickstarter for this, I wouldn't feel comfortable starting one with the goal of hiring some unknown programmer. It'd need a coder onboard from the start. There may also be some reseller licensing issues with the Pi, as I don't think just anyone can resell it however they please. And the Pixi board is still in an early stage. The one that probably would be best would be the cheapest one, but I don't think it's even been released yet. And no idea if it'd even work for a project like this.

We'd really need a FPGA expert and coder to get involved.

@blizz1220

Quote:

A1200 060 (100+ Mhz) 128+ Mb P96 LAN and USB adapter all for the price less than 150 Eur

I'd aim even lower. The Pi ranges from $25-$35.

The A model pixi was previously announced at around $45 (converting from pounds). Although now I see the price is listed as £TBD, so maybe it's more or less than that amount.

Anyway, it could be possible to get the price down to less than $100 for both, which would be really cheap when compared to most Amiga projects.

@Nameless

I thought Pixie was 100 $

Yes , it would be a great machine and this is
worth investigating into , it even has Tv-Out
compatibility so add FlickerFixer to the above
general specs

@blizz1220

The first model they came out with was around 70 pounds, so you are sort of right. But they also listed the A model for half of that, if I remember right. The pixi guy has stated a target price of £35.

Now, I have no idea if he'll actually get to that price, but ideally that's what it costs, or less. The FPGA used in the pixi is pretty cheap, so if someone here was into making FPGA boards, they may be able to match it.

A lot of the current FPGA projects tend to think big-ish, Natami-style full systems, AGA or better graphics, with extra space to spare. A small FPGA is probably all that is needed for something like this, although a little extra legroom for other chipsets would be nice too. And all that said, I still have no idea if the pixi would be suitable, or the right size or speed. But it does seem to be the right price.

Last edited by Nameless on 30-Jul-2013 at 02:10 AM.
Last edited by Nameless on 30-Jul-2013 at 02:09 AM.
Last edited by Nameless on 30-Jul-2013 at 02:09 AM.

@Nameless

what happens to the guys who did the natami ?????
i remember a photo or two even a video ??? of the natami running...
why nobody could help them to make this available ???
a-eon, hyperion, i don't know anyone... this fpga seemed incredible and affordable.
Why nobody is trying to save it and port the amiga os 4 on it ????
Or perhaps somebody do it secretly ?
i can't believe that this gem will disappear forever, we shouldn't accept that.

Last edited by turrican33 on 30-Jul-2013 at 05:56 AM.

@turrican33

The memory controller was a bit slow on the NatAmi prototypes. It might not have been able to compete well against the FPGAArcade Replay board anyway.

As for the Tina board, one of their team posts to AmigaCoding.de and is trying to implement a 128-bit path for all data fetches to take maximum advantage of the DDR2 memory fetching while offering upward compatibility to DDR3. If successful, it would alleviate the memory controller shortcomings present on the NatAmi.

Also on AmigaCoding.de, there is an engineer which is making an improved core for running Classic Amiga software. (Look for the username "Veda".)

@turrican33

I haven't kept track of Natami so much, but the general gist of it, to the best of my knowledge is:

Original Natami developer sort of took over and told everyone else to go home. I think he meant it more as a personal hobby sort of project, yet the others involved had wider plans... or something. Anyway, it did look like feature creep to me, where instead of focusing on working on a single design and actually releasing something, they kept adding extra stuff, new softcore cpu, etc.

I don't think anyone can finish it now besides the original designer. Although I guess those who worked on the softcore 68K stuff could offer up their design to other FPGA projects if they wanted to. I'm not sure if they ever got that finished or not.

It probably would have sped things up if they just used a different CPU and a JIT compiler. I probably am in the minority, but at the time I thought it'd be kind of neat to use Coldfire. I realize it's a bad choice if not using JIT, but the Atari people managed it easy enough ... and it should be a lot easier making JIT for coldfire than other chipsets, due to a lot of operations being equivalent. But I think the best Coldfire chip isn't available for general purchase, unless ordering a ton of them.

@Nameless

even if it looks like, it will never see the light.
i hope we could one day see a result.
The guy who is behind the project should give news even if it's bad news.
This project was a good one to have something very near the most powerfull amiga classic.
too bad.

@thread

Natami end user price from a distributor with a
case and all accessories would probably be
more expensive than FPGA Arcade with 060
daughterboard ... I have no information about
Tina ...

I think everyone could learn from mistakes that
were made with PPC AmigaNG systems ...

Having chepeast possible entry level priced
FPGA model would be a good thing to get new
users interested in FPGA Amiga systems and
with new users there would be new developers
for Aros68k ...

* * * More info about Pixie board * * *

http://www.raspberrypi.org/phpBB3/viewtopic.php?f=63&t=18688

@Nameless

I will start new thread when I do a little more
research on the possibilities of combining
Pi and Pixie ...

Last edited by blizz1220 on 30-Jul-2013 at 01:35 PM.

@blizz1220

I agree entirely in regard to price. I know a lot of people here probably want super fast Amigas, or replacements for their current setups, but if people want to really increase the userbase over a couple of thousand (or however many their currently are), cheap is the way to go. The market to go after is former Amiga users + those who are just curious about the platform -- and they won't spend $500+ on a 3rd or 4th computing device just for fun.

The Pi is a good example, as plenty of people are willing to spend $35 for a little machine just to play with. Chromecast, which was recently just released, is another example as it's sold out in most places -- although that is more for streaming than computer stuff (but I expect it to eventually get hacked).

ASIC would be ideal, but overlooking that, a sub $100 FPGA Amiga type system would be a move in the right direction.

@Nameless

Isn't that market already covered by the Minimig?

@megol

Minimig is Ecs 68000 16 Mhz , this could be AGA and
020 68 JIT on ARM so it would be more like 060 at least.

Some ideas that come to mind are that it would maybe
be better to use Aros instead of Linux and give it a NG
aspect also ... This way you could have a choice if you
want to boot to "real" Amiga 1200 or do you want to
use Aros for the Raspberry ...

Beagleboard seems like it's better quality tha Pi and
Pi can clock to above 1 Ghz but I don't think it would
last very long that way without some soldering here
and there so 800 Mhz (if possible?) would be optiumum ...

Last edited by blizz1220 on 01-Aug-2013 at 03:17 PM.

@megol

What does a pre-configured MiniMig cost? I think it was over $200, wasn't it?

Ideally a Pi + FPGA would be much cheaper than that. The Pi part of it would be around $35.

@Nameless

Quote:

Ideally a Pi + FPGA would be much cheaper than that. The Pi part of it would be around $35.

Manufacturing in small batches is not cheap. Forget about price under 200 EUR.

What is needed:
Suitable ARM+FPGA board design - not available.
Production in higher numbers - impossible for our small community.
68k JIT for ARM CPUs - not available.

@pavlor

IIRC Thomas Hirsch (_THE_ Natami developer) had some idea to sell Natami boards as FPGA development boards to lower costs. I may be mistaken though.