A... Coldfire Projects

[DJBase]

Originally Posted by TheCyberDruid

I meant if Mathias knows him

Ups...sorry.

Originally Posted by alexh

My comments were not just aimed at Atari Coldfire Project. They were also aimed at the NatAmi project which has consistently failed to delivery anything while promising everything. A little bit of spent up anger let out at once

OK, I see. To me it seems now you are not really not liking us, you just cannot imagine that we will finish it.

Originally Posted by alexh

Realistically thought, you're are not going to be able to make an Atari Falcon clone using an existing Coldfire processor and get it to run binary 68k code anywhere near as fast as an 060 based CT63 or Milan because of the incompatibilities?

That is just your analysis! And it is wrong. Already before we starteed, we had a proof of concept, that showed execution of 68k applications inside a native ColdFire system is possible. What do you expect about an computer with 133MHz systembus, DDR2 Ram etc. running under full speed while only some native 68k instructions of some applications have to be trapped? And we are not talking about our recent developments here!

Originally Posted by alexh

• You're using a Coldfire CPU to replace a 680x0 CPU.

Yes of course. That´s our aim. As the fastest availaible 68060s are running at 105 MHz with some overclocking. Thus it´s named Atari COLDFIRE Project

Originally Posted by alexh

[*]You're using an FPGA of potentially fixed size and speed grade to incorporate a HDL design who's size and complexity is as yet unknown.[/list]

Maybe you are not aware about it. Our FPGA design is mainly based on Wolfgangs Suska. He worked with Carol for more than 6 years now at the mapping of all Atari chips. The Suska is availaible for over one year now. I got one at home and it´s woking very good. So nothing is unknown! And we chosed the huge and very expansive (120,- Euro) FPGA to have enough headroom for future improvements.

[TCD]

A YouTube vid says more than 1000 words I bet you know Gunnar

Originally Posted by TheCyberDruid

@Mathias
Do you know Gunnar von Boehn by any chance?

Yes - in a kind. Not personally, but we exchanged a few mails, about our projects, possible support, and especially the V5e. As we at the ACP got first requests, if we like to beat NatAmi, it was important for us to explain, the we do not see any rivalry between our projects. Why are you asking?

Originally Posted by alexh

Push home that applications will have to be recompiled, re-assembled and possibly re-written to get any speed advantage on the Coldfire platform.

Another advantage is that a ColdFire is 25,- Euros each, availaible in high quantities, got several controllers on chip (DDR, PCI, MMU, …) and that it has a future. A single 060 is $450 ordering directly at Freescale at the moment.
Well, recompiling applications will lead to full speed execution at 266MHz - that´s true. But be sure that we are producing a computer which will be able to execute all native Atari applications at a reasonable speed.

Originally Posted by alexh

Push home that while high compatibility with Falcon hardware is the goal it will be a long time coming, if it is at all possible with the first generation FPGA boards.

Did you read anywhere that we promise 100% Falcon compatibility in the next weeks? No, of course it is a "long term aim". We always promised compatibility better than the Hades and usage of every clean GEM application. And of course it will not be possible for every Falcon application, especialy when they require exact timings of the 16MHz 030. But we based our design at the Falcon, and will include the DSP as sonn as it is ready etc.

Originally Posted by alexh

Honesty at this level and my attitude will change.

It will at least change when you see the first working machine

Originally Posted by alexh

Phrases like "make an Atari Falcon clone with nearly full compatibility to an original Falcon" IMO is just not realistic in a short enough timespan using today's FPGA technology at a price point that will be viable.

It is a very good goal, but should include emphasis this is a long term goal for the underlying HDL cores. That compatibility of this type will take many man years.

As said above, the Suska development is going on for years now!

Originally Posted by alexh

I say the same to you: do not make the public believe in unrealistic things.

I respect if you think it is not possible. But please use terms, that make clear that it is your opinion, not a fact. Look, for example, we had a complete boarddesign, before we showed up again in April last year. We did not start again by telling "we plan to do something" we started with "here it is". So I am not sure why you are thinking how you think.
I understand that the 68k <-> CF instruction issue is delicate, aspecially as Coldfusion and Dragon failed. Maybe you should read the Microapl page for some basic techiques http://www.microapl.co.uk/Porting/Co...technotes.html And even if the CF68KLib is not optimized (doesn´t know V4e at all, ...) Didier got good resoults 3 years ago. So please follow our development this year.

[TCD]

I was just asking, because Gunnar does the same 'job' for the NatAmi project as you do for ACP. Mainly because of the stuff you just answered Do you plan to release any vids on your project soon?

[seuden]

Surely alexh has a point. I don't care what anyone says but imo using an fpga is basically hardware emulating something. If UAE is over ten years old (~14 years old) and still not 100% compatible and in the words of many wise men it probably will never be, what are the chances that these fpga projects will be, especially soon?

Originally Posted by seuden

Surely alexh has a point. I don't care what anyone says but imo using an fpga is basically hardware emulating something. If UAE is over ten years old (~14 years old) and still not 100% compatible and in the words of many wise men it probably will never be, what are the chances that these fpga projects will be, especially soon?

Yeah, I did not tell, that 100% compatibility is possible. In fact for some applications - those who use directly the hardware, including timings etc. also as those who use registers directly beside the OS, it will becom extremely hard. That will not be done in some months from now. But please have a look at the working Suska, which is under development for 6 years, and the base for our FPGA configuration.

http://shop.inventronik.de/store
http://experiment-s.de/en/

[Retro1234]

Ive put one of the Videos of the Dragon in the Zone as I cant seem to find them anywhere on the Web, Its hard to tell just how fast it was/is running hes seems to have a TV card in use but it Doesnt look Super fast.

And as far as I can tell everthing running native - without being recompiled?

Originally Posted by TheCyberDruid

I was just asking, because Gunnar does the same 'job' for the NatAmi project as you do for ACP. Mainly because of the stuff you just answered

Ah I see, the typical "german-language-guy-as-coordinator-of-marginalised-computerplattforms-hardware-projects-speech"

Originally Posted by TheCyberDruid

Do you plan to release any vids on your project soon?

We will release videos of our working computers. But "soon" can also mean 12 weeks or something. I will come back and report when they are online.

[cosmicfrog]

just 2 more weeks then

[gilgamesh]

Update on the timeline http://acp.atari.org/news.html

[jotd]

a special version of WHDLoad would be required too I'd love this coldfire board