Atari and Amiga compatibility through hardware

[stormy]

Hmm I admire your passion but I think Amiga/Atari are two different platforms with their own software niches and should be enjoyed as they are. I think it is better to just look at *what you would like to add* on either platform and maybe work towards that, not make them the same thing. I enjoy them because of their differences.

[mikro]

Making something like an "Atari Wine" for the Workbench wouldn't be that difficult but the problem is with hardware compatibility -- many applications, even those for GEM, often access hardware directly. So without further hardware modifications you're looking at building basically an Atari emulator with some parts native, something like what aglarund has done with Mac OS.

[lp]

Apollo 68080 is nice hardware, but I do not have access to develop with it, so it's beyond the scope of anything I'll be involved with.

That's a shame. It already runs some Atari software, just it's limited to very clean apps. Rather than reinvent the wheel it would make more sense if apollo had some fpga/asm gurus to iron out the missing Atari subsystems and write some drivers.

As for uniting the platforms, don't see that happening at the software level. Way easier to fire up the amgia emulator and piss around and it's free. Just saying.

[thomas3]

It's all a waste of effort. The futility is what makes it fun.

[stormy]

Way easier to fire up the amgia emulator and piss around and it's free. Just saying.

So Atari is a waste of effort, gotcha.

It's not that ... It's just your entire premise makes no sense. We had a similar well intentioned Amiga guy here a year ago or so who was talking about porting SoTB to Falcon, when asked if he's consider porting an actual good game he flew off the handle and threatened to throw his rare Falcon in the bin just to spite people on the forum. This thread is starting to feel similar

[shoggoth]

It's all a waste of effort. The futility is what makes it fun.

Not necessarily, but it isn't as simple as putting a DSP in a machine. The architecture itself is quite different. EmuTOS/FreeMiNT/XaAES/MyAES/NAES already runs on the Amiga, as well as on the Vampire V4, and the result is pretty good despite not being compatible at the hardware level.

[lp]

Way easier to fire up the amgia emulator and piss around and it's free. Just saying.

So Atari is a waste of effort, gotcha.

That's not what I'm saying. I'm saying your idea is a wasted effort. I can tell by your questions and lack of Atari knowledge it will never see the light of day.

You dismiss Vampire, yet with a simple reboot it supports both platforms already. That's the smarter approach.

[Eero Tamminen]

It's all a waste of effort. The futility is what makes it fun.

Not necessarily, but it isn't as simple as putting a DSP in a machine. The architecture itself is quite different.

And what makes it even less likely to work, is that many of the interesting programs require CPU and DSP speed ratio to be fairly cycle-accurate to original Falcon (doing explicit syncing would slow down things etc, so many programs, especially demos, don't do it).

[AnthonyJ]

I can make an 030 card with DSP56002 and CirrusLogic GD5434 Video card for my friends ST520 if that would help?

Will that help - probably not? No ST software is going to attempt to use the DSP, and much will fail due to the '030. Almost no Falcon software will run on such a hybrid machine though, unless you're able to make it really look like a Falcon, which that card wouldn't without a lot of work.

Is that Cirrus Logic GPU Atari friendly?

Depends what you mean by Atari friendly. I guess the answer is "probably not" again though.

Looks like a PC VGA chip from a quick google. Could you make a VDI driver for it so you can run cleanly written GEM apps in high resolution - I expect so (like has been done for Vampire). Will it work with anything like Apex Media, or Falcon games - no. For that you'd need to implement a hardware-register compatible interface identical to Atari's custom chip, the Videl (and appropriate OS hooks for it, since you're proposing this for an ST, which doesn't know about Videl).

From the Atari point of view, what Amiga hardware would make the Atari better?

Historically I guess the answer is the hardware sprites and the copper etc which provided functionality which the ST didn't have natively so was done in software, and later the STe blitter was rarely used due to commercial reasons. At this point it's probably irrelevant, with the potential of a fast CPU / more RAM, and with the few coders there are now making use of STe features. As you point out, the big issue would be software support - for which there would be none for Amiga imported HW.

If your goal is to get Atari + Amiga coders sharing the load and so producing more for content for both platforms, here is an idea to try: Implement an Amiga path through the the Atari Game Tools - viewtopic.php?t=31558 - which has potential to make it easy to make high quality games on the STe taking full advantage of it's hardware. I suspect you could write an Amiga back-end for that and end up with it being easier to create games for both platforms, and probably could benefit from some of the tricks already present in the the Atari code-path.

[Atari030]

I'm assuming you were at AUG on the weekend? I saw a stray Falcon among the Amiga's.

I don't think cycle accurate would be an issue. I have two Falcons, one with a CT2b card. CPU and DSP at 50 Mhz. While the CT2 isn't 100% compatible quite a lot runs. I'd imagine its less of a drama under TOS?

[AnthonyJ]

Pretty much. Greater productivity without more effort from the coders. Making it easier and more fun would be best.

Could you help me with direction of what is needed in hardware?

I really don't see that there is anything that can / needs to be done from the hardware side. Even if you ultimately created a hybrid computer that had both sets of HW inside it, I cannot see how you'd get Amiga developers to hit the Atari HW or vice-versa, so it still wouldn't increase the developer base.

[shoggoth]

<snip>

The answers to most of your questions are just a Google away. First and foremost - learn what an ST is, from a hardware perspective. Then take the time to learn how this hardware is uses by common software on the platform. Then do the same for the TT/Falcon if this is your ambition. Then choose a suitable level of ambition for it all.

You won't gain this knowledge by random questions on a forum. You need to invest time in some serious research yourself.

And again, if Atari users would have liked to run Workbench/AmigaOS, or Atari stuff on top of Workbench, they would do so already. There is a choice, and people made that choice. You're not likely to "unite" these crowds, in such case you've failed to understand both of them.

If you want to make a product that appeals to Atari users - make it Atari. Even a plain Amiga with EmuTOS is more "Atari" than some compatibility layer running on top of "the wrong" OS kernel. It makes no sense whatsoever, even if it's a nice idea from a tech perspective. I know - I was toying with a similar project for the Amiga some years ago; basically it executed some command line Atari applications on 68k Amigas. But one has to understand that the primary use for such a thing would be in an Amiga context.