@agami

We now have #emu68k for arm and we have AROS for arm also, wouldn't it be nice having AROS on arm using emu68k as a virtual processor for AROS 68k, it could be used like petunia/trance are used on AmigaOS4/MorphOS

_________________
Indigo 3D Lounge, my second home.
The Illusion of Choice | Am*ga

@Karlos

You "bluntly" missed the point by 628km.

There is a whole lot of hype around 68k today, just like there was a lot of hype about NG 20 year ago.
Plenty hype about the BoXeR, A/Box,PreBox,P.I.O.S, Merlancia, iWin, AmigaDE, QNX-Amiga (GateWay) back in the days.

None and really none of of it had the slightest chance of living up to that hype even if no errors had been made.
Some might have turned into viable products/platform but the hype always was and still is far beyond such earthly endings.

_________________
- We don't need good ideas, we haven't run out on bad ones yet
- blame Canada

@Kronos

the traditional amiga market will not collapse and is growing. There is even more chances ahead because many former amiga users are not aware of current products and at least some of them might buy something again. But it is part of the retro trend, I too do not see the potential for new 68k based hardware to compete with INTEL or ARM

@pixie

that is what Michal (emu68) basically plans, aros 68k or amigaos running on RPi with emu68 as emulation layer. A kind of amithlon for ARM

@Kronos

I do not know what you talk of... what "hype"? There are facts. If you look at amiga 37, the biggest amiga event for many years, 99% about traditional amiga products (hardware and software). I also think there is a retro trend and at least some former amiga users today are not aware what is happening and might also join again. So I think it is a growing market. Will it become mainstream again? Of course not, it is a niche. If you want to get outside niche you need attractive products that offer something new at a affordable price people are willing to spend. Outside retro we currently do not have that.

@Kronos

You're talking about hardware 68K solutions. Nobody is going to build matthey's dream 2GHz 060 ASIC. And nobody needs to, either.

Back when I had both an amithlon compatible machine and a G4 "NG Amiga", I often compiled and tested code on both. It didn't take long for 68K JIT to trounce native PPC except in cases where vectorisation was possible for the G4. I'm quite confident a modern 68K JIT on modern x64 commodity hardware will scale accordingly. Meanwhile what do you have for PPC? Ageing G5 and sky high expensive PA Semi as the peak?

What key advantage does PPC have that a JIT 68K does not? Altivec? Even then only on some machines. As long as you have the ability to do native code calls from your JIT, you can have avx accelerated datatypes and codecs for all the places where it helps.

I put it to you that if your PPC NG OSes were backported to 68K and ran under such an environment Joe Average user would not miss a thing. Would they even notice?

_________________
Doing stupid things for fun...

@OlafS25

Quote:

OlafS25 wrote: @Kronos the traditional amiga market will not collapse and is growing.

Did I ever claim otherwise?

But....

I entered here replying to the idea that an new (non-FPGA) 68k could be competitive which would be only viable at a scale 100 or 1000 times bigger than what the "traditional amiga market" could grow into.

Sure there are a few millions ex Amigans and with a really good and cheap "new" Amiga one could get 1 or 2% to buy into it. Above that noone cares and even most of those who did buy one will loose interest sooner or later.

But the hype wants us to believe if we just all pull together 68k could do what others do while shipping billions of units each year.....

Thats the point, back in the 90s that same crowd believed that Amiga could compete with Windows/x86 if the current brand name owner would just xxxxx.
In the 2000s both red and blue fanboys believed that if the other side rolled over it would be enough to form a viable commercial platform.

In reality anything "Amiga" was is and will be doomed to obscurity ever since C= failed to bring proper updates to it in the late 80s (everything else was just momentum pushing them on for a few more years).

_________________
- We don't need good ideas, we haven't run out on bad ones yet
- blame Canada

@Kronos

as I wrote: If you want to get outside niche you need attractive products that offer something new at a affordable price people are willing to spend. Outside retro we currently do not have that.

Amiga brand certainly still has some value if it could be used for new products. But first we would need products that have a kind of "must have" factor. There is nothing at the moment. What future brings we will see. In any case it must be modern products. I do not see a new 68k hardware there either (except the retro market, but for that you do not need a 1 GHz 68k system, FPGA is more than enough already).

Last edited by OlafS25 on 22-Oct-2022 at 10:37 AM.

@Karlos

See that is (another) problem with "68k for the future", it is just as fractured as NG. I point out the utter nonsense of an 68k ASIC and you hook into it talking bout your virtual 68k-ABI.

Do I know the reality of how fast that is? I switched from an AthlonXP1700(1.4GHz) running Amithon to a 600MHz Peg1 and "speed" for sure wasn't a selling point.

The fact that PPC is dead and failed to bring on the glorious return of the Amiga does not mean hitching ourselves to another dead architecture (emulated in SW, emulated in HW or as real HW) will succeed in that achieving that daydream.


>I< have certain "needs" I want to be fulfilled by something Amiga(releted). Right now MorphOS on old Apples does an o.k. job at these while nothing 68k would come close.

Will they be fulfilled by something different in the future? I don't know but sofar I'd say MorphOS running natively on AMD (or maybe ARM) has the biggest chance of not failing.

_________________
- We don't need good ideas, we haven't run out on bad ones yet
- blame Canada

@Karlos

Yeh, its hard to ignore decades of 68k code, that’s why I’m an advocate for system friendly 68K code, but. Now I most of stuff being made is not written in 68K code, but Blitz Basic II and C, also C++ is becoming increasingly popular. There is also dash of Hollywood code, that targets all platforms.

the problem with many decades of old 68K code, is that is no source code, there license might not actually allow you make changes to it, you can need to establish the history around it. the developers who wrote is not around, when to other platforms, or are dead. You need pretty desperate to become a grave digger.

I think better solution is combination of FPGA and modern CPU’s, FPGA can do hardware banging part, we emulate what most, recompile what’s recompilable, and build new tools and programs.

_________________
http://lifeofliveforit.blogspot.no/
Facebook::LiveForIt Software for AmigaOS

@OlafS25

Quote:

OlafS25 wrote: Amiga brand certainly still has some value if it could be used for new products.

Amiga brand is not known by anybody younger than 40.
It is not known by anybody in Asia.

Many of those that do know it associate it with failure or "toy".

_________________
- We don't need good ideas, we haven't run out on bad ones yet
- blame Canada

@Kronos

there are a lot of people of that age that today even have money. Amiga was sold in millions only in germany

@OlafS25

So?

You sell a 10k units to them in the 1st year and than what?

It is a nice and maybe viable product but for sure nothing that would count as al longterm business strategy.

_________________
- We don't need good ideas, we haven't run out on bad ones yet
- blame Canada

@pixie

Quote:

pixie wrote: @agami We now have #emu68k for arm and we have AROS for arm also, wouldn't it be nice having AROS on arm using emu68k as a virtual processor for AROS 68k, it could be used like petunia/trance are used on AmigaOS4/MorphOS

Not possible.

AROS on ARM isn't running in big-endian mode, so the transparent execution of 68k applications doesn't work.

@OlafS25

Quote:

OlafS25 wrote: @pixie that is what Michal (emu68) basically plans, aros 68k or amigaos running on RPi with emu68 as emulation layer. A kind of amithlon for ARM

Actually he's running on PiStorm, so it requires the original machines.

So, it's something completely different from Amithlon.

@Hammer

Quote:

Using PC VGA PCI cards that initialize VGA X86 bios (part of UBoot's initialization function) is almost an X86 PC.

That started on the real thing when people bought expensive PCI expansion cards in order to stick cheap PCI cards into their Amiga.

The AmigaOne someone simplified the process by getting rid of the Amiga part which people were replacing in their PCI Amigas anyway.

It brought the OS closer to x86 but not running on x86 so still not good enough for some people almost 20 years ago.

@agami

Quote:

These kinds of comments are good for rousing the troops, but are ultimately pointless. Outside of sentimentality, he provides no good reason why it shouldn't run on a PC. Like AROS, (eventually MorphOS?), Haiku, or as Linux or BSD distros do.

These days it's a different game. The world isn't just PC or x86 something and Apple forced to follow along like 15 years ago. ARM really picked up the pace as an alternate architecture.

Now strangely, Apple have gone from RISC to CISC and are migrating back to RISC again. They already moved on from PPC but then they are now moving back to a PPC replacement. If they already solved the CPU problem by going x86/64 why would they repeat the same path they took earlier by going back to RISC than can never beat x64?

But, the reason OS4 shouldn't run on a PC, is because it technically cannot as it is. OS4 is a big endian OS so it cannot naturally run natively as is. It also has dependencies on PPC and some of these are lower level interrupt contexts that are PPC specific. However, like I said in another thread, my work around would be a hybrid solution, with a custom compiler that only emitted big endian x86 codes.

Aside from that, going x86 or x64 now means that people expect to run it on their home PC and any old PC lying around. Most likely they would still need a custom PC build. I wouldn't expect any hardware to work as they would need a mobo with support for all drivers on board.

@Nonefornow

Quote:

OK. Easy enough. I don't think PCs want to run OS4.

Ha.

Just to stir the pot further, the facts as I see them are that most of the significant PPC software for NG has source code (some curios from the warpup/PowerUp era maybe not). This is somewhat different than the 68K situation where there is decades of software that exists only in executable form. Thus transitioning away from PPC to, well, anything, ought to be less of a challenge.

In the end though some people will stick to PPC regardless. It's probably fair to call it a cult following.

This idea of a CLR has legs though, even on the weakest PPC. My 603e with Petunia could execute 68K code much faster than my 040 ever could, evidenced by the fact I could comfortably play the 020 doom attack port at 640*400 or the fact that it relied on a 68K driver for the PCMCIA NIC.

With the right ABI and standards in place one can easily imagine applications, libraries, devices, datatypes and UI classes working across the spectrum of physical and virtual solutions.

Or you can stick to some obsolete niche CPU and just wait for it to execute it's last cycle.

_________________
Doing stupid things for fun...

@Karlos

Quote:

This idea of a CLR has legs though, even on the weakest PPC. With the right ABI and standards in place one can easily imagine applications, libraries, devices, datatypes and UI classes working across the spectrum of physical and virtual solutions.

⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀⠀[ 𝔸ℙℙℝ𝕆𝕍𝔼𝔻 ]



Thread closed.
Everyone else please leave now.

Thanks,
/MEGA

_________________
I HAVE ABS OF STEEL
--
CAN YOU SEE ME? CAN YOU HEAR ME? OK FOR WORK

@cdimauro

no it is not. Amithlon has no chipset emulation either as far as I know. So what Michal wants to do is running AmigaOS or Aros 68k on RPi without needing any host system or integration in original amigas.