I don't really know anything about AROS for 68k, and still have a lingering (probably misconception) thought that it is not nearly as small, fast and efficient, as the Commodore versions of AmigaOS (v1.0 to v3.1), so it isn't really useful on actual Amiga hardware, but is better suited to emulation.

Can anyone here update me on the status of AROS for 68k and give me an idea how far it has improved and how far it still has to go to be considered as good or better than AmigaOS3.1?

Also, does AROS for 68k have all the same limitations as AmigaOS3.1, memory protection lacking, single user, etc. I am guessing that it does have all the same limitations, as it aims to duplicate AmigaOS3.1, not invent a different operating system for old Amiga hardware. Probably not many users would be interested in an alternative operating system for Amiga hardware, that was incompatible with all of the Amiga software, and therefore didn't have any compatible software in the beginning.

[ Edited by amigadave 29.05.2016 - 21:27 ]

> I [...] have a lingering [...] thought that it is not nearly as small,
> fast and efficient, as [...] AmigaOS (v1.0 to v3.1), so it isn't really
> useful on actual Amiga hardware, but is better suited to emulation.

This would seem correct. The new FPGA-based Vampire V2 cards may be fast enough, though.

> does AROS for 68k have all the same limitations as AmigaOS3.1,
> memory protection lacking, single user, etc.

Yes, AROS/m68k has all the limitations AROS has in general, of course. The only advantage AROS/m68k does have over AROS for other ISAs is the binary compatibility with AmigaOS/m68k.

Quote:

Aros is aimed to be source-compatible to 3.1 so implementing many of the libraries and thus of course inherits the weak points too. Aros is more than just 3.1, it includes additional components and features, f.e. implementations of AHI, CybergraphX and MUI (Zune), Poseidon (USB) and much more. The 68k branch is binary compatible so you do not need special compiled software and often can mix 68k 3.1 software (f.e. libraries) and Aros components, partly even replacing the Aros components. And yes, the old 3.1 is better suited for the old hardware, nothing beats the original there (and propably never will). Vampire could be fast enough but my impression is that the users want 3.1 and nothing else, even if the other OS would be better.

Quote:

Thanks for your perspective Olaf, I was hoping you would reply to this thread.

I don't get that impression at all, but you would know better. Yes, some of the fans of the Vampire accelerator are hardcore "Only the original is good enough for me" fanatics, but many of the users interested in the Vampire will be more tolerant and be interested to see how AROS 68k performs and what extra features it has. I believe that if/when AROS is significantly better (not that it isn't already), AND it suits the needs of the user on the specific hardware or emulation being used, more people will adopt and use it. For one thing, I am pretty sure Internet browsing will be better on AROS 68k, than it will ever be on AmigaOS3.1. From what I have learned over the years by skimming through news about AROS, it has advanced greatly the past few years, but still has some missing elements, or undesirable traits, which turn off the average Amiga enthusiast. It obviously has certain advantages as well, but only the smaller user group and developers working on AROS fully appreciate it after all these years. I have stated in the past, on many occasions, that AROS is the most likely candidate for long term success, but I could be wrong, and interest in it might fade away before it is ever popular with a majority of current and former Amiga users. I also advocated a merger of MorphOS and AROS programmers, when the switch to a new ISA was to be made, but that obviously isn't happening, and in hindsight, my wish for that to happen was naive and uninformed.

Edit: If users of 68k hardware, FPGA, or software emulation are only interested in playing old games, then the advantages of AROS 68k are wasted on them, and there is little reason to switch to AROS 68k from AmigaOS3.1 (or 1.3).

[ Edited by amigadave 30.05.2016 - 13:55 ]

the problem is that Amiga OS 3.X is unbeatable optimized for the slow 68k hardware with the specific chipsets like blitter. Aros always was obviously targetting much faster hardware (X86). That starts with GCC that is used to compile Aros but creates much bigger (and I do not know slower) binaries because 68k is not well supported there. I also did some testing and it becomes obvious that Aros is slower with less colours whereas AmigaOS behaves the opposite, the reason (as far as I know) is that it calculates 24bit internal so less means more calculations. To get it on the same level speedwise as 3.X certainly would need a lot of work there (when using it on say AGA). How the situation is right now when using it on truecolor is out of my scope because I neither own Vampire nor similar capable 68k hardware. It would be possible to backport components to 3.X of course (one legal way to get Poseidon on 3.X or RTG without P96) but that would require programmers to do the work and I do not see many left and motivated. Aros requires motivated users and developers participating, sitting back in chair and twiddling thumbs waiting for a perfect Aros 68k adaption will not work and not happen. Gutjahr from amiga-news predicted that but I thought different and believed the advantages of a open platform would finally persuade everyone. I was wrong there (we both learn :) ), so I enjoy what I have and am not too much involved anymore. I will be more user in future I guess :), perhaps do a little programming where interested (f.e. Hollywood) but do not try to persuade anyone in future.

there are basic differences between f.e. MorphOS and AROS. AROS always was aimed to be sourcecompatible not binarycompatible so mixing with 68k never was and is possible. Also (if I read right) there was no implementation of UAE before 2009, so it was not really attractive to many, expecially compared to AmigaOS and MorphOS. Most NG users decided themselves years ago and were not interested to change to Aros later, most people stay with what they have or drop amiga at all, people changing platforms are rare from what I have observed. That is even true for AmigaOS and MorphOS, people mostly dropped amiga when their AmigaOS hardware was dead instead of changing to MorphOS despite cheap hardware. The 68k users on the other hand prefer what they have and would never change to anything else. They have their own environments, with specific patches and software, they will never use anything else, they might install something like Aros additionally but not as their main environment and are not interested to become involved.

Then there is of course the world outside but most users there are not interested in amiga and never will.

> (if I read right) there was no implementation of UAE before 2009

"Date: 2001-04-12 [...]
Fabio Alemagna began to port UAE to AROS so we can at least run Amiga binaries slowly.
[...]
Date: 2001-06-03 [...]
Fabio Alemagna [...] improved AROS' UAE port. UAE now works completely. This means that you can run your old software on AROS if a bit slow."
http://aros.sourceforge.net/news/archive/2001.php

More mentions of UAE:
http://aros.sourceforge.net/news/archive/2002.php
http://aros.sourceforge.net/news/archive/2003.php
http://aros.sourceforge.net/news/archive/2008.php

> so it was not really attractive to many, expecially compared to AmigaOS and MorphOS.

Yes, the reason for this has been the different level of integration of m68k emulation. Even o1i's slightly better integrated Janus-UAE with coherency mode starting in 2009 couldn't quite change that.

@Andreas

Thanks for information. I propably read about Janus-UAE and thought it was first port

68k integration should have always been first priority to Aros devs (as the desktop) but obviously not was. More advanced 68k integration beginning with o1e and Janus-UAE and more and better desktops (Magellan and Scalos) a short time ago. Much too late to influence decisions of users and developers.

> 68k integration should have always been first priority to Aros devs (as the desktop)
> but obviously not was.

AROS has always been primarily developed for little-endian x86, which would have clashed with transparently running (as on MorphOS and OS4) big-endian m68k binaries. Emumiga failed for a reason.

Quote:

Most, but not all.
I never owned a legacy Amiga before using MorphOS.
I was part of a company building 68K based systems when the Amiga was still current and I was throughly familiar with the Amiga.
I am just a huge Motorola cpu fanatic.

And I think MorphOS is distinctly superior to AmigaOS.
A better kernel, more modern sound and graphics, and based on a significantly more powerful ISA.

As such, I would like to see MorphOS become what it could be rather than remain in the Amiga's shadow.
After all, its better.

Back to topic.
With its obvious advantages, AROS68K could become quite attractive as the power of new fpga based 68K systems increases.
Since Cloanto can't develop a new OS based on AmigaOS, that leaves AROS and anything Hyperion might want to create in the future.
As for Hyperion, they have shown no interest in backporting.
So...we have one OS that can continue to evolve and another that will require continued patching.

[ Edited by Jim 31.05.2016 - 13:14 ]

@Jim

If I understand it right MorphOS will not have much left from amiga after ISA change, partly legacy components even dropped (I think datatypes as example). But what will it then differentiate from f.e. a changed Linux with different icons and wallpapers?

I doubt that this will bring any new users to be honest, people are not really interested in OS but software and services (or apps how it is called today). So without a real new and modern infrastructure of modern apps and services (up-to-date browser is only one example) nobody new will use it. And for old amigans it will be too different or modern however you call it.

@Jim

you certainly mean Aros68k?

Aros64 would be AMD64 or ARM64

We will see what happens on Vampire. As I wrote at the moment people obviously prefer the old one but that will come to limits

> I think MorphOS is distinctly superior to AmigaOS.
> [...] based on a significantly more powerful ISA.

Version 4 of AmigaOS runs on the same ISA :-)

> AROS64 could become quite attractive as the power of new fpga based
> 68K systems increases.

The Apollo core is advertised as 64-bit core (with 32-bit compatibility), but wouldn't AROS64/m68k (or AROS/m68k-64?) lack compatibility with existing AmigaOS/m68k binaries?

> Cloanto can't develop a new OS based on AmigaOS

I think they are allowed to do this but can't name it "Amiga"-anything.

@Andreas

Aros is compiled with GCC, the same is true for Aros m68k. You would need apollo as compile option there so it is not possible and then you would need all software recompiled for 64bit, not realistic. The old Amiga OS 3.1 is 32bit anyway.

> Aros is compiled with GCC, the same is true for Aros m68k. You would
> need apollo as compile option there so it is not possible

Yes, of course a generator of the proper 64-bit code would be needed as GCC backend in order to support the Apollo core. GCC is open-source, so this would be possible in theorie.

> then you would need all software recompiled for 64bit, not realistic.
> The old Amiga OS 3.1 is 32bit anyway.

Yes, that's what I said.

Quote:

Corrected

Quote:

We don't really know what it will have beyond what the current OS offers.
What it may carry on is a much more efficient structure than Linux uses.
Currently the OS fits on a CD with room to spare.
That just doesn't happen with Linux distros anymore.
I could even see its core pared down further and installed on a rom.
And as far as Linux goes, while I have one system dedicated to it, its not my preference.

So let us see what happens.
No one said PPC MorphOS was going anywhere, just that a more advanced X64 fork was in the works that would alleviate the restrictions inherent with backward compatibility.
I look forward to it.

Btw - Porting a modern browser will be SO much easier with this OS.

I do not doubt that it will be good and on modern hardware it will fly. But people today are not really interested in OS or hardware anymore, that is for geeks "old" retro guys like us who have seen computer in 8bit area progressing (my first computer was a C64). Today even smartphones have 4 cores, you have plenty of powerful and cheap hardware, the priority has shifted from hardware to software, OS is mainly a tool where the apps run. So even if MorphOS is modern in todays terms and runs on new fast hardware it will not change much as long there is no modern infrastructure of supported services, up-to-date browser is just one example (a important one of course). It is very hard to play in that league without the resources, even morphos devs cannot be cloned :) and they do it in spare time. At the same time when changed to much it alienates the "old users" who see it as a updated amiga. It is difficult to satisfy both groups but if you do not manage to attract the new users outside it could be there without any users at all.

In short, we want it, and it will be fun.
What makes you think that appeal to the masses is an important goal?

Quote:

You don't.

@OlafSch

Quote:

From bigfoot:

Quote:

Yeah, where did you get that idea exactly?

Quote:

By reading your posts.
Mark is one of the best programmers in the development team.
I imagine he might have some information about what the new OS will look like that you and I are not privy to.

They are usually pretty tight lipped about future developments.
Consider the statement reassurence that some of your assumptions are incorrect.

And my explaination? An overly verbose description of a concise reply.

Maybe my message got lost somewhere. I meant to ask Olaf why he thinks it has been said MorphOS for another CPU will be totally different from what it is now?

> Maybe my message got lost somewhere.

It didn't with me at least. Seems Jim thought you were OlafSch :-)

Quote:

because the developers like geit say that:

https://morph.zone/modules/newbb_plus/viewtopic.php?topic_id=9204&forum=11#100844

No PPC integration, no 68k except UAE (like done in Aros), incompatible API (you could also say modernized). I did not find the posting about dropping old stuff like datatypes but I read that.

Ambient certainly will still be there but many other things are changed. Or plans have changed... who knows

"SMP, memory protection, 64Bit support, virtual memory will come with a price. Loosing the compatibly to all thirdparty PPC and all 68K applications, which aren´t in development anymore. It also will split the user/market once again. So a step like this should be made wise and preferable only once. And that is the point, where MorphOS swaps hardware. Such opportunity must also be used to drop insane and old crap the OS is filled with to keep it compatible with old Amiga and the old AmigaAPI simply is stupid in many ways."



[ Editiert durch OlafSch 01.06.2016 - 18:20 ]