@Raffaele

Quote:

Raffaele wrote: @OlafS25 Quote: OlafS25 wrote: @cdimauro I talk about AROS, MorphOS and AmigaOS. They have the same (or better similar APIs). So it is possible (at least theoretic). And it is at least partly possible if wanted. One developer uses AROS to make his MorphOS ports. How nice... MorphOS... A pure microkernel OS that has modern (but still not yet developed) features, must run a Linux Kernel? Full OS? with some flavour of AROS (API? Hosted? Wrapped?) to obtain the same things that could obtain by just developing its QBox part hidden into its guts... It seems a kafkian (*) joke of some sort... (*) Kafkian = adj. Referred to Franz Kafka surreal stories.

No, it doesn't seems... it's that, as usual, you have understood nothing of the discussion.

How nice... will be if you first have read carefully and at least a sufficient comprehension before writing.

Last edited by cdimauro on 09-Nov-2013 at 06:27 AM.

@cdimauro

Quote:

cdimauro wrote: @Raffaele No, it doesn't seems... it's that, as usual, you have understood nothing of the discussion. How nice... will be if you first have read carefully and at least a sufficient comprehension before writing.

Duh! And it seemed to me that I was not talking with you, but just answering with Olaf...

And seems to me he hope it could be added to MorphOS (in line with AROS and AmigaOS4) certain features using this new linux kernel + AROS 68K...

Am I wrong?


However it seems also that you did not read my previous comments and where I want to point to...

As I previously pointed that MorphOS has all the features that a modern OS like Linux has nowadays like SMP, mutlithreading, multiprocessing, Protected memory and so on (obviously except a vaste range of hardware drivers), but the MorphOS team never developed any API to make QBox active, or if they did, they left it completely hidden to public and keep it for themselves.

[EDIT]

Edited to avoid flamewars...

[EDIT END]

Last edited by Raffaele on 09-Nov-2013 at 08:53 AM.
Last edited by Raffaele on 09-Nov-2013 at 08:52 AM.
Last edited by Raffaele on 09-Nov-2013 at 08:50 AM.
Last edited by Raffaele on 09-Nov-2013 at 08:47 AM.
Last edited by Raffaele on 09-Nov-2013 at 08:46 AM.
Last edited by Raffaele on 09-Nov-2013 at 08:42 AM.
Last edited by Raffaele on 09-Nov-2013 at 08:41 AM.

_________________
"When the Amiga came out, everyone [at Apple] was scared as hell." (J.L. Gassée, former CEO of Apple France and chief of devs of Mac II-fx, interviewed by Amazing Computing, Nov 1996).

Olaf talked and asked for a completely different thing. Again: read back the messages, because they are pretty clear about it.

Regarding MorphOS, who cares of the QBox? It's not used! And, more important, don't think that the QBox presence means that ABox can take advantage of its features. It's like asking WinUAE to make use of SMP, memory protection, and resource tracking to improve the Amiga emulation just because it runs on Windows which makes them available...
IMO running a micro-kernel on top of micro-kernel doesn't make sense, especially if the former doesn't make a great use of the latter. To be more clear, it's better to kill QBox and make ABox the real and only microkernel running for MorphOS.

@cdimauro

Quote:

cdimauro wrote: Olaf talked and asked for a completely different thing. Again: read back the messages, because they are pretty clear about it.

I prefer to wait Olaf to confirm if I understood his latter statement or If did not...

Quote:

Regarding MorphOS, who cares of the QBox? It's not used! And, more important, don't think that the QBox presence means that ABox can take advantage of its features. It's like asking WinUAE to make use of SMP, memory protection, and resource tracking to improve the Amiga emulation just because it runs on Windows which makes them available...

Completely misunderstood...

Just let's make QBox the main box running in the system, booting from it and running its own API, GUI and software, complete of MMU, Posix compliant, SMP and Multithreading and live happy... and leave ABox to became only a sandboxed Amiga emulation enviroment... and when it crashes ABox then QBox apart of ABox will be no affected and you could continue running the system... And users just need to let MorphOS rebooting ABox only, with all its Amiga limits and legacy.

Quote:

IMO running a micro-kernel on top of micro-kernel doesn't make sense, especially if the former doesn't make a great use of the latter.

But indeed it is what MorphOS team made with ABox running Exec wrapped in an environment of its own on top of Quark Microkernel.
It is a very elegant and practical solution. It is the solution we all Amigans are trying to implement in order to obtain a complete and modern system without renouncing Amiga...
But without the development of the pure Microkernel Quark box, it remains only a peg-leg solution.

Quote:

To be more clear, it's better to kill QBox and make ABox the real and only microkernel running for MorphOS.

So let's continue using ABox with all Amiga legacy and limits, included the common space memory one address only environment and be happy when a program overwrites memory allocated by another and the entire system crashs pityfully...

(and a rebooting time of 5 seconds is not a benefit when you lost the entire jobs you are working on before the crash)

Last edited by Raffaele on 09-Nov-2013 at 10:36 AM.

_________________
"When the Amiga came out, everyone [at Apple] was scared as hell." (J.L. Gassée, former CEO of Apple France and chief of devs of Mac II-fx, interviewed by Amazing Computing, Nov 1996).

What you ask is already existing. It's exactly like Windows, Linux, MacOS X, ecc. running the old software into the (Win)UAE sandbox...

@cdimauro

Quote:

cdimauro wrote: What you ask is already existing. It's exactly like Windows, Linux, MacOS X, ecc. running the old software into the (Win)UAE sandbox...

No I don't ask for it... WinUAE it is only emulating classic Amiga up to OS 3.9.

I want also using PPC programs like OWB, and also run good things a modern MUI like MUI 4 with MMU features and SMP in QBox, modified to take care of new Microkernel API features (it requires just some minor changes and a recompile if QBox will be well engineered)...

... and parallely running all ancient Amiga RTG programs like TVPaint running in ABox where they will be free to crash and reboot wholenABox.(as long as when QBox will became the main environment of MorphOS then ABox will be hosted in a window of its own).

E-UAE then could be used from QBox or either ABox (I don't care where) and it will be necessary only for playing games and some extensive Chipset-using software, like perhaps as DeLuxe Paint.

Last edited by Raffaele on 09-Nov-2013 at 11:18 AM.
Last edited by Raffaele on 09-Nov-2013 at 11:06 AM.

_________________
"When the Amiga came out, everyone [at Apple] was scared as hell." (J.L. Gassée, former CEO of Apple France and chief of devs of Mac II-fx, interviewed by Amazing Computing, Nov 1996).

@Raffaele

You misunderstood me. It was not about all using the same kernel or all using Linux. It is about doing the same tasks three times (mostly only slightly different). Imagine if all use the same drivers for USB, PCI and so on so drivers could be used on all systems. That would make it easier to support all platforms. Or using the same APIs and work on the development together. That (indirectly) already happens on some libraries between MorphOS and AROS. If that would be extended and would happen between all three camps all would benefit (including the software-developers). Regarding MorphOS I do not want to comment it further only a general thing... being "too proud" sometimes can become a problem either. It reminds of the chinese being the middle of the world (in their thinking) and superior to the west until new gunships appeared at the coast and they finally discovered reality. Sometimes it is better not to feel "superior" and better work together in a common interest. Think about it.

@Raffaele

I have done some google. QBox is certainly a interesting idea but there is no indication that there is any work done on it in recent times (the MorphOS team is small too and they have updated MorphOS (ABox) and added a lot of hardware). So you talk about something that is at the moment propably years away (if they decide to port MorphOS on different hardware). You can dream of a lot of things but at the moment it is just theory.

ARIX will be reality in a couple of days and people can judge it then. Projects like AEROS (and a future version of AROS Vision) that are using Linux for driver support and applications are real too. Linux has a huge driver base and that alone is a huge benefit. Even if MorphOS has a good kernel (even if it would be better than the Linux kernel) they still have to develop every driver and that is a huge disadvantage.

Last edited by OlafS25 on 09-Nov-2013 at 11:26 AM.

@OlafS25

Quote:

OlafS25 wrote: @Raffaele You misunderstood me. It was not about all using the same kernel or all using Linux. It is about doing the same tasks three times (mostly only slightly different). Imagine if all use the same drivers for USB, PCI and so on so drivers could be used on all systems. That would make it easier to support all platforms. Or using the same APIs and work on the development together. That (indirectly) already happens on some libraries between MorphOS and AROS. If that would be extended and would happen between all three camps all would benefit (including the software-developers).

Right... That's better explaining of your thoughts and the general view of it all...

But then there is another solution also except Linux to obtain a common environment or re-inventing the wheel again.

It is a third way to approach a common environment between AmigaOS, AROS and MorphOS.

It lays under the eyes of everyone and nobody noticed it...


Trevor Dickinson, instead of focusing on new PPC machines with a limited market, spending more than one million dollars could had offered MorphOS Team a good amount of money for their microkernel and freeing it to make Quark available on AmigaOS and AROS too, so then all Amiga-like exising Operating System could benefit of a small efficient Kernel complete of modern features, and we all had gained a common background environment free from Amiga Legacy and free from Linux enormous monolithic kernel.
Once Quark Microkernel was free, then every Amiga camp could had developed Microkernel own API to evolve its own OS of reference, developing a common set of hardware drivers that share the same microkernel inheritance, and/or even decide to porting it on other architectires such as X86 or ARM Cortex...

Quote:

Regarding MorphOS I do not want to comment it further only a general thing... "too proud" sometimes can become a problem either. It reminds of the chinese being the middle of the world (in their thinking) and superior to the west until new gunships appeared at the coast and they finally discovered reality. Sometimes it is better not to feel "superior" and better work together in a common interest. Think about it.

It's not a matter of "superiority"... As long as everyone is complaining that Amiga to obtain new features should abandon old legacy system and embrace Linux as it grants all innovations without re-inventing the wheel again and again in all Amiga camps, I just asked myself "WHY?"...

Just remember we already have the common solution. We already have the solution and it is called Quark Microkernel.

Last edited by Raffaele on 09-Nov-2013 at 11:54 AM.
Last edited by Raffaele on 09-Nov-2013 at 11:52 AM.

_________________
"When the Amiga came out, everyone [at Apple] was scared as hell." (J.L. Gassée, former CEO of Apple France and chief of devs of Mac II-fx, interviewed by Amazing Computing, Nov 1996).

@Raffaele

I do not know if MorphOS team would have been interested at all and not judge the kernel as one of their unique advantages compared to AmigaOS or AROS (similar to MUI4 f.e.). I am also the opinion that it is a waste of money to invest in new PPC hardware and instead the money should have been invested in f.e. porting the OS to a more common hardware platform. But that is Trevor D. money and decision. But even "if" the kernel would be open and all would use it there would still be the problem with missing drivers (and specialists that are able to write drivers are rare in the community). So something based on Linux is the best bet today to get more hardware support. But people should wait "a little than just one more week" and then make judgements.

Quote:

Raffaele wrote: @cdimauro Quote: cdimauro wrote: What you ask is already existing. It's exactly like Windows, Linux, MacOS X, ecc. running the old software into the (Win)UAE sandbox... No I don't ask for it... WinUAE it is only emulating classic Amiga up to OS 3.9. I want also using PPC programs like OWB, and also run good things a modern MUI like MUI 4

Even the PowerPC can be emulated. In fact, it happens in other emulators.

Quote:

with MMU features and SMP in QBox, modified to take care of new Microkernel API features (it requires just some minor changes and a recompile if QBox will be well engineered)...

So do you want new applications running natively on QBox, based also on libraries like MUI 4? It requires a lot of effort both to update/enhance QBox and to the developers too.

Quote:

... and parallely running all ancient Amiga RTG programs like TVPaint running in ABox where they will be free to crash and reboot wholenABox.(as long as when QBox will became the main environment of MorphOS then ABox will be hosted in a window of its own). E-UAE then could be used from QBox or either ABox (I don't care where) and it will be necessary only for playing games and some extensive Chipset-using software, like perhaps as DeLuxe Paint.

That means exactly the same thing that happens with (Win)UAE running on Windows, Linux, and MacOS X.

Why should I care about QBox doing the same thing, but with far less drivers?

Quote:

OlafS25 wrote: ARIX will be reality in a couple of days and people can judge it then. Projects like AEROS (and a future version of AROS Vision) that are using Linux for driver support and applications are real too. Linux has a huge driver base and that alone is a huge benefit.

The same happens with Windows: many many drivers and applications are already available. And albeit not so much stable, AROS runs hosted on Windows too...

Quote:

I do not know if MorphOS team would have been interested at all and not judge the kernel as one of their unique advantages compared to AmigaOS or AROS (similar to MUI4 f.e.).

In fact buying QBox and making it free doesn't make sense. Even because free microkernels already exists.

Quote:

I am also the opinion that it is a waste of money to invest in new PPC hardware and instead the money should have been invested in f.e. porting the OS to a more common hardware platform.

I agree.

Quote:

But that is Trevor D. money and decision. But even "if" the kernel would be open and all would use it there would still be the problem with missing drivers (and specialists that are able to write drivers are rare in the community). So something based on Linux is the best bet today to get more hardware support.

But I disagree here.

There are already plenty of Linux distros, which can run the old 68K software on top of them...

@cdimauro

Of course... if that is all we would need nothing. Pascal is creating a version of Aros Vision that runs from a USB-Stick and directly starts FS-UAE based on a Linux-distribution. But people want use not only 68k software but also new native. But as I said "just one more week"

@cdimauro

Quote:

cdimauro wrote: Even the PowerPC can be emulated. In fact, it happens in other emulators.

Actual Speed of PearPC PPC emulation on an X86 is 1:13. If you like to emulate a 1,00 GHZ processor G4 at clock speed of 76 MHz then you are welcome, and be happy with it...

Quote:

Quote: Raffaele wrote: ... and parallely running all ancient Amiga RTG programs like TVPaint running in ABox where they will be free to crash and reboot wholenABox.(as long as when QBox will became the main environment of MorphOS then ABox will be hosted in a window of its own). E-UAE then could be used from QBox or either ABox (I don't care where) and it will be necessary only for playing games and some extensive Chipset-using software, like perhaps as DeLuxe Paint. cdmauro wrote: That means exactly the same thing that happens with (Win)UAE running on Windows, Linux, and MacOS X.

We all know your passion for Classic Amigas, but unfortunately the best part of amigans on these sites are evolving their preferences from the "classic chipset" paradigma and prefer far better the PPC and NG systems, and this does not means we are stuck on just PPC systems.
So we do not care of WinUAE and playing tiny Amiga games on emulated 68xxx systems.
Thanks, but no, thanks.

Drivers are the last of our problem actually.

First step is to integrate well new Operating Systems with the existing hardware that at this moment is still not supported entirely. Look at AmigaOS and its lack of USB 2 support, for example.

Only when all things will starting working flawlessly and the drivers and the 3D engines will be integrated fully in AmigaOS, resulting in no bugs, then and only then we could think of changing actual architecture, and migrating to X86 and ARM, or else we will encounter again the same problems even being in possess of all kind of drivers available for other systems...

Last edited by Raffaele on 09-Nov-2013 at 03:50 PM.
Last edited by Raffaele on 09-Nov-2013 at 03:48 PM.
Last edited by Raffaele on 09-Nov-2013 at 03:48 PM.
Last edited by Raffaele on 09-Nov-2013 at 03:47 PM.
Last edited by Raffaele on 09-Nov-2013 at 03:45 PM.
Last edited by Raffaele on 09-Nov-2013 at 03:43 PM.

_________________
"When the Amiga came out, everyone [at Apple] was scared as hell." (J.L. Gassée, former CEO of Apple France and chief of devs of Mac II-fx, interviewed by Amazing Computing, Nov 1996).

@Raffaele
pearpc compares well to bochs... it is slow

Try dolphin emulator as good example of ppc emulation.

_________________
AROS Broadway - AEROS - Aminux - AmiCloud - indieGO! Appstore - AmiWallet - VAN lossless video codec - AMC Amiga media Center -KrypUnite - LibertyNet - MinX - amigaNX

@phoenixkonsole

Quote:

phoenixkonsole wrote: Try dolphin emulator as good example of ppc emulation.

Or, get a real PPC and screw the emulation. zoom-zoom

@thread

Should Linux kernel power future AOS solution ?

NO!

AmigaOS, or Amiga OS, or AOS, is moving along just fine.

The really interesting thing about AmigaOS is what will evolve as it moves to
multicore operation.

To not travel down this road would be unforgivable, unless of course you're
afraid of falling off the edge. 'Oh no Mr Bill. Here comes the edge. Oh NOoooo.'

_________________
Tinkering with computers.

@Signal

what "multicore" support do you mean? It will break 68k compatibiliy

and when you see that the road is going more and more in the desert and noone around anymore you should change direction

Last edited by OlafS25 on 09-Nov-2013 at 05:15 PM.

Quote:

Raffaele wrote: @cdimauro Quote: cdimauro wrote: Even the PowerPC can be emulated. In fact, it happens in other emulators. Actual Speed of PearPC PPC emulation on an X86 is 1:13. If you like to emulate a 1,00 GHZ processor G4 at clock speed of 76 MHz then you are welcome, and be happy with it...

It doesn't mean that ALL PowerPC "emulations" will be slow. Take a look at the Rosetta results: http://www.macworld.com/article/1049163/intelimacs.html

"ROSETTA iMac Core Duo/2GHz iMac Core Duo/1.83GHz iMac G5/2.1GHz
Adobe Photoshop CS2: Suite 2:50 3:02 1:16
Microsoft Word Scroll: Scroll 1:58 2:05 0:57"

As you can see, an OLD 2Ghz Core Duo runs at about HALF (not 1/13!) the speed of a 2.1Ghz G5.

Consider that:
- they are old (January 2006!) machines;
- both the o.s. (OS X Tiger) and Rosetta ran in 32-bit mode (x86).

And that now we have:
- double the clock;
- more efficient micro-architectures (Nehalem and successors);
- the new CPUs with new instructions which can help emulating big-endian machines (thanks to the very welcome MOVBE) and Altivec (AVX is a much more efficient and flexible SIMD unit/ISA; the coming AVX-512 will be MUCH better);
- 64-bit architectures and software, which can GREATLY enhance both the emulation and the quality of JIT generated code.

As a very rough and conservative estimation, I think that the performance can be 3 times better.

So, didn't you like a 3Ghz G5 performance? That's for a newer Rosetta version, OK, but I think that a new project can achieve good results, so having a good user experience.

Quote:

Quote: cdmauro wrote: That means exactly the same thing that happens with (Win)UAE running on Windows, Linux, and MacOS X. We all know your passion for Classic Amigas,

You're wrong: my passion is for the Amigas. I don't know of any "Classic" Amigas.

Quote:

but unfortunately the best part of amigans on these sites are evolving their preferences from the "classic chipset" paradigma and prefer far better the PPC and NG systems, and this does not means we are stuck on just PPC systems. So we do not care of WinUAE and playing tiny Amiga games on emulated 68xxx systems. Thanks, but no, thanks.

And this might be even a better situation. Newer applications have much better chances to be recompiled natively to the new platform/architecture, so running faster than a JIT-translated PowerPC application.

There'll be a transition period to have almost all applications ported, such as happened to MacOS X x86.

Quote:

Drivers are the last of our problem actually. First step is to integrate well new Operating Systems with the existing hardware that at this moment is still not supported entirely. Look at AmigaOS and its lack of USB 2 support, for example.

And isn't the USB 2 support... a DRIVER implementation? It took 7 years to have it, and you have the bravery to say that the drivers aren't the last of the problems now?

Quote:

Only when all things will starting working flawlessly and the drivers and the 3D engines will be integrated fully in AmigaOS, resulting in no bugs, then and only then we could think of changing actual architecture, and migrating to X86 and ARM, or else we will encounter again the same problems even being in possess of all kind of drivers available for other systems...

OK, so may be my grandchildren will have a chance to see it.

@Olaf: I wait. But I don't like Frankensteins, as I said before.

@OlafS25

Quote:

OlafS25 wrote: @Signal what "multicore" support do you mean? It will break 68k compatibiliy

And (old) PowerPC code too, IMO. 68K applications are "first class" applications, much like the PowerPC ones...

The same compatibility breaking will happen with 64-bits and memory protection.

Quote:

and when you see that the road is going more and more in the desert and noone around anymore you should change direction

With the hope to don't take a bad new one...

@OlafS25

Quote:

OlafS25 wrote: @Signal what "multicore" support do you mean? It will break 68k compatibiliy and when you see that the road is going more and more in the desert and noone around anymore you should change direction

Or build a city.

What can I say?

You guys are the AmigaOS Next Gen experts. You are also very good at deciding which path
others should take.

_________________
Tinkering with computers.