Everyone already knows that my preference is that the MorphOS Dev. Team continue perfecting MorphOS3.x on PPC only hardware and that they release MorphOS3.x for one or more G5 Mac models, before they decide to move on to some other architecture.

This thread is about WHEN the switch to a different architecture happens, what would you prefer the MorphOS Dev. Team to do? Would you prefer that they just do a straight port of MorphOS3.x to x86 and/or ARM hardware? Or would you prefer to have the MorphOS Dev. Team to break with the old limitations of the Classic Amiga hardware and create a new MorphOS3.x (or 4.x) that has more modern features? Like maybe memory protection, or SMP, or greater than 1.5gb RAM capability? Set up the new MorphOS to include what ever the Q-Box was supposed to have and just run older Amiga and MorphOS apps boxed in a virtual environment, or some other means that allows some backward compatibility, while allowing modern features that will benefit MorphOS in the future.

I know that the Q-Box was abandoned some time ago, but I don't remember, or understand the reasons why. I don't need them explained to me either. A simple "Yes" it is possible for a future MorphOS version to break away and provide some of these more modern features, or "No", the decision was already made years ago that MorphOS will never be anything other than what it is now, no matter what platform it is ported to and running on.

Maybe the Team decided that it is not worth the work and effort to break backward compatibility and create the Q-Box, or something different. Maybe they decided to never make MorphOS into more than it currently is, and prefer to just make it the best it can be in this form instead.

I don't know, maybe there was some announcement made in the past that I don't know about. Sorry I haven't researched it more thoroughly before starting this thread. I am just curious to know what the users of MorphOS3.1 on this forum site think, or what they already know with regard to future plans. Curiosity you know.

Well, you could look at thge reference takemehomegrandma made (or just take his word that ARM is the way to go).

Personally, I'd side with you in that I'd like G5 support (and maybe even e6500 support). And SMP via an additional box does make sense.

The thing is, with two seperate concurrent boxes running, I don't think you'd have to break with compatibility. Abox would just run on one core, while Qbox ran on the remaining cores (kind of a weird combination of SMP and ASMP).

And if an ISA change does ever occur, I'm with those who favor ARM. X86 is just too convoluted.

is there desktop ARM hardware available out there? Is this hardware compatible with established standards for hardware? Is ARM hardware cheaply available?

What benefits does it provide over x86?

These are genuine questions btw, not trying to troll something.

> maybe there was some announcement made in the past that I don't know
> about. [...] I am just curious to know what the users of MorphOS3.1 on this
> forum site [...] know with regard to future plans.

https://morph.zone/modules/newbb_plus/viewtopic.php?topic_id=7930&forum=3&start=54
https://morph.zone/modules/newbb_plus/viewtopic.php?topic_id=8121&forum=9&start=18

> Is ARM hardware cheaply available?

Yes.

amigadave,
Quote:

My POV is within this document: http://via.i-networx.de/q86.htm

Valid Zylesea, but it puts us in direct competition with AROS.
If I had to pick a competitor, I'd rather it be something like AOS4.1.

AROS is free and developing at a steady pace.
Its not as stable as MorphOS, but who is to say what state it will be in by the time MorphOS can be ported to X86?

While AROS is availble for ARM, that's not their primary platform.

We dominate PPC, why not be the best alternative for ARM?

What tangible benefits derive from ARM as opposed to x86, from the end users perspective?

runequester,
Quote:

I'm not sure about the "end user perspective", but I can give you a few benefits.

First, many ARM processors are true Socs. All X86 processors require some companion chipset.

Second, ARM generally draws a LOT less power. Enabling longer run times on battery operated devices as well as more economical/quieter operation on desktop devices.

Third, there are a lot more suppliers as ARM is available via license while X86 is limited to a small number of companies.

> All X86 processors require some companion chipset.

http://en.wikipedia.org/wiki/Vortex86 does not.

Sorry Dave, but who cares ? For all we know the decision might be already made. X86, ARM or just stick to PPC because ( if I am not mistaken ) it's a hobby and they don't really care where it will end with PPC.

> it's a hobby and they don't really care where it will end with PPC.

I don't think so. When it was already pure hobby work, they cared where it would end with Pegasos and Efika 5200B and subsequently ported MorphOS to PPC Macs. Going by that, I guess they also care where it will end with PPC Macs and PPC in general. Just click the links in posting #4 to see evidence of this.

...and at Alchimie, Fab said:

"Once we're done with that, we'll look into an architecture change, either X64 or ARM, whatever is interesting when we get there, which should take a few years"

> ...and at Alchimie, Fab said: [...]

Yes, that's exactly what I'm referring to.

Andreas_Wolf,
Quote:


Interesting. Missed that one, but then I don't think a single core i586 would be all that powerful.
Low power draw though.

Quote:

Please, there is no need for you to tell others what "my word" is (this is not the first time you do this). My point is (was a few years ago) that you shouldn't rule out ARM for what it is now (then), when its development is so fast and there are (were) clear signs that ARM will be something completely different a few years down the line (which it will take for MorphOS to make a jump anyway).

I would love for MorphOS to go x86. My main desktop system has been a Core i7 for ages, and I love it.

That said, I own both incarnations of Efika MX (both the smarttop and the smartbook). They are a bit old now, but they are very nice, and MorphOS would have made a great home even on these machines (HW wise they are running any Sam 440 into the ground, and they were extremely cheap in comparison), and very much has happened since then. Today's fastest ARM CPU's kicks any old PPC based desktop/laptop's butt, especially so when you add the price tag to the equation, and I think it's safe to say that we have yet to see the peak of Cortex A9 generation of CPU's, let alone Cortex A15, and certainly not upcoming 64 bit ARMv8 generations we will see coming from a several manufacturers for years and years to come.

ARM has very nice media centric offerings and is all about efficiency and clever use of resources. Just like MorphOS, and in that sense it's a perfect marriage IMHO. There is however no tradition in the ARM context of providing open and general desktop/laptop computers, and while I am confident that there will always be systems like "Efika MX" available (by Genesi or others), based on *future* ARM CPU models as well, they will be easy to count, and it probably won't ever become comparable to today's x86 market. Question is if that is needed? Look at AROS; it only supports a very narrow (and quite oldish) selection of HW despite the seemingly limitless and constantly expanding supply of cool HW coming out all the time. HW support doesn't come automatically, "x86" does not mean salvation in itself. Drivers and HW documentation will always be a challange. Would a "x86 MorphOS" support the system I have in front of me and use everyday? Hardly, probably not in fact, but that's still what everyone would expect, right?

However, the Linux that Genesi provides with the Efika MX really supports the HW, *including* the proprietary accelerator controllers of the CPU! Hence there are no problems watching HD videos in various codexes, "multi media" is fast, browsing the web is fast enough even with Firefox, etc. All thanks to the drivers provided, which is all thanks to Genesi's close connection to the CPU manufacturer. Maybe access to a company like Genesi is more important than access to x86 per se? Maybe one, two or a few *fully* supported ARM computers over the years would be preferable to an endless and constantly expanding flora of anonymous/proprietary x86 motherboards containing all new combinations/generations of controllers that is very difficult for a small outsider like the MorphOS team to support, and if/when it finally is (through tedious reverse engineering), it has gone out of production since long?

If they really can go x86 in a good way, I would love it. But to me, x86 means top of the line, the edge in power, that's why I buy it. Those systems aren't cheap though, and since I already have one, I would hardly buy a second one (for $2,000+ or so) just because the controllers in my present one doesn't happen to match the ones MorphOS would choose to support. However, buying a second, fully supported $200 ARM box for MorphOS is no brainer though, I'd do that in an instant.

But maybe that's just me...

Quote:

Extremely attractive price/bang for the buck ratio, extremely energy efficient, tiny footprints, completely noiseless, yet perfectly capable of doing what 99% of consumers spends 99% of their computing time to doing; watching HD movies, youtube, listening to music, facebooking, twittering, (video)chatting, e-mailing, playing (simpler) games, office productivity, etc.

Quote:

Yes there are, from more than one supplier, both motherboard or "desktop" systems, as well as portable netbooks. However, this is not even close to be comparable to the x86 market. Besides, most ARM devices are locked, proprietary ones with very limited options of running third party SW. The upcoming Windows 8 ARM devices will be locked, for example.

Oepabakkes wrote:,
Quote:

The point of the whole thread is to find out what the users here in these forums think and what they would prefer to see happen.

That does not mean that the MorphOS Dev. Team will listen to any of us, but it does not hurt anything to express our opinions and discuss the pros and cons of each, here in the forums. It is part of why these forums are here, to discuss MorphOS and it's users and what they would like to have happen in the future.

If the MorphOS Dev. Team is thinking of pursuing "X", or "Y", and 99% of the users would rather see them support "Z" next, the Dev. Team might not change their minds, but at least they will know what most of the users would prefer to see happen.

There is no harm in having discussions about future plans and brain storming ideas between ourselves. Maybe something will be mentioned by one or more of us that no one has thought of before and it will sound good to the Dev. Team, so they will investigate the idea.

I have no false impressions that the decision on which way to go next is in anyone's hands but the Dev. Team members. We as users can either decide to continue following and using what they create, or we can use what we currently have while we look for something else to replace it when ever it quits working in the future.

I think everyone will agree that unless the PPC architecture makes some kind of miraculous recovery, it is dead in the desktop and laptop arenas. This means that the MorphOS Dev. Team will be forced to switch to a new architecture sooner or later. All I was asking in this thread was for members here to express what they would like to see the Dev. Team do when that day arrives, a year from now, or 5 years from now. When is not the question, it is what should they do when that day comes to make a switch away from PPC.

@Zylesea & Jim

I'm going to quote myself here because of pure laziness:
http://www.amiga.org/forums/showthread.php?t=60755

"Judging from what people have requested over the last decade, and also judging from what seems to be the ambition from some OS developers, "Moving Forward" may at some point mean the incorporation of some of the "modern" OS features, like true SMP, true MP, 64-bit, etc. And since this *will* require a clean-slate break from the Amiga legacy anyway (it must happen if you decide to go there), with a clearly defined border line marking the "before" and "after", the seemless "Amiga compatibility" scrapped post that line, starting anew, I must ask the question:

At that point, why continue the PPC path?

If you are to break the "Amiga" anyway, why not do it on some other architecture. My point is that even ARM seems to beat PPC. Performance wise, and from a desktop Point of View, x86 is even more attractive.

Isn't a platform migration the natural thing to do at that point? There is no "Power" in PowerPC, not in the year 2012 and beyond! ARM and/or x86, but not PPC!

Right? "

A clean slate would be required to incorporate features like these and make an architectural jump in a "fairly easy" way. This will probably also mean handling 68k apps in a similar way as AROS does (by their way of using UAE together with various scripts or whatever, to automatically/seamlessly run each app in it's own little "UAE box" (meaning 1 app = 1 emulated Amiga computer)).

I think this approach would be totally acceptable for a 4.x branch of MorphOS, incorporating all those new interesting core OS features (more addressable memory, true SMP, true MP, 64-bit, etc, that simply *can't exist* in a legacy Amiga context without breaking the Amiga part of it (no matter what some people claim)), and do an architectural jump at the same time.

I also hope that one of the project goals for a MorphOS 4 would be to explicitly work towards making the OS (and SDK) more easily portable or even fully ISA independent altogether.

"we'll look into an architecture change, either X64 or ARM, whatever is interesting when we get there. [...] Meanwhile, we get into the 64 bit compatibility to prepare the ground. We will make it run in a virtual machine [...]. That's for development [...] so that we can start working on it without having a machine available yet."

http://www.amiga.org/forums/showthread.php?t=59765

Hmm... Develop now, choose architecture sometime later...

> I own both incarnations of Efika MX (both the smarttop and the smartbook).
> [...] HW wise they are running any Sam 440 into the ground

Just to illustrate what that statement means specifically, let me put this link to a benchmark comparison here:

https://morph.zone/modules/newbb_plus/viewtopic.php?forum=3&topic_id=7675&start=164

I know that the poor floating point result of the Efika MX was due to softfloat being used. Now that "the Linux that Genesi provides with the Efika MX really supports the HW", would you please run the nbench benchmark on your Efika MX with hardfloat enabled, so the comparison becomes more fair?

> Today's fastest ARM CPU's kicks any old PPC based desktop/laptop's butt

Today's fastest ARM CPUs are still slower than the PPC970 CPUs in most (i.e. 2.0+ GHz) G5 desktop machines. Geekbench example:

Dual-core 2.5 GHz PowerMac G5: 2366 points
Quad-core 1.4 GHz ARM Cortex-A9: 1811 points
Dual-core 1.3 GHz Apple A6: 1572 points
Dual-core 1.5 GHz Qualcomm Krait: 1540 points
Dual-core 1.4 GHz ARM Cortex-A9: 1229 points

http://browser.primatelabs.com/geekbench2/622225
http://browser.primatelabs.com/android-benchmarks
http://browser.primatelabs.com/ios-benchmarks

> we have yet to see the peak of Cortex A9 generation of CPU's

The world is already turning to Cortex-A15 performance generation of ARM CPUs with the Qualcomm Snapdragon S4 (Krait core) and just recently the Apple A6 (custom core). I know that there're vendors behind the curve with their first Cortex-A9 chips still to come to mass market (like Freescale with i.MX6), but I doubt somewhat that they will be able to beat current Cortex-A9 chips in performance.


Edit: corrected A6 clock rate.

[ Edited by Andreas_Wolf 08.10.2012 - 21:54 ]

Quote:

I know that making and maintaining these charts is your hobby Andreas, so please don't hate me for not bothering with this. I don't have the systems at hand, and both Efika MX (I don't know if Genesi still has any left in stock) and Sam 440 has lost its relevance anyway. My point wasn't about raw cpu power either, but about the impact of the clever use of integrated accelerator controllers for performance boosts, and the amount of real life usability you get in comparison to its low price. I think I paid $129 (or so) for my Smarttop, and it's a magnitude more usable than this guy's $1,500 AmigaOne 500 (yes I know, that's 460 and not 440, but anyway). No need to fall back to using Ibrowse for speed, Firefox works great, no need to recode videos to a different codec/resolution to be able to watch them, the Efika MX CPU happily decodes and renders most codecs in 720p HD (it actually can decode 1080p streams but not render them at a 1080 HD resolution screen (only 720) due to a stupid bottle neck in the CPU that was removed in the i.MX53 version). All this on a rather bulky Linux distro that doesn't come close to MorphOS's lean elegance.

Quote:

Yes I know, but this is a fact that may very well be changing as we speak, at least if we speak slooooow enough...

(But it happened quite some time ago if you consider the *entire* meaning of my sentence that you quoted: "Today's fastest ARM CPU's kicks any old PPC based desktop/laptop's butt, especially so when you add the price tag to the equation" )

Quote:

This was my point exactly, far from every chip manufacturer settles with just the bare core reference design provided by ARM, many consider that to be merely a starting point and pour a lot of engineering into their own chip design to pimp it up significantly in various ways. We have seen that many times with older ARM core designs, and this will of course continue in the future as well with new core designs.

>>> All X86 processors require some companion chipset.

>> http://en.wikipedia.org/wiki/Vortex86 does not.

> Interesting. Missed that one, but then I don't think a
> single core i586 would be all that powerful.

There's also this since April 2012:
http://en.wikipedia.org/wiki/Atom_(system_on_chip)

Compared to Power Architecture SoCs, those Atom SoCs seem to lack things like SATA and Ethernet, but then they have 2D+3D GPU.

>> would you please run the nbench benchmark on your Efika MX with
>> hardfloat enabled, so the comparison becomes more fair?

> please don't hate me for not bothering with this. I don't have the
> systems at hand

Pity. But I can live with that I guess ;-)

>>> Today's fastest ARM CPU's kicks any old PPC based desktop/laptop's butt

>> Today's fastest ARM CPUs are still slower than the PPC970 CPUs in
>> most (i.e. 2.0+ GHz) G5 desktop machines.

> it happened quite some time ago if you consider the *entire* meaning
> of my sentence that you quoted: "Today's fastest ARM CPU's kicks any
> old PPC based desktop/laptop's butt, especially so when you add the
> price tag to the equation" ;-))

Appending "especially when..." to a statement means the statement is true also without the appendix, and *even more so* with the appendix. Your statement however is false without the appendix, so what you probably meant to say (I can't be sure though) is "as long as" instead of "especially when".

> far from every chip manufacturer settles with just the bare core
> reference design provided by ARM

Yes, and we can even quantify it exactly: regarding ARMv7, there're exactly 3 companies that do their own core(s) from scratch instead just licensing them from ARM Ltd. These are Marvell (Sheeva PJ4/PJ4B), Qualcomm (Scorpion, Krait) and Apple (A6 core).
Btw, I don't think "reference design" is an apt term for what is happening here. It's not like ARM ISA licensees take a core design from ARM Ltd. and try to enhance it or adapt it for their needs. In fact, they develop their own microarchitecture from scratch with the ARM ISA as specification which the microarchitecture is an implementation of. This is in contrast to the ARM core licensees who take and use cores from ARM Ltd. as is.

Yes the business model of licensing IP to others to allow/enable them to make their own CPU's has definitely been key to ARM's success, and will to a major part be the explanation to the upcoming booming of the architecture (it may seem like it already has happened, but I think we have seen nothing yet). It's a success saga, and it's only at the beginning.