I remember the old X86 v. PPC wars of years ago for the NG Amiga. When they started, both the X86 and the PPC had a good future, although the PPC was moving to a narrower niche market.

Now, they are both entering the niche phase with the demise of the dominant desktop market.

Since the dramatic increase of the X1000 PPC chip prices, I think we need to explore the alternatives. At the moment, the dominant processor architecture is ARM. It is renamed in various guises for Apple, Android and other devices, but its application is expanding. There are already four-core plus graphics chips in general use. There are proposals to produce a low-power server solution based on ARM (just as there were earlier PPC and Atom proposals). There are, I understand, more devices with ARM processors than any other.

It brings with it a wide range of open source software and device drivers. However, there are few general ARM platforms. Even the Raspberry Pi is not open source hardware and software (because of the Broadcom chip).

I think that OS4 should be ported to an ARM system. There is room for both proprietary hardware platforms and the opportunity to repurpose those Android platforms that will eventually allow the free choice of Linux distributions.

Microsoft and Intel have more or less put paid to the ARM laptop/notebook, Apple have proprietary, closed ARM systems. Google are promoting Chromebooks and Android tablets that mght be repurposed for the Amiga OS. There is no pile 'm high, sell 'em cheap general purpose ARM platform equivalent to the desktop PC motherboard. There is room for a specific Amiga desktop implementation, al la X1000. Amiga OS could even be ported to the Raspberry Pi as have other (mainly Unix/Linux) operating systems.

What do you think? Please; no processor rants. We've already done that!

Regards
Roy Leith

Last edited by royleith on 16-Mar-2013 at 10:50 AM.

@royleith

The need to free AmigaOS 4 from PPC is so obvious that I can imagine only two reasons because it hasn't been ported to X86 years ago or to arm more recently.

1) It's simply too difficult for Hyperion's developers. "Not worthy" is not an answer, because anything that increases by an order of magnitude the chances of your product being run, IS worthy.

2) Hyperion has an under the table agreement with one or more PPC hardware makers and they make money on that hardware being sold, too - not just on OS4 sales. If you think about it, there's no justification for buying, say, an X1000 - if not for running OS4.

@royleith

If there was a processor architecture switch I'd still rather see x86 than ARM. I don't think x86 desktops are going away anytime soon and I've yet to see an ARM alternative desktop system.

x86 also makes sense because the same video, audio and network hardware is available on that architechture. On ARM it is always hardware that is currently alien to Amiga OS.

@Rob

Arm has much more sense:

1- You can build very inexpensive entry level hardware with very well specs
2- there is not (not much at least) desktops computers with Arm so it's a niche to conquer; you know, who strikes first strikes twice. There is more sense for and custom dektop Arm than a custom desktop x86
3-Moving to Arm means to have a step more towards being mobile/tablet (a huge market compared to desktops)

Anyway I'm a PowerPC fan and think that we can make the same with PowerPC but if we have to switch i think that Arm is better for the reasons that i hace write above

_________________
No PowerPC, No Fun
Make Amiga Great Again

@resle

Quote:

resle wrote: @royleith The need to free AmigaOS 4 from PPC is so obvious that I can imagine only two reasons because it hasn't been ported to X86 years ago or to arm more recently. 1) It's simply too difficult for Hyperion's developers. "Not worthy" is not an answer, because anything that increases by an order of magnitude the chances of your product being run, IS worthy. 2) Hyperion has an under the table agreement with one or more PPC hardware makers and they make money on that hardware being sold, too - not just on OS4 sales. If you think about it, there's no justification for buying, say, an X1000 - if not for running OS4.

So what is the MorphOS camps excuse after all they have been in the PPC game longer than Hyperion as far as Amiga Next Gen?

I think it has nothing to do with either of your suggestions, I think originally it was a decision on guessing what platform had the brighter future and they guessed wrong, then I think it was about being more compatible at the API level for classic WB3.9 and prior software.

I also think that Hyperion or MorphOS camp don't have the money to spend on developing and marketing a total port and redevelopment of the respective operating systems to run on x86 architecture.

_________________
Amiga A1200T - TF1260 - R9200 - Indivision AGA MK3
Amiga A500 - PiStorm EMU68
Pegasos 2 G4 - AmigaOS 4.1 FE / MorphOS 3.16

@royleith

http://mac1.no/artikkel/3107/ukens-apple-reklame-toasted-bunnies
http://www.osnews.com/story/3997
http://www.youtube.com/watch?v=WQXvnWfOKlA

Last edited by NutsAboutAmiga on 16-Mar-2013 at 12:26 PM.
Last edited by NutsAboutAmiga on 16-Mar-2013 at 12:24 PM.

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

@royleith

The problem with ARM hardware is that the interesting parts of some designs are not public and hardware vendor supplies only some binary objects. The biggest problem with ARM hardware are drivers since there's no common low level firmware in most public devices. Just check out mobile phones, these phase out in a year or so, no time to write drivers for that.

I prefer x86 hardware as there's more information available to write drivers and these are compatible with various machines, and x86 hardware shares common firmware. AROS for x86/x64 kinda runs on top of everything. If you think you have problems with driver support on AROS x86 these would be elevated by 3 if you used ARM hardware as information is scarce, there are various revisions, there are models without fpu, there are no many desktop machines, these are not as easy to buy as x86 hardware and hardware versions get obsolete faster. Moreover, any x86 hardware runs rings around ARM hardware for many years. x86 has more JITs and better support.

E.g.: Raspberry Pi struggles to play DVD-like MPEG2 if you don't use binary blobs supplied by vendor to use the video acceleration. These binary blobs are usually designed for linux.

All in all: I want a desktop computer. I want it to be fast, cheap and powerful. Easy to replace by something compatible too. If I buy a obscure ARM platform today, it probably won't exist in 2 years and my drivers won't work there. In contrast in x86 you can use the same PCIe gfx card, you have BIOS/EFI and everything is so fast that lack of GPU/video documentation can be solved by brute force. ARM cpus are quite weak.

_________________
The only spanish amiga news web page/club: CUAZ

@royleith

It was expected at one point the design of the Pentium, mede it impossible to go higher clock frequency, at later stage the x86 family started to become more Risk like.

the 680x0 Family is a big endian architecture.
The PowerPC Family is also a big endian architecture.
But the x86 is little endian architecture.

So what is the problem?

The PowerPC put number in memory like we read them

in memory its stored 1000 is and value for the number is 1000,
but on x86 if the number stored as 10000 then the value is 0010,

the x86 and 68000/PowerPC don't keep the bytes in the same order, this complicates everything, shore Apple has Rosetta, to convert the code, like we have petunia, but it way more complicated.

So basically you lose the possibility to run 68000 programs on AmigaOS4 if it was ported to x86, what about ARM, well its be endian architecture it means that assembler code can be little or big endian, ARM is not as fast and powerful as PowerPC architecture yet, is for mobile devices, thats cool, and it has video hardware acceleration built in, many of ARM devices out there is faster then some of the low end computers the Sam440 for example, maybe also the Sam460.

Now 100000 euro question should the time be spent on porting AmigaOS to ARM or should the time be spent on updating AmigaOS to support SMP?

Note there are many PowerPC processors that are sheep, and not that expensive yet, they are after all used in many thing you do not think of as computers, open up your PS3 or Xbox, look inside your DVD/BlueRay or network router, you find that chips are every where, pa semi chip was horrible choice as was out of production before we started to use it, and also in think X1000 too way to long to get to market this also possibility made it even harder to get pa semi chips at nice price.

A architecture switch will ether require more JIT's to be written or that every one recompiles there programs for two architectures. You might end up in possibly where ARM based computer can't run PowerPC programs or PowerPC computer can't run Arm based programs, basically split the tiny marked we have.

Last edited by NutsAboutAmiga on 16-Mar-2013 at 01:12 PM.
Last edited by NutsAboutAmiga on 16-Mar-2013 at 01:10 PM.
Last edited by NutsAboutAmiga on 16-Mar-2013 at 01:06 PM.
Last edited by NutsAboutAmiga on 16-Mar-2013 at 12:56 PM.
Last edited by NutsAboutAmiga on 16-Mar-2013 at 12:53 PM.

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

either ARM or x86 would be good. But, sadly, it will never happen for OS4, which will continue on PPC on life support untill somebody pulls the plug or the patient dies.

Last edited by WolfToTheMoon on 16-Mar-2013 at 01:11 PM.

_________________

@Fransexy

Quote:

Arm has much more sense: 1- You can build very inexpensive entry level hardware with very well specs 2- there is not (not much at least) desktops computers with Arm so it's a niche to conquer; you know, who strikes first strikes twice. There is more sense for and custom dektop Arm than a custom desktop x86 3-Moving to Arm means to have a step more towards being mobile/tablet (a huge market compared to desktops) Anyway I'm a PowerPC fan and think that we can make the same with PowerPC but if we have to switch i think that Arm is better for the reasons that i hace write above

Making a custom desktop ARM board for OS4.x would soon mount up when you start trying to add in PCIe or integrating RadeonHD and compatible sound. I don't think you'd end up with something much cheaper than current SAM hardware that way.

The point with x86 is that there would be suitable boards available off the shelf. OS licenses could be locked to the hardware just like MorphOS so it wouldn't have to be detrimental to the Amiga dealer ecosystem either.

Anyway all this is pretty moot because Hyperion call the shots and they've repeatedly stated that they will stick to PPC for the forseable future.

Last edited by Rob on 16-Mar-2013 at 01:19 PM.

@WolfToTheMoon

Quote:

untill somebody pulls the plug or the patient dies.

Whit Amiga Inc out of the picture I don't see how thats going to happen, but anyway thanks for spreading food, and trying to derail this thread.

PowerPC is not dead, nor is AmigaOS.

Its also funny to see RiscOS made a comeback on Raspberry pie, as the ARM made a comeback, ARM was virtually abandoned in moeble industry.

Last edited by NutsAboutAmiga on 16-Mar-2013 at 01:51 PM.

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

@NutsAboutAmiga

Quote:

Whit Amiga Inc out picture I don't see how thats going to happen, but anyway thanks for spreading food, and trying to derail this thread.

It's been happening for the last 10 years already, users leaving and not coming back.

Quote:

PowerPC is not dead, nor is AmigaOS.

They're not dead, but they're both slowly dying.

_________________

@royleith

The most important thing regarding our computer is always the same: USES.

What do we want to do with our Amigas?

Either way we look: past, present or future, answers are similar:

- Video
- Music
- Internet / network
- office /productivity
- Hardware availability and good price (choices)

All in all, I think the best option is ARM. Why?

- The future is multi-device and tactile. No discussions: either right or wrong (there is a wide room for discussion), it is what the entire world already chose. NOTHING WILL CHANGE THAT.

- It has the cheapest hardware options (and that includes options starting to compete in power computing with desktop processors, and being a lot cheaper)

- It has the wider hardware configuration options too

- Yes, there are driver restrictions but A LOT LESS than in the x86 world: We know very well how difficult has it been to get drivers just for the graphic chips (even the best efforts from AROS developers only achieved some success with Nvidia cards, and no luck for years with Ati cards). Imagine the same for ethernet, sound, usb and wifi cards. Do we magically and dramatically change the situation we have been suffering in the last 15 years? It simply WONT HAPPEN. There will no way for us to support MOST of the hardware.

Meanwhile, in the arm camp THERE ARE ALREADY OPEN SOURCE DRIVERS FOR MOST OF THE AVAILABLE CHIPS. With ports, driver adaptations and optimizations sometimes in just a couple of days.

- It has the most interesting open source applications, either for:
- video playing (you can play 1080p movies with devices for under 60$. How much does that cost for a complete x86 system? sure nothing less than 150 at the best)
- Video editing: simplified apps for any user with powerful results.
- Image creation and manipulation (painting and image editing tools)
- Music creation: even the most popular and recognized artists are using the new tactile music apps for their songs (just do a quick search on youtube).
- Games and emus: EEEEEEeeeeeeeendlessss choices

As a final touch, it is the best opportunity for Amiga OS to enter in the 2 BIGGEST AND MOST IMPORTANT hardware features which WE already missing (and that will be essential for the future (eerrr not, better say:) THE PRESENT:

MOBILE DEVICES
TOUCH SCREENS

Yes, there will be always nainsayers proclaiming "I hate touch screens, I don´t want them for Amiga OS". Who f**ing cares? is the same as: chunky vs planar, 16:9 vs 4:3, tft vs tube, x86 vs ppc. Scream as much as you want and deny reality surrounding you: THE WORLD ALREADY MOVED ON.

How much will we live as a viable platform ignoring that? do we want to die and as the time pass by just be like a C64 scene? I DONT.

As I see, there just 2 options for Amiga future:

CHANGE!

or

DIE

Last edited by vision on 16-Mar-2013 at 01:45 PM.

@NutsAboutAmiga

Stop with endianess crap, please. It is known that ppc can work big endian as little endian.
Obviusly, hyperion gets benefits from hardware sales, if not, it's nonsense to not broader the market porting the os to platforms that have more future.

_________________
Amiga 1200 desktop. Apollo 030/50 Mhz 8mb ram + ClassicWB + Wb 3.1
Amiga 500 + ACA500plus 8mb + 30gb CF
Raspberry Pi 3b+ and Amibian 1.4
Mac Mini G4 1GB Ram with the butterfly!!

@WolfToTheMoon

Quote:

They're not dead, but they're both slowly dying.

Dying means they are still alive, unlike some other "Amiga" solutions.


However, I agree with need of CPU platform change, here I prefer x86 over ARM. Fast, efficient, best for desktops/laptops and soon also suitable for tablets/smartphones.

@_ThEcRoW

Quote:

Obviusly, hyperion gets benefits from hardware sales, if not, it's nonsense to not broader the market porting the os to platforms that have more future.

x86 port wouldn´t be that hard task, maintaining compatibility with OS3/OS4 applications is much bigger challenge.

@ALL

http://www.gartner.com/newsroom/id/2301715

Not that it matters much as OS4.x has its chosen targets, but for a desktop based machine x86 is the way.
Cheap, abundantly available/accessible, powerful, with the best software support (given our need to rely on the open source world, which predominantely has x86 optimisations and code paths vs. the more generic code for other architectures).
There's also open source drivers for most hardware out there. Granted they still need to be ported to the systes theyre to be used on, but they do exist. Its not by coincidence that AROS has the widest hardware support for any amiga-oid system (and x86 within AROS).

Both Sony and Microsoft have chosen x86 for the upcoming consoles because it makes like easier on development/developers.

As for byte ordering, its entirely possible to use big endian on x86 (amithlon for example). Granted you lose some performance, but you'll still end up with something faster (and cheaper) than elsewhere.

Again though its still moot, as its out of our hands. If you want more input into how a system develops then you'll need to try elsewhere.

@_ThEcRoW

To be fair, the endian-ness issue is not a trivial one. If you wanted to port the OS to x86, you'd need a good solution for backwards compatibility, unless you didn't care about it at all. Amithlon, for example, solved the issue by emulating big endian memory so that all structures that were shared between the emulated 68K system and the underlying x86 host were big-endian. It was a clever solution that required compiler support, though the case could be made that a wholly x86 native system without any emulation outside of an embedded UAE sandbox would be a simpler prospect. One that already exists (JanusUAE on AROS).

_________________
Doing stupid things for fun...

Ah, come on, we should just infiltrate Hyperion and free AmigaOS 4 source code to be ported to X86 by the community.

The problem is that we should infiltrate it physically, because I'm pretty sure they keep the source files on 720k floppies..