Blasphemy (I know, I know ...... but ?????)

[Fransexy_]

amigadave wrote:

KThunder wrote:
i have quite a few posts were i just say
[color=ff0000]AROS[/color][/b][/i]
there i did again
aros is a platform independamt rewrite of amiga os3.1 which currently has a port to native x86 and hosted port on linux

But what software runs on AROS?  Is there a list somewhere that will show me what works and what does not?  I always hear about which new programs are now available for MorphOS and AOS4.0, but never AROS.

Educate me please.

There is Hollywood (the first commercial app for AROS)
there is amIRCos an irc client
There is Mosaic a very primitive web browers
There is lunapain a deluxe paint clone
There are a lot of more
you can take a look at AROS archives

[uncharted]

Fransexy_ wrote:

I thinked and continues thinking that the apple switch to intel is a dead end.Yes maybe initially their sales will increase (it´s the the improvement of the death) but after this initial boosts it will be sunk

Why?

[melgross]

I'm willing to bet dollars for donuts that OS X is a vastly more complex and sophisticated OS than any Amiga version that is being worked on.

The numbers that came out were that Apple has 350+ programmers on the main thread of its development.

That compares to, how many for the Amiga?

The point is that it wouldn't take nearly as much work to do it.

And, I'm not talking about some sideways version, as was brought up, but the main development model.

The PPC is now officially a dead end street. The G4 models have been relegated, for several years, to Freescale's embedded line. Now that Apple is no longer going to be buying them, any features that were useful for desktops will be expunged form the future line as being unnecessary, and wasteful of silicon, and power.

It saddens me to think that, but it's true.

All major operating systems that can be considered as being desktop systems now run on x86. This will have to as well.

With Parallels virtualization software, and the virtualization now being built into Intels chips, Amiga OS can be run on a Mac Mini quite well. Better than on any proposed hardware that will likely never see the light of day anyway.

It would therefore not require anyone to buy hardware. If you have hardware, you're set to go.

Fransexy_ wrote:

melgross wrote:
That's not even in the slightest bit right. I'm getting a lot of interest from people who have been interested in switching, over the years,  who are ready to take the plunge.

These will be the only machines able to run all operating systems at full, or near to full, speed.

Now, if the moron's who are crapping their way through finishing some form, any form, of Amiga OS, would get off the pot, and move it to x86, it would run there, on the Mac,  far better than on any piece of junk that MIGHT ever see the light of day as Amiga native hardware.

Apple a more big an active company has taken year of development of their OS for x86 in the background (or do you think that apple magically decided the stwitch an in a night made the port?) And even more OSX is based on BSD kernel that has an X86 port for much more years.An in these years a lot of apllications has been made for osx.
And do you want that amigaos will be ported to x86 in a night day from a small company with a small userbase and a small software library??? you are a crazy man, baby

[SamuraiCrow]

@melgross

Amigas have always run on embedded controller versions of the 680x0 series so why would embedded PowerPC models be any different?  They don't have anything missing from the desktop models except that they are more frequently multicore and less frequently involve out-of-order execution.

With the Cell processor being used in game systems and a triple-core PPC in Microsoft's own XBox 360 and the Nintendo Revolution it would seem that all of the most computationally intensive applications (games) will be running on PowerPC.

It seems, then, that Intel is getting relegated to the desktop in a market where embedded systems and consoles are taking over.  Which one's the dead end?

[melgross]

Ok. First of all, embedded cpu's are not usually multi-core chips. Embedded chips, for the most part, do not have a need for great performance. The greatest need is for low power consumption, and reliability.

If you go through the catalogs of the embedded lines of several manufacturers, you will notice that.

Apple, as part of the Power alliance, influenced the development of the various chip lines, including supplying microcode and engineering. In turn, the manufacturers developed features that were exclusively needed by Apple. Altivec was one of those developments.
While I'm not saying that Freescale (and IBM) will strip these chips of all of these features, anything that hasn't proven to be required by the other embedded applications wou't remain. It's survival of the fittest.

Despite what you may have read, the Cell is only marginally a PPC chip. While the Xenon is closer to the older G4 line, with improvements, and deletions, the Cell is vastly different. It requires a very different programming model.

Intel will sell 300 million chips to desktop, workstation, and minicomputer manufacturers this year. Those numbers increase about 10 to 15% a year, particularly with Apple now using them as well. They will also sell tens of millions of x86 cpu's optimised for embedded applications to auto manufacturers and others. And AMD will do the same on a smaller scale.

Both IBM and Freescale will sell PPC chips in the tens of millions to the same kinds of manufacturers.

Together, MS and Sony will use perhaps 8 million Xenon and Cell chips this year, and perhaps 25 to 30 million each year until new models come out. The Revolution will not be using the Xenon, it will be using a much simpler chip. These chips, unlike those designed for the computer industry, and even for the embedded one, where chip designs change more slowly, will not be enhanced for the life of those consoles. While both IBM and Sony will sell some to Toshiba for their large screen TV's (the other intended use for the Cell), any workstations that will be based on them will have, at best, a VERY limited market.

Which one's the dead end?

SamuraiCrow wrote:
@melgross

Amigas have always run on embedded controller versions of the 680x0 series so why would embedded PowerPC models be any different?  They don't have anything missing from the desktop models except that they are more frequently multicore and less frequently involve out-of-order execution.

With the Cell processor being used in game systems and a triple-core PPC in Microsoft's own XBox 360 and the Nintendo Revolution it would seem that all of the most computationally intensive applications (games) will be running on PowerPC.

It seems, then, that Intel is getting relegated to the desktop in a market where embedded systems and consoles are taking over.  Which one's the dead end?

[SamuraiCrow]

Speaking of not changing, the Playstation 3 will probably not change for the life of the product for compatability reasons.  The Commodore 64 also followed that same pattern.  The C64 was the best selling single-model computer in history.

Just because the chips change less doesn't mean they'll be any less competitive in the console market than they are today.  If anything, that gives IBM the incentive to be more creative as the Cell processor and the Kilocore processor indicate.  IBM is making great strides since Apple forced them out of their comfort zone.

[Waccoon]

melgross:  The PPC is now officially a dead end street.

My take as well, as the problem isn't the CPU, it's the chipset and all the other logic that makes for a "real" computer.

SamuraiCrow:  Amigas have always run on embedded controller versions of the 680x0 series so why would embedded PowerPC models be any different?

PowerPC isn't widely applauded for compatibility.  By comparrison, the infamous 68000 didn't change very much over the years.  Complex bus logic is also built into CPUs these days, and external logic must be matched accordingly.  It's expensive, especially without good chipset support.

SamuraiCrow:  It seems, then, that Intel is getting relegated to the desktop in a market where embedded systems and consoles are taking over. Which one's the dead end?

Coding practices are very different.  Desktop-style programming demands heavy abstraction.  Embedded systems and consoles encourage the use of special instructions and features that are not widely available, even on other devices using the same family of CPUs.

There are good reasons why the CPUs are developed the way that they are for each market, and the architectural directions they take are based on market demands.  It's not easy or practical to swap processors between markets, especially if the development tools are designed for a particular purpose.  Availability of tools is a major issue, since that's how you make things forwards compatible, not just backwards compatible.  I seem to recall that forwards compatibility isn't a major issue in the console/embedded market.  If it is, developers usually take the easy way out and use Java.  *shudder*

melgross:  Despite what you may have read, the Cell is only marginally a PPC chip. While the Xenon is closer to the older G4 line, with improvements, and deletions, the Cell is vastly different. It requires a very different programming model.

I've heard the PPE core in the Cell is more or less used as a preprocessor to keep the SPEs fed.

SamuraiCrow:  Speaking of not changing, the Playstation 3 will probably not change for the life of the product for compatability reasons. The Commodore 64 also followed that same pattern. The C64 was the best selling single-model computer in history.

The C64 also didn't have a real OS, people had to give up on their software when moving to the Amiga (or whatever), and there were a lot more games available than apps.  Even commercial programs required some POKEing every now and then.  What a pain.

With all this open-sourcing going on, and the fact that the Amiga pretty much started the public domain and shareware scene, it shocks me that many Amigans still want to shackle themselves to closed hardware.

SamuraiCrow:  If anything, that gives IBM the incentive to be more creative as the Cell processor and the Kilocore processor indicate. IBM is making great strides since Apple forced them out of their comfort zone.

They can afford to be more daring as backwards compatibility isn't a major priority, and if it is, IBM is being commisioned to do custom design for a specific platform that wouldn't be available to other companies without a hefty price tag.  If you buy a Cell processor, you're paying Sony, not IBM, and you have to pay IBM indirectly for the work they did on the processor.  More middlemen, plus, you still have to have a custom motherboard made.  Specs look good on paper, but real-world practicality is lacking.

All of this to do... what?  Web browsing?  E-mail?  Running 68K emulators?

Why bother with so much custom hardware if it will take forever to make OS tools to use it correctly?  What about the budgets required to make those tools and port them every time the architecture changes?  It doesn't make sense to hype a specific processor type unless you intend to hard-code it, and that is suicide for any real OS.  The whole point of an OS is so programmers don't have to worry about the hardware.  How abstract the OS needs to be depends on the usage, and desktop computers are really designed to do just about anything.

Release an OS first, get it in the hands of developers, get the basics out of the way, and then you can worry about cutting-edge hardware concepts.

[Fransexy_]

Ok. First of all, embedded cpu's are not usually multi-core chips. Embedded chips, for the most part, do not have a need for great performance. The greatest need is for low power consumption, and reliability.

Sure? Well, this CHIP is aimed at embedded market and has 1025 cores

[Fransexy_]

They can afford to be more daring as backwards compatibility isn't a major priority, and if it is, IBM is being commisioned to do custom design for a specific platform that wouldn't be available to other companies without a hefty price tag. If you buy a Cell processor, you're paying Sony, not IBM, and you have to pay IBM indirectly for the work they did on the processor. More middlemen, plus, you still have to have a custom motherboard made. Specs look good on paper, but real-world practicality is lacking.

You are wrong.Sony not only wants that other OS´s run on their CELL based Playstation but to made the cell the next multimedia chip.Sony, toshiba and ibm wants that their cell be used in all types of devices from computers to wasihg machines.So it will be not difficult to abtain cell processors

[melgross]

It entirely means that they will be less competitive. The computer market is nothing at all like it was back when people bought the C64, and put them in the closet after a couple of months because there was nothing useful they could do with them. and please don't say that that wasn't done, because it was the big joke at the time. I remember it very well. The C64 was the best selling single model in history — at the time. today, almost any single model, from anyone,  will outsell it.

People were told how easy balancing their checkbooks would be with a computer. When they found that it wasn't, and that there wasn't much else they could do with them, they either threw them away, or relegated them to the closet or attic.

Perhaps, using Amiga software, and hardware, you aren't aware of the modern computer scene, but chips change constantly. A one year old chip is already consigned to the trash.

The G4 is totally obsolete, useful for nothing. The G5 requires too much in ancillary support in the machine. The programming for the G5 is also different on several significant levels. Anything optimized for a G4 would have to be partly re-written to work properly on a G5. Sure, the older code will work, but as we found out on the Mac, portions will actually run slower, if the re-write isn't done.

And who will be using the G5's these days, other than some very small specialized manufacturers?

Your wish for the chips you mention is nothing more that that. It simply isn't relevent that the consoles will be competitive. Competitive with what? Each other? The hottest PC games already are superior to the ones coming out on the latest consoles. And so it will always be.

If the Amiga community lives with the expectations you have, then it will continue its rush to the grave.

The only chance it has, is to take advantage of the superior hardware out there, either Mac, or PC.

Any chips unique to the Amiga are totally irrelevant these days. CPU's and GPU's have made certain of that. No one will fund the development of new chips for what amounts to a few tens of thousands of users.

The Amiga OS teams have to get into the real world. The Amiga OS is today, no more than a hobby OS. As a program run under virtualization, it will sell to people who are interested in such things. But, if it is dependent on new hardware, then it has no hope.

Some understanding of this must be acknowledged. All of the infighting has done no one any good.

Most likely, there will never be any money to be made here, and the project should be relegated to Sourceforge. At least that way there will be some people working on it who really care. Unlike now.

SamuraiCrow wrote:
Speaking of not changing, the Playstation 3 will probably not change for the life of the product for compatability reasons.  The Commodore 64 also followed that same pattern.  The C64 was the best selling single-model computer in history.

Just because the chips change less doesn't mean they'll be any less competitive in the console market than they are today.  If anything, that gives IBM the incentive to be more creative as the Cell processor and the Kilocore processor indicate.  IBM is making great strides since Apple forced them out of their comfort zone.

[melgross]

That's not aimed at the normal embedded market. I didn't say that no chips for that market were multi core. That's intended for super computers, and the like.

And are you telling me that you expect a future Amiga to run on that? Because if you do, then you have no credibility.

Fransexy_ wrote:

Ok. First of all, embedded cpu's are not usually multi-core chips. Embedded chips, for the most part, do not have a need for great performance. The greatest need is for low power consumption, and reliability.

Sure? Well, this CHIP is aimed at embedded market and has 1025 cores

[melgross]

It doesn't matter much what Sony, or IBM wants. There are no tools available to work with these chips. And just what OS is going to be used with them, other than a Linux implementation, which, so far, hasn't succeeded in getting any of the SPE's to work? So, they have one in-order two thread enhanced (sotra) G4, with a truely wonky memory model that no one has ever used before, that can only handle 512MB of RAM without the SPE's.

Great!

So, you want the Amiga OS to be re-written for that, rather than for x86?

[melgross]

I agree with what you say.

The PPE actually does some work of its own. But without the SPE's, there really isn't much point to it.

Waccoon wrote:
er*

melgross:  Despite what you may have read, the Cell is only marginally a PPC chip. While the Xenon is closer to the older G4 line, with improvements, and deletions, the Cell is vastly different. It requires a very different programming model.

I've heard the PPE core in the Cell is more or less used as a preprocessor to keep the SPEs fed.

[SamuraiCrow]

melgross wrote:
It doesn't matter much what Sony, or IBM wants. There are no tools available to work with these chips. And just what OS is going to be used with them, other than a Linux implementation, which, so far, hasn't succeeded in getting any of the SPE's to work? So, they have one in-order two thread enhanced (sotra) G4, with a truely wonky memory model that no one has ever used before, that can only handle 512MB of RAM without the SPE's.

Great!

So, you want the Amiga OS to be re-written for that, rather than for x86?

The memory model has been used on the RAM expanders for C64s and 8086-based IBM PCs before and works.  The reason it works so well is that interleaved memory is really fast when accessed sequentially and a DMA-based approach delivers that kind of sequential access.

As for being rewritten for the x86 I'd be cheering for the AMD64 if it weren't for that funky little-endian byte ordering.

[melgross]

Interleaved memory? Do you understand the memory model?

It's called a "ring memory" scheme. It comes from Rambus.

It has nothing to do with the C64, or the 8086.

SamuraiCrow wrote:

The memory model has been used on the RAM expanders for C64s and 8086-based IBM PCs before and works.  The reason it works so well is that interleaved memory is really fast when accessed sequentially and a DMA-based approach delivers that kind of sequential access.

As for being rewritten for the x86 I'd be cheering for the AMD64 if it weren't for that funky little-endian byte ordering.

[SamuraiCrow]

RAMBUS memory isn't the only memory that will work with the Cell processor as Mercury Semiconductor is trying to make DDR memory work with the Cell as well using a custom northbridge.