Do you approve of PPC (in some form) as the future of Amiga?

[Heiroglyph]

Amithlon was as close as we ever had to an x86 Amiga and that was killed by stupidity.

If it had support and a few revisions, it could have gotten somewhere.

So did Phase5 push us into PPC all by themselves?  Why are we still there?

[Heiroglyph]

My comment about Amithlon was due to the fact that it had a transparent way to use 68k code at hyper speed for an immediate improvement, while also offering a way to use native x86 code.

Had this approach matured, we could have shifted towards the OS being x86 code similar to how OS4 and Morphos have shifted to PPC.

My biggest problem with Aros as it stands is lack of 68k software support.  As it is, Aros is pretty much like any Linux distribution minus stability, drivers and memory protection.  They both have UAE and the majority of Aros apps are on Linux too.

Seamless 68k integration could change my outlook dramatically, but I'm not sure what the status is beyond launching UAE when you click on an adf.

PPC is just an expensive dead-end and FPGAs can't get us to even current hand-held device speeds.

[Heiroglyph]

I dont see anyone wanting to do the work to port Amiga to x86. I cant see anyone believing theyd ever see a return on their invesement to do that.

Aros is doing a pretty good job of that, but I agree there is no money in it.

It would have to be a labor of love, that's for sure.

[Heiroglyph]

Since we already have AOS4 and MorphOS, I'm not sure about the point of this question. PPC has been a part of our future for quite a while.?

I think he was getting a judge of how the users feel about PPC and is it still the way of future, or should we consider other CPUs for future development.

I'm firmly in the x86 camp.  Nothing will ever be as fast per dollar again.

[Heiroglyph]

Wrong! A medium sized and priced fpga should allow an enhanced 68k processor to achieve around 150MHz.

Wow, a whole 150Mhz?

I know MIPS aren't everything, but it's one indicator with numbers that were readily available on Wikipedia.

Assuming 060 performance, not 020/030 as most fpgas currently are:
68060@66 MHz = 88 MIPS
68060@75 MHz = 110 MIPS
68060@110 would be roughly 220 MIPS

Intel Pentium III @ 500 Mhz = 1,354 MIPS
ARM Cortex A8 @ 1.0 GHz = 2,000 MIPS
PS3 Cell BE @3.2 Ghz = 10,240 MIPS
XBox Xenon PPC @3.2 GHz = 19,200 MIPS
Intel Core 2 QX6700 @ 2.66 GHz = 49,161 MIPS
Intel Core i7 i980EE @ 3.2 GHz = 147,600 MIPS

x86 is clearly fastest per dollar now and will be for the foreseeable future.

Even the cheap embedded x86 chips are faster than the PIII 500MHz.

You can buy an off the shelf x86 that is 1677 TIMES FASTER than the fastest 060 available (66Mhz) for less than the cost of the 060 card.

I'm not sure why anyone would want to hobble Amiga with slow CPUs, they used the 68000 because it was the fastest for the money at the time.

Before Amiga went under they were already looking for alternative CPUs, so I'm not sure why there is such a strong attachment to a basic component that was near end of life already.

[Heiroglyph]

Seamless compatibility with existing software?  How useful will a lightening-fast system be without much software to run on it?

At 1000+ times the speed, I'm sure you could handle a compatibility mode ala MacOS or MorphOS.

[Heiroglyph]

While I agree that the 68060 was a great CPU, there is a point where brute force and momentum overcome elegance.

I like the idea of using 680x0 code as a sort of bytecode JIT like Java or .net does.

It would give the ease of programming and backwards compatibility but wouldn't be run inside a complete sandbox like UAE.

The 68k code would seem to run far faster than on a real 68060.

If you wanted to use the native faster CPU you could and the 68k could call x86 libraries seamlessly.