error from motorola ? switching 68k to ppc

[Hewitson]

I disagree. Most Amiga owners, in 2019, still do not have an 040.

[meynaf]

Most do not have a 68000 anymore either.
Many have A1200 and this is 68020, whose emulation would have required JIT which did not exist in 1994.

[Hewitson]

Originally Posted by meynaf

Most do not have a 68000 anymore either.
Many have A1200 and this is 68020, whose emulation would have required JIT which did not exist in 1994.

Irrelevant as in 1994 most people still had a 500. A 486 CPU would have emulated this perfectly, as well as paved the way for cheap, powerful Amigas in the future.

The wrong decision was made. Anyone who thinks the move to PPC was the right one is kidding themselves. Apple realised they'd made a mistake a long time ago, it's time for the Amiga scene to do the same.

[meynaf]

Originally Posted by Hewitson

Irrelevant as in 1994 most people still had a 500. A 486 CPU would have emulated this perfectly, as well as paved the way for cheap, powerful Amigas in the future.

So you want what ? That A500 owners throwed their A500 in the bin for a machine that costs a lot more and isn't even more powerful ? Or that they bought an accelerator board that does not make their machine faster ?

People who had an A500 just wanted to keep it. Else they would have bought an A1200 like i did, which would have been a magnitude better than a 486 emulating an A500 - and cheaper than 486, especially a DX4/100.

Not to mention little-endian x86 does not integrate very well with Amiga hardware.

And i doubt by the time they could afford the development time of such an emulation.

Originally Posted by Hewitson

The wrong decision was made. Anyone who thinks the move to PPC was the right one is kidding themselves.

Right, but switching to x86 would have been even worse.
PPC at least were faster than what they were supposed to replace.

[Hewitson]

Originally Posted by meynaf

So you want what ? That A500 owners throwed their A500 in the bin for a machine that costs a lot more and isn't even more powerful ? Or that they bought an accelerator board that does not make their machine faster ?

People who had an A500 just wanted to keep it. Else they would have bought an A1200 like i did, which would have been a magnitude better than a 486 emulating an A500 - and cheaper than 486, especially a DX4/100.

Not to mention little-endian x86 does not integrate very well with Amiga hardware.

And i doubt by the time they could afford the development time of such an emulation.

Right, but switching to x86 would have been even worse.
PPC at least were faster than what they were supposed to replace.

I think you are completely missing my point. I'm not suggesting that people throw their machines in the bin at all. I'm suggesting that instead of choosing PPC as the successor to 68k, x86 should have been chosen.

It would have been capable of 68k emulation as well as running new, native x86 Amiga software.

AmigaOS 4.1 would be written for this platform enabllng people to turn dirt cheap PC's into next-gen Amigas, instead of requiring a ridiculously expensive, outdated and underpowered machine.

x86 is still evolving and getting faster and faster. PPC died (as a personal computing platform) many years ago. It has no future.

[meynaf]

Originally Posted by Hewitson

I think you are completely missing my point. I'm not suggesting that people throw their machines in the bin at all. I'm suggesting that instead of choosing PPC as the successor to 68k, x86 should have been chosen.

How can 486 be chosen as successor to 68040 as they are comparable in speed (with even a slight advantage to 68040) ?
x86 by that time were a regression compared to 68k and it would still have been the case today if 68k development didn't stop.

They were looking for something faster than 68040 - which the PPC is.
Mere (actually unproven) fact of being able to emulate 68000 is clearly not enough, don't forget that A500 wasn't competitive anymore in 1994.

Originally Posted by Hewitson

It would have been capable of 68k emulation as well as running new, native x86 Amiga software.

Ask yourself why even today we don't have x86 based accelerator boards - not even together with 68k like ppc.
Speed doesn't do everything.

Originally Posted by Hewitson

AmigaOS 4.1 would be written for this platform enabllng people to turn dirt cheap PC's into next-gen Amigas, instead of requiring a ridiculously expensive, outdated and underpowered machine.

Nobody would have bought this. Anyone accepting to use x86 as cpu would have bought a PC instead.

And no, dirt cheap PC's could not be turned into next-gen Amigas without having to also emulate the hardware, and for this even 300Mhz machines struggled.

Originally Posted by Hewitson

x86 is still evolving and getting faster and faster. PPC died (as a personal computing platform) many years ago. It has no future.

But what is current future of x86 ? Wouldn't have it been better to directly switch to ARM ?

[turrican3]

Meynaf,

If somone tried to make a 68060 at 1 Ghz, could it be possible ??
What would be the limitations, the cost ???
I know that it will never happen in the real world.
I should add sadly from my pov.

[NorthWay]

A 1GHz 68060 is kinda unlikely considering that Motorola used 42nm(?) in their latest spin.
If you instead call it a compatible then you need a lot of testing and a lot of money. The Vampire has gone a long way towards your target, but they are AFAIK heavily optimizing for FPGA limitations.

I predict/encourage a Kickstarter in the future to implement an HDL based 68020+ full/semi ASIC implementation. IMO it's just a question of time/$$$/technology before that happens.
I have seen quotes thrown around of $5000 minimum for a small batch.

[meynaf]

Originally Posted by turrican3

Meynaf,

If somone tried to make a 68060 at 1 Ghz, could it be possible ??
What would be the limitations, the cost ???
I know that it will never happen in the real world.
I should add sadly from my pov.

Technically there is nothing preventing building 68060 @1Ghz and more.
If using latest tech it maybe could even reach that speed nearly unmodified...
But the world isn't driven by technical considerations.

[AmigaHope]

Originally Posted by meynaf

If answer to any of the above is 'no' then for me it does not qualify as "anything good"...

A mature quantum GPU could do movie-grade raytracing in the blink of an eye by simultaneously testing probability for all potential rays in a pixel. It could conceivably get so fast that you could render thousands of angles for a full glasses-free 3D display that wasn't limited to stereoscopic.

[meynaf]

Originally Posted by AmigaHope

A mature quantum GPU could do movie-grade raytracing in the blink of an eye by simultaneously testing probability for all potential rays in a pixel. It could conceivably get so fast that you could render thousands of angles for a full glasses-free 3D display that wasn't limited to stereoscopic.

And after building such a fantastic machine they will discover that the output image is so bad that it's not usable due to the random nature of quantum mechanics