@Bosanac

That’s what I’m thinking too, and that most likely will be a disaster for both Apple and ARM in the long run.

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

@kolla

Apple could certainly afford it, but I think a purchase by any single entity would be blocked.
I think what SoftBank was hoping for was for a public/private alliance to come forward, ideally made up of a tinker, tailor, soldier, and spy.

Nvidia or Apple could be the soldier, but you'd need a major university like Carnegie-Mellon or MIT as the tinker, someone from the manufacturing sector like TSMC or Samsung to be the tailor, and a government funded R&D group like ARPA or CSIRO to be the spy.

We'll see who snags up the larger percentages in the IPO.

_________________
All the way, with 68k

@agami

Aren’t governments, research and educational sector skipping ARM and going straight on RISC-V already? Didn’t I read about NASA working with SiFive to make RISC-V the “go-to” architecture for space exploration?

It doesn’t matter that ARM is correct and certain entities have been breaching the licenses - the point is that a lot of people have been under the impression that ARM was "safe”, and now it has become obvious that it isn’t, and so they are looking elsewhere.

Last edited by kolla on 13-Jan-2023 at 08:05 PM.
Last edited by kolla on 13-Jan-2023 at 08:05 PM.

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

@kolla

Quote:

Aren’t governments, research and educational sector skipping ARM and going straight on RISC-V already?

Correct. Which is why SoftBank's initial hopes were dashed, and their current gambit is the IPO.

_________________
All the way, with 68k

arm is now pc like x86.
pc is job and is boring.
amiga either have to be job and be good enough to sell it to out customers and use instead of lnx/win/osx
either have to be hobby and use real cpu that is not pc which means no x86 no arm no riscv.

@ppcamiga1



Has to use a CPU that is not PC!? What? Arm will work just fine unless Softbank screw up with harsh licensing. Our problem is AmigaOS 4.x development has stalled. In the meantime we can have fun developing for 68k or using Arm chips to emulate accelerators or whole systems! There's no law against it! Do you expect people to soldier on using POWER CPUs when the PPC chips finally dry up? No one will continue to pay £2,000+ for AmigaOne hardware!

_________________
"Art challenges technology. Technology inspires the art."
John Lasseter, Co-Founder of Pixar Animation Studios

@ppcamiga1

How can you dismiss ARM and RISC-V when they are very much the closest you can get to PowerPC? Both are RISC and big-endian 32bit capable. Just like PowerPC.

Do I have an “ARM PC”? No.
Do I have a “RISC-V PC”? No.
Do I have a “PowerPC PC”? Yes, several!

So which is closest to “being just like x86”?

Last edited by kolla on 14-Jan-2023 at 10:20 PM.

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

@ppcamiga1

Quote:

ppcamiga1 wrote: amiga either have to be job and be good enough to sell it to out customers and use instead of lnx/win/osx either have to be hobby and use real cpu that is not pc which means no x86 no arm no riscv.

You were so close to actually making a reasonable statement here.

What the hell is a "real CPU"?

Is it one that is produced in very small numbers so the per-unit price is prohibitively high?
Oh wait, let me guess: A real CPU is big endian 32bit. Like the 68k, which was to use your phrasing, a "job", the original PowerPC, also a job, and now the FPGA-based 68080, which is not a job.

I guess any sufficiently popular system, with a mature enough ecosystem is automatically a job. And job = PC (yuck).
- ARM and RISC-V hobby boards, job
- Arduino and Teensy boards, job
- Lego Mindstorms, job

But here's the interesting thing you are missing. One could actually design a 32-bit big endian RISC-V if they'd like. Would that CPU be "real"?

Or is it more likely that most of the time you have no idea what you're talking about. Let's face it, you peaked in the early '90s and every year since then has been a cold hard reminder of just how inadequate you are.

_________________
All the way, with 68k

@agami

Quote:

agami wrote: @ppcamiga1 Quote: ppcamiga1 wrote: amiga either have to be job and be good enough to sell it to out customers and use instead of lnx/win/osx either have to be hobby and use real cpu that is not pc which means no x86 no arm no riscv. You were so close to actually making a reasonable statement here.

I think it actually is a reasonable statement. If your hobby is Amiga retro computing then it has to be 68k, not x86 or riscv. Anything else is diluting/polluting it.

On any other platform this would be reasonable. People who are into retro PC's don't feel the need to shove a foreign CPU into them. People who want a modern PC (for work or play) don't waste their time trying to make a retro PC into a modern one.

Quote:

What the hell is a "real CPU"?

For the Amiga, an actual 68k chip.

Quote:

I guess any sufficiently popular system, with a mature enough ecosystem is automatically a job. And job = PC (yuck).

'Mature' or not, Job = work = PC (yuk). It's been that way since the IBM PC in 1981, and even before. Some people wanted to make the Amiga = work too, but today that makes no sense.

Quote:

- ARM and RISC-V hobby boards, job - Arduino and Teensy boards, job - Lego Mindstorms, job

A hobby by definition is not work. But these platforms are not Amiga.

Quote:

But here's the interesting thing you are missing. One could actually design a 32-bit big endian RISC-V if they'd like. Would that CPU be "real"?

Once implemented in silicon, yes. But it would not be Amiga.

Quote:

Or is it more likely that most of the time you have no idea what you're talking about. Let's face it, you peaked in the early '90s and every year since then has been a cold hard reminder of just how inadequate you are.

Congratulations you just crossed the line. You are now officially a cyberbully.

I don't know why some 'Amiga' fans want to ram 'mature ecosystems' down our throats. If we wanted to do that we would be working with them, not tinkering with old vintage computers from the 90's. Funny thing is this attitude doesn't seem to affect other retro platforms, only the Amiga. It's almost as if they do it to overcome their own feelings of inadequacy. After so many years of being insulted for using an 'immature' Amiga, they finally see an opportunity to join the 'mature' crowd, then to prove how 'mature' they are they have to insult anyone who is happy playing with 90's technology.

@bhabbott

Quote:

Congratulations you just crossed the line. You are now officially a cyberbully.

Wow, um, that was so unexpected.
Do I get badge or something?

_________________
All the way, with 68k

@bhabbott

Quote:

bhabbott wrote: @agami Quote: agami wrote: @ppcamiga1 [quote]ppcamiga1 wrote: amiga either have to be job and be good enough to sell it to out customers and use instead of lnx/win/osx either have to be hobby and use real cpu that is not pc which means no x86 no arm no riscv. You were so close to actually making a reasonable statement here.

I think it actually is a reasonable statement. If your hobby is Amiga retro computing then it has to be 68k, not x86 or riscv. Anything else is diluting/polluting it.
[/quote]

And PowerPC? It’s as if you forget who you’re “defending” here…

Is PowerPC any more Amiga than ARM is? If so - why?

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

"If your hobby is X, your CPU has to be Y"

Oh, so our hobbies have to conform to someone else's expectations now, so they? Unless it's using a physical 68K, it's a fraud. I'm glad I still have my actual 68K hardware then or I'd have to rescind my license...

So, in summary, I don't agree at all with ppcamiga's view that PPC is somehow the right way to go these days, but I wouldn't try to insinuate that he's wrong for liking it or wanting to use it. Where he's wrong, in my opinion, is that the PPC is somehow the the natural continuation.

The only thing, in my opinion, that should be 68K is the userland programming model. It doesn't mean that you have to run on an actual 68K CPU (though that's obviously a desirable thing). Emulation, FPGA or real silicon, it's all good. It does seem inevitable the PiStorm will allow 68K code on bonafide Amiga machines to do things that no "real" 68K could hope to. Yet, that's no different than comparing the fastest 68060 accelerators to the basic 7MHz 68000 found in an A500 as far as I can see. That's the reason I really want to see affordable versions for all the major Amiga models.

Last edited by Karlos on 16-Jan-2023 at 10:16 AM.
Last edited by Karlos on 16-Jan-2023 at 10:15 AM.
Last edited by Karlos on 16-Jan-2023 at 10:14 AM.

_________________
Doing stupid things for fun...

@kolla

Quote:

Is PowerPC any more Amiga than ARM is? If so - why?

I would say so because it had an accelerator with a PowerPC. It was the first non-68K foreign CPU on an Amiga accelerator card. Though, if Phase 5 were truly attempting to revive the Amiga where Commodore left off, an Itanitum may have been a better choice.

ARM on the other hand has lagged behind in the Amiga world. And only in the last few years has arrived on Amiga boards. Even then it's in limited capacity. As a co-processor on video cards or used as a 68K emulator. But not directly accessible like PowerPC was through libraries. The closest I last found was being able to run ARM binaries externally.

Perhaps they are avoiding the kernel wars again with an ArmUP going against a WarpARM. While it would be great decompressing JPEGS at 1.5Ghz and doing 3d rendering even faster than the fastest PPC that can be plugged into an Amiga, does the Amiga market want that? It would introduce another binary format and Aminet would need extending again. The "real" Amiga market, even the high end market with ZZ900 boards and SSD on SCSI with 100Mhz 68060, still has a focus on 68K. To the point that new accelerators aren't even trying to replace the 68K or add a co-processor to run software at modern speeds, they are emulating the 68K to run 68K code as fast as possible.

@Hypex

I think emulation makes the most sense here anyway. The dual processor architecture of PowerUP was a pain. Optimizing software for least amount of context switching was essential. The moment MorphOS and OS4 appeared they dropped the 68K in the original boards immediately and no matter how fast an 060 you may have had, it was the right move. Taking away the complexity of context switching allowed many 68K applications to run a lot faster. A case in point, the 020+ doom attack port that I used to play reasonably well at 320*200 on CGX on my 040 could run at 640*400 at a comparable speed under petunia on the 603e.

The writing was on the wall there and then that emulating a 68K inside an actual Amiga was something doable. I have no doubt the Pi4 running Emu68 is going to kill it.

_________________
Doing stupid things for fun...

from developer point of view Amiga powerpc was natural continuation of amiga 68k.
32 bit big endian cpu allow integration and use 68k libraries in ppc code.
everything work ootb and switch to ppc gcc amiga code need less time and work
than switch from 68k sas c to 68k gcc.
almost 30 years old code complie and run with almost no changes.
compare this to aros x86 or arm when 28 years after start of developing system there is still no
working mui clone.
thats why ppc is more amiga than arm.
carlos switch to c or c++ and you will see it.
kolla lern c or c++ and you also get it.

making software for emulator when no one use real cpu
is simply stupid waste of time and work.
just compile to native code.

amiga either have to be hobby which means no pc which means no x86, no arm, no risc v
either have to be work which means good enough to sell to customers instead of win/lnx/osx or ios/android.

@ppcamiga1

Lol. Remind me how the *vast majority of all Amiga software ever written* runs on OS4?

_________________
Doing stupid things for fun...

@ppcamiga1

All the C/C++ code I have worked with compile equally well on ARM as on PowerPC and for that matter, 68k. Only issue I ever had on ARM was when OpenSSL developers once upon a time made an assumption that all ARM is little-endian, so my big-endian 32bit ARM systems did all encryption “wrong-endian”. That was fixed, like 20 years ago.

For around a decade, I compiled and maintained binary packages for Gentoo/Linux on 32bit big-endian ARM.

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

@ppcamiga1

Lol. Remind me how the *vast majority of all Amiga software ever written* runs on OS4?

"from developer point of view Amiga powerpc was natural continuation of amiga 68k"

No it wasn't. It's completely different. The only things it has in common are the 32-bit word size and a big endian memory model. Let's see what's different:

Load/store architecture.
There are no effective address operands (kind of the same point as above, or rather the consequences of it).
Link register for subroutine call.
No extended precision for FPU.
Bits are enumerated in the reverse order relative to 68K and just about every other sensible architecture ever invented.

There are more that I forget, especially in terms of compatibility between series, MMU foibles and other more specific things.

PowerPC wasn't bad due to the above points and it had some neat features at the assembly language level. However there's nothing at all natural about it for someone used to 68K assembler.

It is somewhat simpler from a C perspective, of course, but then, what isn't? If you've ever written code in C that can't be modified to recompile for systems with bigger pointers or that have the opposite byte order then you've not done a particularly great job.

_________________
Doing stupid things for fun...