@liquidbit

Quote:

liquidbit wrote: "Lack of money will make a port unpayable" -I don't agree, there are a lot of money from people which are looking the next opportunity to invest, we saw a lot of waste latelly.

The level of investment required would be equal to, if not more than, all the money-time invested in all things Amiga over the past 20 years.

Therefore, the sums of money that have "lately" been invested (wasted), would not give us anything meaningful, even if you could somehow rally all the wallets around a singular mission.

Last edited by agami on 17-Dec-2023 at 02:19 AM.
Last edited by agami on 17-Dec-2023 at 02:18 AM.

_________________
All the way, with 68k

@Karlos

I think the issue with most Amiga NG folk, regarding the viability of Raspberry Pi or indeed any ARM-based system, is the lack of a PCIe x16 slot to accommodate off-the-shelf Radeon graphics cards.
Which is why I can see the appeal of the referenced RISC-V system, having the aforementioned slots.

On the other hand, should the PiStorm/emu68 developers attain the required performance out of the Pi's onboard GPU, then there's plenty there for anything that a 2023 MorphOS/AmigaOS 4 system is able to leverage. Rendering the overkill Radeon card, redundant.

_________________
All the way, with 68k

@Karlos

The AMD port was 1st taunted a long time ago, nothing real was been seen until 4 years ago, nothing real has been seen ever since.

With the way the team tends to roll that proves nothing one way or the other.

What is known is that bigfoot did work on better 3D drivers for slightly newer cards which also can be seen in various ways.

- they realized that an AMD port is beyond the scope of their resources (would also apply to ARM) and gonna ride out old Macs and the X5000 into the sunset

- they decided that an AMD port needs to support GFX cards one would actually want to pair with a low end AM4 mobo/CPU


@agami

Add PCIe to ARM or RISC-V and you removed 1 stumbling block while killing their only benefit (super small and super low power).

-> "it might not be worse than AMD"

Karlos also added "ARM might make me feel a little less stupid" so we are now a 2 compelling argumnents for porting to anything but AMD

_________________
- We don't need good ideas, we haven't run out on bad ones yet
- blame Canada

@agami

And do what with their Radeon Cards? My 2007 PC (Q9450 2.666GHz) is paired with a bottlenecked GTX780ti and runs games like Doom2016 at 60fps/1080p still (few dips, not worse than 45). It even ran Eternal OK until the driver's stopped being compatible but there's a different story for you.

Meanwhile, what are the NG crowd rubbing on their more recent Radeon hardware? WBPattern seems to be the most obvious answer pouring over used screenshots. Gaming outclassed by a PC from 2007. Real nice effort.

@Kronos

I never once said stupid. I said "embarrassing". It's embarrassing to go from being the king of a puddle, to a nobody on the ocean. One way or another, half key benefits of MorphOS are lost on x64. You either end up losing the much taunted resource efficiency into resource paucity (you can't utilise a fraction of it), or you keep those and lose decades of seamless backwards compatibility.

Personally I'd feel better underutilising a cheap low power consumption SoC then I would a modern x64 ans dedicated GPU. Unless the latter was also going to multiboot something , anything that can use it. But that's just me.

I'd still rather use a modern x64 system with all of the above faults duly noted than a decades old G4 that might not turn on the next time I power cycle it. And I'd rather use somthing like RPi with the investment put into drivers for the hardware it has than worrying about a PCIe 16x dedicated GPU to add to the CPU that I can't use a fraction of the power of because it's bottlenecked.

Last edited by Karlos on 17-Dec-2023 at 08:20 AM.

_________________
Doing stupid things for fun...

@Karlos

Quote:

Karlos wrote: @agami And do what with their Radeon Cards?

And what to do with a x00MHz fake 68k or even a 100MHz real one?

Everything "Amiga" is pointless and could be done easier/better with a bottom end PC these days. Even without going into emulation as every old game or app has a shovelware JS version somewhere on the net.

Also what is the "benefit" of MorphOS that is been lost? Running efficient on inefficient HW, vs running less efficient on efficient HW

In the end MorphOS will be single threaded for the foreseeable future and thats why I would want to run it on the CPU with the best single core performance, which is neither be found on ARM (except maybe Apple, but thats a non starter) nor RISC-V.

Power consumption is marginal if the other cores are disabled the proper way, same for the Radeon while displaying a mostly static Ambient.
Might get even better if they support APUs.

A rPI running endian swapping code (and for some parts even fully emulated code) on a single core would perform better than the best PPC, but not by a big enough margin to keep up with current level of usability for the long term.
!!!!_IF_!!!! they go beyond single core doing it on AMD should be somewhat easier.

_________________
- We don't need good ideas, we haven't run out on bad ones yet
- blame Canada

@Kronos

Quote:

A rPI running endian swapping code

And why would it need to do that? Emu68 runs big endian.

Quote:

thats why I would want to run it on the CPU with the best single core performance

To do what with? Go on, I'm all ears as to what software catalogue you have that needs anything like the full performance of a single core on x64 to run respectably that doesn't already run more effectively on an OS for the same hardware already?

Last edited by Karlos on 17-Dec-2023 at 08:54 AM.
Last edited by Karlos on 17-Dec-2023 at 08:53 AM.

_________________
Doing stupid things for fun...

@Karlos

Quote:

Karlos wrote: And why would it need to do that? Emu68 runs big endian.

On a CPU that "prefers" little endian. Sure every modern CPU has some big endian support these days, but deep down they are little endian and running big endian code will come at a cost.

Quote:

To do what with? Go on, I'm all ears as to what software catalogue you have that needs anything like the full performance of a single core on x64 to run respectably that doesn't already run more effectively on an OS for the same hardware already?

Yep, throw away everything Amiga since Windows/Mac/Linux is way better for everything....

The logic of "I can't use the full performance that why I want it on HW with less then 50% the performance" somehow escapes me.

Last edited by Kronos on 17-Dec-2023 at 09:14 AM.

_________________
- We don't need good ideas, we haven't run out on bad ones yet
- blame Canada

@Kronos

Quote:

And what to do with a x00MHz fake 68k or even a 100MHz real one

That's easy. Do all the things the original amiga was never able to do effectively or that you couldn't even conceive of doing outside of UAE.

_________________
Doing stupid things for fun...

@Karlos

Quote:

Karlos wrote: @Kronos Quote: And what to do with a x00MHz fake 68k or even a 100MHz real one That's easy. Do all the things the original amiga was never able to do effectively or that you couldn't even conceive of doing outside of UAE.

A.k.a. things that could be done much better, easier and faster on a low end Windows-box

_________________
- We don't need good ideas, we haven't run out on bad ones yet
- blame Canada

@Kronos

You are labouring under the misapprehension that I think MOS on x64 is a bad idea. I don't. It's PPC that is the bad idea. I'd go further. as we approach 2024, it's about the worst idea possible.

I just happen to think that ARM offers a better alternative to PPC than x64 and you haven't given a single viable reason why that isn't a case today.

And let face it a GHz class virtual 68K today is infinitely cooler than a real PPC from the mid naugthies. PPC has precisely feck all going on in that regard.

Last edited by Karlos on 17-Dec-2023 at 09:26 AM.
Last edited by Karlos on 17-Dec-2023 at 09:24 AM.

_________________
Doing stupid things for fun...

@Karlos

Sofar you haven't gone beyond "it might not be worse" when it comes to ARM vs AMD.

Power saving will be marginal and may be even negated by the fact that rPI are running so much slower (single core) on a much bigger node (28nm for rPI4).

Size is a thing until you realize that small form factor/ single board computer also exist for AMD.

Long time support is everyone's guess, but with AMD you do have a wide range of HW AND can update single components at a time.

Quote:

Karlos wrote: And let face it a GHz class virtual 68K today is infinitely cooler than a real PPC from the mid naugthies.

I guess we have to agree to disagree on this one. Fake GHz 68k is what I had in 2001 (Amithlon) and it was nice at that time but thats all there was to it.

_________________
- We don't need good ideas, we haven't run out on bad ones yet
- blame Canada

@Karlos @Kronos


For me it has never been about the hardware, it has been about the people and the experience of it all. Even back to the Amiga in the 80s.

I can do all my work, effectively and run the 2 to 5 68k programs along with a full native solution with an ever improving experience.

The MorphOS Team are outstanding people, that have made using the platform a fun experience. I can report bugs that get fixed and if I have a important issue that needs to be addressed they give me direct answers so I can make decisions, since this is my work box I use as my daily driver for 80% of my tasks.

I have donated to the Team and other developers. With PolyOrga, it went from decent organizer, to a great CRM system over the last few years with Frank helping me initially when I was in trouble and Fred caring and adding good stuff and direction to it.

Jaca is the most famous programmer in the AmigaLand. His work with Wayfarer and Iris and MorphOS (multimonitor, smb, etc) is just phenomenal.

Even with AmigaOS4 being dead (no OS updates or enhancements, etc) users of the OS still hang together and are very helpful to each other with the hand they are dealt. That is a nice community and I can see why it is fun for them to work it out and stick together. Again because of the people and the experience.

Would I pay to see MorphOS to another platform? Sure. But, with the Mac PPC stuff being the absolute most powerful PPC that is readily available for free to a few hundred but, I'm good for now. Particularly as the MorphOS Team, esp. Jaca makes Wayfarer and Iris better and faster almost every other day now.

@Kronos

Amithlon was indeed amazing and it was what made me realise back then that PPC was a ultimately mistake. I had 68K code I'd written outperforming the PPC native version on a G4 800 MHz, on a machine costing a fraction of the price that G4 box did, and thar performance disparity has only grown since.

However, today, I dunno, I want something more. I want cake and the ability to eat it. That means I want to sit at a desk with the RSI inducing A1200 wedge, playing native chipset games via the Pi camera passthrough on a nice 1080p RTG display. Run OctaMED SS perfectly while using *all* the 64 channels, with no input latency or audio breakup which is the bane of UAE on machines far better equipped, all routed out through real Paula and her dirty nonlinear sound. An emulated 68K with a hardware interface into the system provides all this.

More importantly, I want a 68K ecosystem where a binary I wrote can run on the entire freak show menagerie of compatibles that exists today, including NG machines rubbing on PPC because 68K is the Lingua Franca of the Amiga, PPC was just an accident of history, but as long as NG machines are able to run 68K object code, they are arr still included.

Last edited by Karlos on 17-Dec-2023 at 09:56 AM.
Last edited by Karlos on 17-Dec-2023 at 09:50 AM.
Last edited by Karlos on 17-Dec-2023 at 09:48 AM.

_________________
Doing stupid things for fun...

Excuse the gazillion typos. I'm not well, lol.

_________________
Doing stupid things for fun...

@Karlos

Quote:

Karlos wrote: @Kronos Amithlon was indeed amazing and it was way made me realise back then that PPC was a ultimately mistake. I had 68K code I'd written outperforming the PPC native version on a G4 800 MHz, on a machine costing a fraction

Sure, but lets not forget that the Articia based systems were really slow compared to "proper" PPC HW, and even the Peg2 left a lot of potential on the ground.

I could stuff some PowerBook motherboard(or AMD/ARM) + KeyRah into a A500/600/1200 case if I wanted to but I fail to see the point.

As for 68k being the "Lingua Franca" that would also mean constricting myself to the AmigaOS3.x APIs and relying on questionable support in current compilers, so no I'll happily pass on that one too.

_________________
- We don't need good ideas, we haven't run out on bad ones yet
- blame Canada

@Kronos

Nothing there that couldn't be fixed. All it needs is the vision and inclination, but that's what's lacking most. Evidence of that being, well, here we are still on PPC. I mean, isn't that even longer now than the original Amiga line ran on for? Got to be catching up if not.

_________________
Doing stupid things for fun...

@Karlos

Quote:

Karlos wrote: @Kronos Nothing there that couldn't be fixed.

Nothing there that couldn't have been fixed a decade ago.

Just like moving from PPC to AMD or ARM.


Neither did happen, and only one has a realistic chance to happen within the current decade.

_________________
- We don't need good ideas, we haven't run out on bad ones yet
- blame Canada

@Kronos

And why is that?

_________________
Doing stupid things for fun...

@matt3k

Quote:

The MorphOS Team are outstanding people, that have made using the platform a fun experience. I can report bugs that get fixed and if I have a important issue that needs to be addressed they give me direct answers so I can make decisions, since this is my work box I use as my daily driver for 80% of my tasks.

I have nothing but respect and admiration for the devs working on it, also 4.x, 3.x and the multitude of AROS related projects. It doesn't mean I can't or won't question what I see as poor decisions.

I contributed to OS4, it doesn't mean I don't recognise the absolute debacle it is right now. It beggars belief that the MOS Devs would countenance supporting the vanity x5000, a machine made of pure unobtanium that has no chance of being anything but a footnote.

68K has a history and gravitas, a coolness factor. It's undergoing something of a retro renaissance too. PPC, with a few interesting offworld usages, will never, ever be anything more than an "also ran" architecture. Sticking to it is a guaranteed ticket to total obscurity. And maybe that's fine if that's what everyone wants, but when hardware dries up, it's game over.

When the last classic 68K Amiga dies, it will carry on in UAE and various standalone hardware configurations and people will still have fun with them.

How many people will stick to PPC/NG in QEMU I wonder?

Last edited by Karlos on 17-Dec-2023 at 11:00 AM.

_________________
Doing stupid things for fun...

@Karlos

Quote:

It beggars belief that the MOS Devs would countenance supporting the vanity x5000, a machine made of pure unobtanium that has no chance of being anything but a footnote.

AFAIK that was "sponsored" by Trevor, whether with actual money or just the HW I do not know.

It allowed bootstrapping PCIe support before the PCIe G5 were supported.
Sure that also applied for the SAMs.

In the end it tickled someone into doing it which is pretty much the reason why anything gets done in the community these days.

_________________
- We don't need good ideas, we haven't run out on bad ones yet
- blame Canada