kolla Quote:

It is actually the other way around. Hardware exists to accommodate software, it is software that "does the work". To come up with fancy new hardware and not doing it in coordination with software developers is at best very inefficient and at worst utterly pointless. This is why there so much dedicated software for the AC68080+SAGA (not!) and also why so many "Raspberry Pi clones" failed and are no longer around, even though their hardware were both faster and cheaper than Raspberry Pi.

I will not argue the chicken and the egg problem with you. Besides, the 68k Amiga still has some development software support and a large library of software. Some Amiga software has aged better than others but the retro 68k games are hot right now and 68k games are the best of 2D games before early 3D games which have not aged as well. The 68k and Amiga still have some development software support which has gained momentum due to the hot retro market but support will likely die without new 68k hardware. The old hardware is dying and emulations is not a compiler target.

The AC68080+SAGA is not competitive enough with a FPGA AC68080@100MHz and is losing sales to the competition like the PiStorm. "Raspberry Pi clones" are not compatible enough with the RPi standard system usually lacking the RPi GPU which was the same problem as PPC Amiga1 hardware but the PPC CPUs are now lower performance than using emulation and much more expensive. Competitive compatible hardware is required to attract customers.

kolla Quote:

The Amiga operating system wasn't initially developed on Amiga hardware.

What little Amiga software is developed today if often not developed on an Amiga either. Even worse, most Amiga software is not used on Amiga hardware either. EOL!

NutsAboutAmiga Quote:

Well the problem there is lot people disagree here, lots of people, keep writing code, that does not support plugins, different render options, or audio options, that pretty much stuck on AGA, and Paula sound.

It is not about "disagree" but about freedom, preference and compatibility. Some developers choose to old school bang the 68k Amiga hardware. PPC hardware was chosen not to be 68k Amiga compatible. New PPC software is not compatible with 68k Amigas either. Division of the Amiga market was a choice that weakened the already diminished Amiga market. New ARM and x86-64 AmigaOS and AmigaOS like ports will further divide the Amiga market and there is unlikely to be any new hardware. New low cost semi-modern 68k Amiga hardware could reunite the Amiga market instead and offer better compatibility.

NutsAboutAmiga Quote:

There is not just an application problem, there is an OS security problem as well, if we are talking about making AmigaOS an option to the main stream operating systems. one program being allowed to access another programs stack, the IPC system is also too open, and the OS invitation to patch and hack, track system API’s, some times wherry useful, for debugging, and for optimizing and customizing, but also horrible risky when comes to spyware. Also, we have the issue of system files that can be replaced and sometime already is replaced by competing companies,

PPC Amiga NG did not fix these security issues?

NutsAboutAmiga Quote:

We also have the issue of OS not being written for L1 cache, which is probably why we do not have stable SMP support on AmigaOS. Without multi core support, we are too lame. I also do not particularly like how multi monitor support work in AmigaOS, the screen idea worked great for one monitor, but not for two.

PPC Amiga NG did not add SMP support?

Are you saying cache coherency is the problem even though the X1000, X5000 and A1222 CPUs support hardware cache coherency between cores? The difficulty has nothing to do with PPC having one of the weakest memory consistency models of any desktop ISA ever? A stronger memory consistency model like TSO would not help considering AROS x86-64 using TSO has come the closest to Amiga NG SMP?

What did PPC Amiga NG add? Not enough to be NG? Not anything that the 68k AmigaOS can not do with more performance?

minator Quote:

I actually think the AxRT idea is the future. It gives you Amiga API compatibility but effectively replaces the kernel with Linux. So, it's running native on the CPU/s and it's not "hosted". Build a new Workbench on this and you'll have an Amiga environment that doesn't have to look or feel anything like Linux. Granted it'll probably break a load of stuff but any move modernise the system will do this. However, doing this means you get the memory protection, multi core support and 64 bit support. Best of all you don't have to care about hardware support, it'll run on whatever Linux runs on. That's a wide range whatever architecture you chose.

So divide the Amiga market more with another incompatible NG offering? Are "memory protection" and "multi core support" with isolated Amiga processes/tasks really a NG Amiga solution?

minator Quote:

My Raspberry Pi feels ridiculously fast for an Amiga, imagine what a native Amiga-like system running on one of the new Nvidia DIGITS machines would be like...

So basically like WinUAE on a 15 year old PC?

@minator

Quote:

minator wrote: @agami I actually think the AxRT idea is the future. It gives you Amiga API compatibility but effectively replaces the kernel with Linux. So, it's running native on the CPU/s and it's not "hosted". Build a new Workbench on this and you'll have an Amiga environment that doesn't have to look or feel anything like Linux. Granted it'll probably break a load of stuff but any move modernise the system will do this. However, doing this means you get the memory protection, multi core support and 64 bit support. Best of all you don't have to care about hardware support, it'll run on whatever Linux runs on. That's a wide range whatever architecture you chose. My Raspberry Pi feels ridiculously fast for an Amiga, imagine what a native Amiga-like system running on one of the new Nvidia DIGITS machines would be like...

I agree. The future is AROS and AROS related options such as AxRT.

However, MorphOS is the present. It runs on desktops and laptops. It has an excellent range of productivity software, and it gives Amigans the necessary dose of nostalgic je ne sais qoui.
Its main problem is that the hardware, while fairly inexpensive, is slowly dying.

I wish AxRT were as mature as MorphOS. And how crazy would it be if the MorphOS Team announced that they have created a MorphOS MUI run-time layer for Linux, a la AxRT.
Fantasy, but not out of the realm of possibility.

Last edited by agami on 15-Jan-2025 at 03:16 AM.

_________________
All the way, with 68k

@Matt3k

Quote:

Matt3k wrote: @Agami Sure, PPC is on life support for sure and they do need to get that done. But, if I look at my PowerBook G4 as an example... It is available, perfectly fast and responsive for everything but browsing and has similar issues to the G5 for larger sites. So for everything the speed and availability of processing is almost instantly responsive and cheap to acquire still. So I hope they further develop the software to save me labor from work still in a few areas and then focus on the shift. One simple thing like recurring meetings in Iris saves me time and work. Adding a new processor doesn't fix it. I don't disagree that new hardware is needed, I just don't think it is a first priority from my experience. If they put all their effort over the last 5 years to shifting cpu's the applications would be useless... Think we agree more than not on this :)

Yes. We are generally in agreement. We're really mostly debating the nuances.

I also have 15" 1.67GHz PowerBook G4. Sadly it has developed a super annoying backlight flickering issue, I just don't see it as a good use of my funds to spend money on repairing a machine where something else could fail any day.

Also, I don't just want to 'get by' on a MorphOS system. I want to "rock" the system.

_________________
All the way, with 68k

@kolla

Quote:

kolla wrote: @agami Quote: I'm here for the people. So you're an anthropologist?

Close. I'm a sociologist.

_________________
All the way, with 68k

@matthey

Quote:

There is a chicken and the egg problem for hardware and software. However, hardware is necessary for software so needs to come first. It is a large hardware user base that attracts developers. The Amiga 1000 lacked software until the more affordable Amiga 500 grew the user base enough to attract developers.

A1000 had been expensively manufactured in a Sanyo VCR factory. C= didn't use their best mass production manufacturing facilities when C= manufactured A1000.

In 1985, the C= hired a new CEO and president, Thomas Rattigan, who attempted to steer the company back to its retail roots. Rattigan settled on developing the A500 at the low end and
the A2000 at the high-end.

A500 was manufactured at Commodore’s Braunschweig (Germany) and Hong Kong facilities.

A1000's manufacturing was hobbled by then-CEO, Marshall F. Smith.

----------------
Steve Jobs courted major business software vendors such as Microsoft (for GUI version of MS Excel, MS Word), Aldus (PageMaker), Adobe and 'etc'.


https://youtu.be/PI23HAEN63c?t=508
The Rise of Microsoft Excel: Part 1, includes the rise of Lotus 123 beating SuperCalc and VisiCalc

Microsoft has to figure out a plan to beat Lotus 123 since the existing MS's text UI Multiplan wasn't enough to win against Lotus 123. Multiplan, SuperCalc, VisiCalc, and Lotus 123 are text-based UI.

In 1984, Mac was still struggling due to weak software e.g. lack of spreadsheet applications.
Microsoft's decision to ditch MS-DOS code for Odyssey (codename for Excel) was made in March 1984, two months after Mac 128K was launched. Bill Gates gambled on delivering a good GUI spreadsheet for Mac, Mac 512K road map, dominating Mac's spreadsheet market and using it as a beachhead for other GUI environments e.g. Windows. MS Excel team has cooperated with Mac 512K's unreleased task switcher. Bill Gates and Steve Jobs have a joint MS Excel for Mac launch.

Lotus delayed Jazz for Mac.

Mac 512K's unreleased task switcher with MS Excel countered Lotus Jazz's integration.
Microsoft has access to internal Apple's MacOS road map and beaten Lotus Jazz on market release. Microsoft is aware of 1MB RAM Mac Plus's 1986 road map for larger MS Excel Spreadsheet use case.

1986 Macintosh Spreadsheet Market
Excel: 160,000 copies sold with 89% market share.
Jazz: 10,000 copies sold with 6% market share.
Other spreadsheets: 5% cumulative market share.

Microsoft's Mac experience would be patterned for MS Windows 2.x/MS Excel plans.

QuarkXPress was released in 1987 for the Macintosh that was developed in 1986.

The "killer business app" sold the Mac when coupled with a stable high-resolution business GUI standard.

In 1986, the Amiga had Professional Page while the Atari ST had a Publishing Partner (PageStream).

Amiga OCS's stable high-resolution mode A2024 has a production scale of 5000 units which is a token i.e. good for demos, but not mass production.

Atari ST's stable high-resolution mode needs a special monitor or an expensive multi-sync monitor.

PC VGA has a built-in scan doubler for 200-line resolution modes, which promoted mass production for simpler 31 kHz VGA monitor clones.

Amiga ECS would encounter similar problems as Atari ST's expensive multi-sync monitor.
ECS Denise's 31 kHz/60 hz productivity mode targeted VGA monitors. ECS Denise's 31 kHz/50 hz display mode wouldn't work on strict PC VGA monitors.

_________________
Amiga 1200 (rev 1D1, KS 3.2, PiStorm32/RPi CM4/Emu68)
Amiga 500 (rev 6A, ECS, KS 3.2, PiStorm/RPi 4B/Emu68)
Ryzen 9 7950X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB

@matthey

Quote:

New low cost semi-modern 68k Amiga hardware could reunite the Amiga market instead and offer better compatibility.

show us this "low cost semi-modern 68k Amiga hardware"
native 68k with graphics at least at ps1 level at rational price
until you get it stop trolling and start working on it

@ppcamiga1

Quote:

show us this "low cost semi-modern 68k Amiga hardware" native 68k with graphics at least at ps1 level at rational price until you get it stop trolling and start working on it

I guess the market must be flooded with PPC then...

Last edited by pixie on 14-Jan-2025 at 09:05 AM.

_________________
Indigo 3D Lounge, my second home.
The Illusion of Choice | Am*ga

@Hammer

AtariST vga output for high resolution monochrome is usable with standard vga monitors. Is one of the things that amiga lacked, a good vga output directly from the board.

_________________
Amiga 1200 desktop. Apollo 030/50 Mhz 8mb ram + ClassicWB + Wb 3.1
Amiga 500 + ACA500plus 8mb + 30gb CF
Raspberry Pi 3b+ and Amibian 1.4
Mac Mini G4 1GB Ram with the butterfly!!

@kolla

I agree with you. But as a Amiga user that are trying to learn programming on the amiga for 68k workbench, I can say that it is not easy to find good tutorials on how to do it the right way. There is so much outdated material in the RKRM that you do not know how to do things. And also which compiler should I use,

the lattice C 6.58, or the GNU C 2.95, or the Storm C, each with its own problems and all three of them outdated.

I love the new book that Thomas did about Amiga DOS programming. I bought it right away just to support him. We need more of this initiative. I would love a rewrite of the RKRM books for the latest 68K OS.

and also a modern compiler that is still supported would be very nice.

BR
JJ

_________________
A1 X1000 is here !!!

@agami

I would just buy another PowerBook off of ebay, they are cheap and plentiful still. Shipping will be manageable. Or just hook an external monitor up to the laptop and you can setup dual monitor and ignore the laptop screen. I did that until I got my G5 the way I wanted it and it worked great.

If you end up getting a G5 PCIe, get the 2.0 or 2.3 as they are both single processor/dualcore. They run cooler and you can always swap out processor and put in a 2.5 from you dead system. That is exactly what I did. You can also rebuild the cooler as Chris Edwards did, it runs super cool now, but it is too much work for me. Chris did a video on the 2.0 PCIe as well that he updated the processor to.

Maybe I just don't use it enough and I'm not in the know, but I don't see AROS flavors being any more of a future than MorphOS. Last I used AROS I had to rabbit hole the good stuff, so I would just use a unix flavor at that point. Maybe, the world is different now though.

The sheer effort to get MorphOS where it is today with the applications and OS, that is a lot of heavy lifting that requires a tremendous amount of dedication to getting it done. 3.19 if I had to guess comes out this year is how many versions in itself at this point to provide the experience it does, now the hundreds of collective updates of the major apps of the os. I might just be in the dark, but I don't see it.

Last edited by Matt3k on 14-Jan-2025 at 09:26 PM.

Quote:

by agami on 13-Jan-2025 0:52:59 Quote: OneTimer1 wrote: Most people are here for the old software and nostalgia. I'm here for the people.

That was nicer than some people deserve.

Quote:

by agami on 14-Jan-2025 3:16:34 The future is AROS and AROS related options such as AxRT.

My hopes are on something with a Linux kernel and an Aros API and GUI on top, maybe AROS hosted, but maybe I should give AxRT a try. From its presentation it only looked like a development system for Aros programs that seemed me to be not AmigaOid enough.



Quote:

if the MorphOS Team announced that they have created a MorphOS MUI run-time layer for Linux, a la AxRT ...

They are to quite about what they are doing, the next version of Morphos may run on ARM ox x64 and it seems it will have a kernel different from before (Laire's Quark is not used any more) but there are no open discussions about it. The had demoed it on some shows but they should spread what they have as demo so people would know they are alive.

@matthey

Quote:

New PPC software is not compatible with 68k Amigas either. Division of the Amiga market was a choice that weakened the already diminished Amiga market.

You already know, this but for sake of repeating it, AmigaONE was supposed to be a accelerator card for the A1200, and it was not supposed to be a complete Computer, the failure was that Accelerator was never made. So last minute they found a replacement.

if they had gone with A1200 dongle, that meant ONLY A1200 uses can buy it, (excluding A4000 users like always and new users, I wont be here now, I be Linux user for sure.)
and that it software might been compatible, because the chip wont have been removed.

its not like 68080 / Vampire did not divide the 680x0 market, with new Audio chip, and with new instruction set, and its new graphics modes. And its special 3D chip.

And it’s not like A600GS / AmiBench, also has issues with incompatible graphics libraries and lack of support for picasso96 and cybergraphics, that almost all graphics programs use.

or the incompatible AROS 680x0 binaries that do not work on AmigaOS.

Or the ProDad P.OS programs that also do not work on AmigaOS.

There are lots of examples of incompatible or dead API’s that no one use anymore, like EGS, and RTG Master.

So you focusing in on PPC makes little sense here, but you do it to argue.

Quote:

What did PPC Amiga NG add?

If you’re interested in what new features that’s in AmigaOS4.x I’m sure you probably already read the change logs, and features list, but if you are not interested, I’m not going to google it for you.

Quote:

Not enough to be NG?

What is NG? Next Generation, it’s what is planed for the future, not what is currently available, I’m sure all dreamt about flying cars, and rocket that land on mars, or jetsons vacuum tubes. It turns out that not everything you dream about comes true, some things do come true, but it does not happen as soon as you want it to happen. Or when you wake up the next day.

I remember dreaming about A/BOX and BeBox, back in the day...
I also dreamt about BoXeR and Inside out.

Last edited by NutsAboutAmiga on 14-Jan-2025 at 09:17 PM.
Last edited by NutsAboutAmiga on 14-Jan-2025 at 09:14 PM.
Last edited by NutsAboutAmiga on 14-Jan-2025 at 09:06 PM.
Last edited by NutsAboutAmiga on 14-Jan-2025 at 09:03 PM.

_________________
http://lifeofliveforit.blogspot.no/
Facebook::LiveForIt Software for AmigaOS

@agami

Quote:

I agree. The future is AROS and AROS related options such as AxRT.

Well, I think the idea is the future, it gives the most features for the least effort.

Quote:

Its main problem is that the hardware, while fairly expensive, is slowly dying.

Exactly. I don't have MorphOS or anything I can run it on. It'd probably be pretty cool on the RPi.

Quote:

I wish AxRT were as mature as MorphOS. And how crazy would it be if the MorphOS Team announced that they have created a MorphOS MUI run-time layer for Linux, a la AxRT. Fantasy, but not out of the realm of possibility.

If anyone could do it, it'd be the MorphOS guys.


@matthey

Quote:

So divide the Amiga market more with another incompatible NG offering? Are "memory protection" and "multi core support" with isolated Amiga processes/tasks really a NG Amiga solution?

If you want an OS that in't stuck in the 80s, it's a necessity.

Quote:

So basically like WinUAE on a 15 year old PC?

What on Earth are you talking about?

Raspberry Pi 5 is as fast as a 10 year old PC, but it has to run a 68K emulator.
I was comparing it to a native system running on brand new hardware - The Nvidia DIGITS system has 20 of the latest Arm cores, a big GPU and 128GB of unified RAM. It's quite obviously designed to compete with Apple's Max chips so it's going to be a monster.

Now imagine if you had a modern Amiga-like system that could take advantage of it all.

Last edited by minator on 14-Jan-2025 at 11:03 PM.

_________________
Whyzzat?

pixie Quote:

I guess the market must be flooded with PPC then...

PPC AmigaNOne systems have the CPU performance of the lowest end SiFive RISC-V and ARM in-order CPU cores, hardware priced under $100 USD for a SBC, while having the price of high end x86-64 systems. It is like selling a low production golf cart for the price of a low production Lamborghini. Sadly, an older, cheaper and much loved 68k golf cart design was swapped out for an expensive crap PPC golf cart design that shares no parts with the original. The PPC Efika with a PPC603e@400MHz level of CPU performance and 128MiB of memory was higher end than any 68k system but low end PPC systems were notoriously bad, it did not sell well and it was replaced by ARM with Thumb-2 hardware which has a smaller footprint like the 68k. Thumb(-2) was licensed from Hitachi SuperH which is a RISC copy of the 68000 Hitachi 2nd sourced. The golf cart competitors were smart enough to copy the good old 68k design to make a better and cheaper golf cart but the old fat and inferior PPC design for the golf carts are used instead even after decades of hoping to sell the inferior design using the good old brand name. Fortunately, most customers realized the bait and switch and that the brand name was likely nefariously obtained so stayed away.

Interestingly, the AC68080@100MHz can come close to the performance of Efika PPC e300 CPU core in a low cost FPGA with the better memory performance. A $500 USD higher priced FPGA may allow a 150 MHz to 180 MHz clock and a $1000 USD higher priced FPGA may allow a 200 MHz plus clock and better memory performance which may outperform the Efika PPC CPU. The economics are worse than the low production PPC AmigaNOne hardware using outdated commodity hardware though. An AC68080 ASIC may clock to 2 GHz from an ASIC that costs $0.20 USD to produce. The very cheap transistors in an ASIC could be used to provide much more than 10 times the performance of a high end FPGA too. The value difference measured in performance/$ is as close as comparing a golf cart to a mass produced Porsche.

jonssonj Quote:

I agree with you. But as a Amiga user that are trying to learn programming on the amiga for 68k workbench, I can say that it is not easy to find good tutorials on how to do it the right way. There is so much outdated material in the RKRM that you do not know how to do things. And also which compiler should I use, the lattice C 6.58, or the GNU C 2.95, or the Storm C, each with its own problems and all three of them outdated.

I would throw Storm C out for the version of GCC included having bugs and offer VBCC as another alternative. There are newer versions of GCC which are useful for porting code but they are bigger, slower and unnecessary for creating new Amiga programs. Each Amiga compiler has advantages and disadvantages with none supported like they should be but there is no new 68k hardware to inspire development.

Lattice C 6.58
+ lots of existing C code and examples use it
+ good Amiga support and tools
+ fast compiling
+ mature and mostly bug free
- mediocre code generation
- no C99 support and poor C++ support
- not a cross-compiler

GCC 2.95.3
+ the best 68k integer backend ever gives excellent code generation
+ fast compiling compared to newer versions of GCC
+ has a mix of early C and GCCism support but older than C99
+ cross-compiler
- not Amiga friendly and difficult to install
- C++ support is old and should consider a newer version of GCC if needed

VBCC (latest version)
+ still developed and supported
+ good Amiga support (the two primary developers are or were Amiga owners)
+ cross-compiler and easy to build with few dependencies
+ easy to install for a cross compiler
+ basic C99 support
+ the best peephole optimizing 68k assembler in VASM
+ good FPU support, especially for the 68060 and 68040 where older compilers lack
+ good front end and support code
- below average backend code generation but partially made up by good support code
- slow compiling
- no C++ support

Disclosure: I developed for VBCC so I may be biased.

jonssonj Quote:

I love the new book that Thomas did about Amiga DOS programming. I bought it right away just to support him. We need more of this initiative. I would love a rewrite of the RKRM books for the latest 68K OS. and also a modern compiler that is still supported would be very nice.

ThoR did not write the new RKRM for financial gain as the 68k Amiga is EOL without new hardware. I could only offer a semi-modern compiler in VBCC that is still supported for the same reason. The 68k Amiga is fortunate to still have this considering the lack of 68k Amiga hardware. It defies logic almost as much as continued PPC support for an even smaller user base after more than two decades of pitiful sales.

@minator
@agami

I think what @agami wanted to say is relatively inexpensive. Not sure, will let him clarify.

After hating Apple stuff initially, I'm now a fan... It is all widely available even today and given how well they are built, no real issues in the next few years...

Where else can you spend $200 all in and have a solution for any Amiga flavor that you can install in 7 minutes and it just works and major software auto updates. The just working part is a huge accomplishment that I commend them for. It is a great experience to install and use it daily.

The PCIe single processor 2.0 and 2.3 are very well made and I have yet to see one even fail.

My Powerbook has been used for many years and is holding up just fine as well.

Granted the 2.5 QUAD systems had many issues, especially with the motherboards as they fixed the coolers that were bad in the PCIX systems.

The 2.7 is a real nice system if you put heatsinks from a 2.3 PCIX system or rebuilt the cooler. It is a trade off between the 2.7 PCIX and 2.5 PCIe as the PCIe has a faster bus but the 200MHz gain is nice as well.

I still think the 2.0/2.3 converted to a 2.5. Is just the best option, drives 2 monitors (as multimonitor was enhanced not long ago and still more improvements I believe are planned).

I think MorphOS really thought this out and worked very hard to get their. It would make sense that once you have a great program library and a mature OS, you can port that over and keep running. Everything else requires much more work to figure out the chicken and egg problem, where they already made the omlet... :)

NutsAboutAmiga Quote:

its not like 68080 / Vampire did not divide the 680x0 market, with new Audio chip, and with new instruction set, and its new graphics modes. And its special 3D chip.

The AC/Vamp hardware is compatible with 68k Amiga software so no division. Most AC/Vamp software that uses new features still works on 68k Amiga hardware as well like (RiVA and DevilutionX).

https://aminet.net/package/gfx/show/RiVA-0.54 Quote:

Two binaries are provided in this archive, a classic m68k compatible build (68040+) and an AMMX build, exclusive to Apollo Core Gold 2 and newer.

https://aminet.net/package/game/actio/DevilutionX Quote:

DevilutionX is open source, reverse engineered version of Diablo game (not included) by Blizzard Entertainment (1997). The Amiga port was created by Artur Jarosik and Marlon Beijer. ASM & AMMX optimization by Samuel Devulder. It is based on https://github.com/diasurgical/devilutionX contributed by various authors. ... For smooth game-play an Amiga with Vampire accelerator is recommended.

PPC NG systems emulate the 68k like ARM and x86-64 systems but slower and virtual Amigas on ARM and x86-64 hardware have built in chipset compatibility for better 68k Amiga compatibility. Some Amiga fans still hallucinate about the greener pastures of incompatible Amiga NG division instead of learning from past mistakes.

NutsAboutAmiga Quote:

What is NG? Next Generation, it’s what is planed for the future, not what is currently available, I’m sure all dreamt about flying cars, and rocket that land on mars, or jetsons vacuum tubes. It turns out that not everything you dream about comes true, some things do come true, but it does not happen as soon as you want it to happen. Or when you wake up the next day.

Promise the future and deliver the past. For example, SMP has been developed for over a decade now and even has an AmigaOS version number. The funny thing is that SMP may be easier on 68k hardware. Hardware development allows to fix things like the AC68080 allowing TAS/CAS and CAS2 locked RMW memory accesses to work. I believe it is also possible to get SMP working on 68k Amiga hardware. The problem is that SMP in a limited FPGA 68k Amiga SoC is almost useless because multiple cores would require too expensive of FPGA to be practical.

minator Quote:

If you want an OS that isn't stuck in the 80s, it's a necessity.

Isolated processes/tasks means minimal advantage from multitasking, no inter process communication and no ARexx. I'd rather have my 1990s 68k AmigaOS. We can have multi core support which the 68k Amiga hardware already supports with copper and blitter and we can have partial memory protection. SMP and improved security may be possible with better 68k Amiga compatibility too. The first thing NG systems want to do is throw away compatibility when it is the last thing you want to throw away for the Amiga market.

minator Quote:

What on Earth are you talking about? Raspberry Pi 5 is as fast as a 10 year old PC, but it has to run a 68K emulator. I was comparing it to a native system running on brand new hardware - The Nvidia DIGITS system has 20 of the latest Arm cores, a big GPU and 128GB of unified RAM. It's quite obviously designed to compete with Apple's Max chips so it's going to be a monster. Now imagine if you had a modern Amiga-like system that could take advantage of it all.

No?

year | CPU core(s) | DMIPS/MHz/core | memory | fab process
2024 A18/? ? LPDDR5X 3nm (Apple A18 SoC)
2018 Cortex-A76 6.6 LPDDR4X 16nm (RPi 5 SoC with OoO CPU cores)
2012 Cortex-A53 2.3 DDR2 40nm (RPi 3 SoC with in-order CPU cores)
2011 P5020/e5500 3.0 DDR3 45nm (X5000 SoC with OoO PPC CPU cores)
2009 PPC460EX 2.0 DDR2 90nm (Sam460 SoC with OoO PPC CPU core)
2008 QorIQ-P1022/e500v2 2.4 DDR3 45nm (A1222 SoC with OoO PPC CPU cores)
2008 Core_i7_920 7.0 DDR3 45nm (1st gen OoO Core i3/i5/i7/i9)
2007 PA6T 2.2 DDR2 65nm (X1000 SoC with OoO PPC CPU cores)
2005 MPC5200B/e300 1.9 DDR 90nm? (Efika SoC with OoO PPC CPU core)
2002 PPC970 2.9 130nm DDR (G5 OoO PPC CPU)
1997 PPC750 2.3 260nm (G3 PPC CPU)
1995 PPC603e 1.4 preSDRAM 350nm (OoO PPC CPU)
1994 68060 1.8 preSDRAM 500nm (in-order 68k CPU released in 1994)
1993 PPC601 1.4 preSDRAM 600nm (OoO PPC CPU)

The 1st gen Core i7 specs are closer to the X5000 P5020 SoC specs with similar chip fab process and memory but the Core i7 having more than twice the performance. The Cortex-A76 has a decade newer silicon and still can not reach the 1st gen Core i7 in performance even though it more than doubled the X5000 PPC P5020 performance. Do you understand yet what makes CISC CPU cores so powerful or do you need me to explain it again? Would you believe that x86-64 has more variations in common instruction length, less optimal instruction alignment and less orthogonal of an ISA than 68k too? Do you at least understand how impressive the performance of the CISC cores on the list are considering the specs, especially the chip fab process?

Last edited by matthey on 16-Jan-2025 at 03:02 AM.
Last edited by matthey on 15-Jan-2025 at 06:41 AM.
Last edited by matthey on 15-Jan-2025 at 06:20 AM.

@Matt3k

Quote:

Matt3k wrote: @agami I would just buy another PowerBook off of ebay, they are cheap and plentiful still. Shipping will be manageable. Or just hook an external monitor up to the laptop and you can setup dual monitor and ignore the laptop screen.

I have a Mac mini 1.5GHz with MorphOS.

As I mentioned before, and maybe I'm spoiled by my use of modern Linux/Gnome/Wayland, macOS, and Windows 11, but the noticeable lower performance of 1.5GHz G4 is turning me off. And I don't want to be turned off.

I know many people find the performance of MorphOS on G4 to be more than fine, but I find the older I get the less patient I am.


Quote:

If you end up getting a G5 PCIe, get the 2.0 or 2.3 as they are both single processor/dualcore. They run cooler and you can always swap out processor and put in a 2.5 from you dead system. That is exactly what I did.

The very second one of these G5 PCIe machines pops up on my remote continent island, I will be doing exactly that.

Last edited by agami on 15-Jan-2025 at 03:38 AM.

_________________
All the way, with 68k

@minator

Quote:

minator wrote: @agami Quote: Its main problem is that the hardware, while fairly expensive, is slowly dying.

I meant, and corrected to say 'fairly inexpensive'.


Quote:

The Nvidia DIGITS system ... Now imagine if you had a modern Amiga-like system that could take advantage of it all.

I do that every day, and have done so, in vivid detail, for the past 25 years.

_________________
All the way, with 68k

@matthey

you do not have any ppc amiga and anything you wrote about it is
is as usually pure propaganda bs

AC68080@100MHz is 5 times slower than Efika

@matthey

Quote:

What did PPC Amiga NG add

It was explained many times.

native os, cpu many times faster than fastest 68060 and 68080, fast 2D, fast 3D, working MMU, working FPU

it is at least as fast and as comfortable as cheap pc from win 95 era
something that real 68k never reach