@manga303

Quote:

manga303 wrote: Yes find all this together and a PiStorm with a Pi4 or Pi5 is burning like 2000Mips.

Just a clarification, PiStorm and PI5 will most likely never work as they changed the GPIO pins in a way making into unusable for this use case.

Maybe some PI clone will work, maybe they go back to the old way with PI6 but as of today PI4 is end of the line.

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

@Hans

Quote:

At least nobody I know who's still an Amiga computer fan. It's about letting go of those parts that don't make sense any more (like the old custom chipsets)

With the right abstractions, there should be no reason why supporting legacy chips ought to be a problem. I'm in a minority view here but one of the reasons why I think OS4.1 is still "AmigaOS" proper, aside from the fact I've seen the repository and history and *know* the lineage, is the fact it still runs on my A1200 and I can still run a lot of metal banging old code on it

That's a strong sense of connection. Shame to lose it.

_________________
Doing stupid things for fun...

@Kronos

Thank You very much
it is good to know that PI4 is the last

@ppcamiga1

Don’t worry my darling. Raspberrypi will keep building and selling Pi4 and CM4 until 2034 - so you need to keep taking your pills for at least 9 next years ;)

Or longer since we will at some point switch to something better than RaspberryPi ;)

cpu other than in pc
is the last thing from original amiga hardware
that I really care
I will be using it to the end

but may consider x86 or arm amiga like solutions on it
if it will be no more than 15 years behind windows

@michalsc

szulc stop trolling start working on mui on aros

@ppcamiga1

I don’t care about you, remember? ;)

@michalsc

szulc stop trolling start working on mui on aros

@ppcamiga1

Do it yourself. Oh you cannot? Not skilled enough? Sorry to hear ;)

No, actually not sorry at all. I have a good laugh reading your troll posts :) keep up the only thing you are good at

@Kronos

Pi 5 doesn't work with PiStorm as is, but then there's nothing preventing it from running a standalone Emu68. It wouldn't be the first 68K Amiga clone without the custom chipset.

That said, the main attraction of the PiStorm is the fact that it plugs into your actual Amiga.

_________________
Doing stupid things for fun...

@Karlos

I like the idea to have a 68k amiga system based on arm hardware with emu68 as layer. That could be potentially a fairly powerful system costing not much more than a RPi. Some will moan because of no chipset but I think it adds more interesting options. The more the better I would say

@Karlos

Quote:

Karlos wrote: @Kronos Pi 5 doesn't work with PiStorm as is...

That's a pity. The Pi guys need to fix this if they want Pi 5 to have broadest appeal.

OTOH Pi 4 is enough surely? Oh wait, I forgot. Nothing is ever enough for an Amiga fan!


Quote:

That said, the main attraction of the PiStorm is the fact that it plugs into your actual Amiga.

Indeed. Many of us have zero interest in a standalone 'solution'.

@bhabbott

We're a funny old lot. You can please some of the people all the time, all of the people some of the time and your Amiga user base almost none of the time...

_________________
Doing stupid things for fun...

michalsc Quote:

I don’t see much place for AOT benefits in case of amiga software for m68k, mostly due to lack of sufficient amount of metadata in executables. Furthermore, things like postponed writes or instruction reordering in order to get more performance can be achieved in JIT too. So no, I do not see a necessity for AOT translation.

I agree that AOT can not be a complete solution for using existing 68k Amiga software on ARM. Metadata could be included for new software. Debugging data identifies function offsets in the code and improves disassembly accuracy but is still lacking for AOT compilation/patching. I improved ADis with the goal to make most AmigaOS friendly 68k programs accurately disassemble and reassemble with optional peephole optimizations so I have a pretty good idea about this. The 68k VLE increases the difficulty but is much easier than x86 disassembly. Even fixed length encoded PPC is difficult to reliably AOT patch as PPC executables patched for the A1222 due to the lack of a FPU is not reliable either. I am not advocating for AOT, just saying it has advantages over JIT that have more recently made it preferable in applications where pre-compiled executable are not possible. JIT can have issues too and more issues with more optimizations. I want real hardware and native executed code that provides superior value, receives compiler/developer support and has a future. JIT/AOT compilation is tolerable for old software but trying to move the Amiga forward with it is cringe worthy.

OlafS25 Quote:

there is the retro branch that is connected with 68k and will live in coming years there were OS versions running on faster hardware but inheriting most of the limitations like no memory protection and 32bit

The retro 68k Amiga is here to stay alright. It is the majority of the Amiga user base. Ignoring 68k Amiga compatibility will result in a repeat of the PPC AmigaNOne failure.

OlafS25 Quote:

so NG today is somehow retro too, just not feels like but to get something modern you must make a break. But even that not automatically gives you more developers, drivers and software, you just look better on paper NG OS as a platform on its own has no chance in my view

The retro 68k AmigaOS and hardware spec is marketable while PPC AmigaOS 4 and NG hardware is dead with a replacement needed. An Amiga NG without adequate compatibility is sure to repeat the mistakes of PPC AmigaNOne and Amiga NOwhere. The potential Amiga NG options are listed below.

1. Build a virtual NG AmigaOS
2. Build a new NG LE AmigaOS for ARM and/or x86-64 with sandbox 68k Amiga compatibility
3. Build AmigaOS NG on top of new compatible 68k Amiga hardware

Option #1 above seems to be the current direction for the AmigaKit A600GS and A1200NG motherboard but I believe it is the worst option. There is nothing NG about it as it does not add 64-bit support, memory protection/MMU support, ISA enhancements, SMP, etc. Adding NG support for a virtual machine is very difficult and will receive little support as it is amateurish. It requires the least investment but I do not see anyway to an Amiga NG.

Option #2 above is a legitimate choice and perhaps the best choice from a technical perspective but it moves the NG Amiga away from everything the 68k Amiga was and further than the PPC AmigaNOne which failed because it was not compatible enough making it a very risky choice from a marketing perspective. The AmigaOS is only 1/3 of the Amiga with the 68k CPU and chipset ignored. I do not see anyway to sell a NG AmigaOS for enough to pay for development and there is AmigaOS competition like the free AROS x86-64 and likely MorphOS x86-64/ARM by the time a port would arrive.

Option #3 above may be the most difficult choice from a technical and financial perspective but it is the best choice from a marketing perspective. Building on existing 68k Amiga hardware has compatibility and economy of scale advantages as both 68k retro and NG users use a different AmigaOS but the same hardware. Hardware customization for the CPU and chipset may improve compatibility and make NG features easier. I believe this option would receive the most Amiga user and fan support as it is the most faithful to the original Amiga spirit. Natami and AC/Vamp FPGA projects have/had strong support but came up short by failing to plan for and create an ASIC. Existing 68k CPU and chipset cores already exist potentially reducing the work and time needed. This is the only Amiga NG choice I believe is viable.

@ppcamiga1

Quote:

@michalsc szulc stop trolling start working on mui on aros

Maybe when you're paying him €50 an hour, you can order him what do with his time.

Last edited by Rob on 23-Dec-2024 at 11:52 PM.
Last edited by Rob on 23-Dec-2024 at 11:52 PM.

@ppcamiga1

Quote:

ppcamiga1 wrote: @michalsc szulc stop trolling start working on mui on aros

Classic AmigaOS 4.1 FE (PowerPC) doesn't pass the Deluxe Music 2.0 test.

AROS 68K doesn't pass the Deluxe Music 2.0 test.

_________________

@matthey

option 3 is dead in the water and not as many people want it as you seem to think.

_________________
Support AROS sponsor a developer.

"AROS is prolly illegal ~ Evert Carton" intentionally quoted out of context for dramatic effect

@matthey

Quote:

GPU intermediate languages for unified shaders also switched from JIT to precompiled AOT like compilation. Javascript JIT is still used but there are alternatives. JIT is losing popularity due to lack of efficiency/competitiveness and whole virtual machines are practically dead. If performance and value are the goals, native code for real hardware can not be beat.

For the PC, 1st startup shader compile still occurs for games like Unreal Engine 5 based The First Descendant. Ryzen 9 7950X's 16 cores with 32 threads have an advantage i.e. all 32 threads are active during shader compile process. The compiled shaders are saved on mass storage.

For each game, 1st startup SPE compile also applicable for RPCS3 (PS3 emulator) with all 32 threads being active on Ryzen 9 7950X.

Game consoles don't have this problem since the target hardware is known.

---------

For PC games, Denuvo DRM can also use a virtual software CPU with a custom instruction set as a defense against typical reverse engineering methods. PC has to brute force these games with Denuvo DRM with a virtual software CPU. CPU's large cache size is important for these use cases.

_________________

@matthey

Quote:

The retro 68k Amiga is here to stay alright. It is the majority of the Amiga user base. Ignoring 68k Amiga compatibility will result in a repeat of the PPC AmigaNOne failure.

1. FYI, AmigaOS 4.1 FE has a userland 68K emulator.

2. The major problem with AmigaNG PowerPC is the performance vs cost issue when compared to the competition.

AmigaOne PowerPC's price failed the A500 aspect.

3. When AmigaOne PowerPC was released in the early 2000s and the internet's video social media revolution, AmigaOne PowerPC didn't have Amiga apps that made Amiga a winner in the video editing market niche. Amithlon X86's fast 68K emulation doesn't solve this aspect.

All Amiga video NLE software is anchored to their respective 3rd party hardware addon dongle in the form of a Zorro II/III card. Wedge 68K Amigas wouldn't be able to participate due to missing the Zorro II/III slot .

I'm aware of Toaster Flyer and MovieShop NLE software.

Having 68K doesn't guarantee Amiga compatibility which made Amiga a winner in the video market niche.

68K Mac's Adobe Premiere NLE 68K can still run on Emu68-enhanced Shapeshifter/Fusion and old-world PowerMacs.

For Mac68K NLE, I tested Avid VideoShop 2.0.3 and Adobe Premiere 4.0.

The Amiga is not Mac.

---------------------------
In the early 2000s, my K7 Athlon XP-based gaming PC's GeForce 4 Ti VIVO (video-in, video-out, s-video out, s-video in) included an MPEG2 video editing software bundle and nForce 2's motherboard provided FireWire port. GeForce 4 Ti includes MPEG-2 acceleration.

My GeForce 4 Ti 4200 VIVO is Leadtek WinFast A250 LE MyVIVO
https://www.neoseeker.com/Articles/Hardware/Reviews/winfasta250le/
MyVIVO Features:
MPEG I/II video compression Codec,
MGI Videowave 5.0 Software for video editing from various video sources.
GeForce 4 Ti 4200 VIVO can be overclocked into Ti 4600 level.

A gaming PC nuked video editing workstations like DraCo.
A gaming PC with a GeForce card nuked SGI's 3D workstations.
NVIDIA used economies of scale in PC's gaming market.
The same for the MacOS platform when it expanded beyond the DTP market.

I just don't stare at the OS.

Last edited by Hammer on 24-Dec-2024 at 05:04 AM.

_________________

@matthey

Quote:

Option #3 above may be the most difficult choice from a technical and financial perspective but it is the best choice from a marketing perspective. Building on existing 68k Amiga hardware has compatibility and economy of scale advantages as both 68k retro and NG users use a different AmigaOS but the same hardware. Hardware customization for the CPU and chipset may improve compatibility and make NG features easier. I believe this option would receive the most Amiga user and fan support as it is the most faithful to the original Amiga spirit. Natami and AC/Vamp FPGA projects have/had strong support but came up short by failing to plan for and create an ASIC. Existing 68k CPU and chipset cores already exist potentially reducing the work and time needed. This is the only Amiga NG choice I believe is viable.

Dead in the water i.e. need a company with a revenue size of PRi e.g. semi-custom RP2040 for the microcontroller market and independent from Broadcom'ss ARM-based SoC.

30,000 Amiga users pitch in $50 would be $1,500,00.

30,000 Amiga users pitch in $100 would be $3,000,000.

30,000 Amiga users pitch in $1000 would be $30,000,000.

10,000 Vampire users x $500 would be $5,000,000.

68060 will need modernization work.

_________________