Matt3k Quote:

I agree in that AmigaNG has become Amiga for me as well. Classic is used for fun and NG is used for work and entertainment.

PPC AmigaNG has been out for over 20 years compared to 68k Amiga for less than 10 years. The last memory is at the forefront of the mind too. For hundreds of thousands of people, the Amiga is a 68k Amiga which THEA500 Mini sales show. The PPC Efika was almost as cheap as THEA500 Mini at one time and didn't sell enough to stay in production. I expect the majority of NG users have left the AmigaNG market long ago and likely only number in the low to mid thousands counting MorphOS and AmigaOS4 together. MorphOS should be helped by a port to x86-64 as PPC is long dead but it will probably remain a very niche product as they will try to survive by selling an OS where most of the competition is free and/or has a huge advantage in software.

Matt3k Quote:

I use it as my daily driver with dual screens on a PowerMac 2.5GHz PCIe and PowerBook 1.67 17 systems and they do a great job. For me the Amiga has become more about the software as the day of bespoke hardware that was mindblowing in Amiga Land is dead and gone sadly. We are asked to over pay for slow hardware that a simple PPC Mac would blow away at pennys on the dollar simply to reclaim losses at this point.

It is true that old used PPC Mac hardware is cheaper than niche AmigaNOne hardware but the old PPC Mac hardware does not "blow away" the AmigaNOne hardware. The AmigaNOne high end X5000 surpassed a G5 Mac in many performance categories and the supposedly low end A1222 has better integer performance and memory performance than the majority of PPC Mac hardware, most of which was low end and gave PPC a bad reputation.

CPU core | DMIPS/MHz/core
---
CellPPE 0.6 (PS3)
Xenon 0.6 (XBox360)
PPC603e 1.4
PPC601 1.4
68060 1.5
ColdFireV5 1.8
PPC7410 1.8 (early G4)
e300 1.9 (Efika)
PPC440/460EX 2.0 (Sam440/460)
PA6T 2.2 (X1000)
PPC750 2.3 (G3)
Gekko 2.3 (Nintendo Gamecube)
Broadway 2.3 (Nintendo Wii)
Expresso 2.3 (Nintendo Wii U)
QorIQ-P1022/e500v2 2.4 (A1222)
PPC970 2.9 (G5)
P5020 3.0 (X5000)

There has been some progress in PPC integer performance from PPC Mac days but it has been especially poor considering the chip process improvements. After PPC failed on the desktop, PPC designs were targeted at the embedded market where low power and area are more important than performance. The 68060 also became a balanced embedded design after Apple moved to PPC yet the performance is very good on the ancient 500nm chip process. The Pentium P54C was measured at 1.88 DMIPS/MHz yet the 68060 outperformed this Pentium by 40% in per MHz integer performance with the ByteMark benchmark and no 68060 specific instruction scheduling (40% more than 1.88DMIPS/MHz is 2.63DMIPS/MHz). At most, PPC only doubled the 68060 integer performance as the 68060 to the PPC P5020 is a roughly 450nm chip process advantage (91% shrink) while from the P5020 to a RPi 5 Cortex-A76 at 6.6DMIPS/MHz more than doubling performance there is only roughly a 29nm chip process advantage (28% shrink). There were PPC memory improvements, multi-level cache improvements, clock speed improvements and other integration improvements with the chip process improvements that significantly increase performance but PPC core performance didn't see much improvement, even with cores designed by IBM.

Matt3k Quote:

I remember when 2.0 came out and many who were used to 1.3 hated it because it didn't 'feel' like AmigaOS to them anymore, but after they got used to the mass amounts of changes it was enjoyed and adopted. That is how I see MorphOS, it took the best aspects of 2/3.x and took it to the next logical level. and made it useable and developable. They used MUI, Turboprint, and other core elements that were the right choice to build on.

I liked AmigaOS 2.0 enhancements from the beginning but it was buggy and rough until AmigaOS 3.1. MorphOS has pioneered and innovated AmigaNG enhancements while AmigaOS 4 originally remained more faithful to AmigaOS 3. Both strategies have their advantages and disadvantages while splitting the tiny AmigaNG market hurt both.

Matt3k Quote:

Sadly stakeholders have just squandered everything when they should have just backed MorphOS in the 90's as it should have been instead of the madness that followed. That is where I would spend my money if I was Hyperion or other stakeholders, right into MorphOS as it really should have been done in the first place.

It may happen yet if Ben is successful at continuing to hold AmigaOS 4 hostage. There are other options to build on AmigaOS 3 including the Enhancer software and AROS. Where the AmigaOS is to go is the big question though. MorphOS will join AROS in supporting x86-64 giving the best performance option. AmigaOS 4 could be ported to ARM for low cost hardware, become emulation only or create 68k Amiga hardware. While ARM hardware is the obvious choice that so many people call for, the AmigaOS on a RPi with little software probably doesn't make the top 5 OSs and maybe not the top 10 in that crowded OS space. The Amiga is at its best on 68k hardware with chipset compatibility and emulation won't attract enough developers and users to make the platform successful.

Matt3k Quote:

TBH, AOS4 has made just to many mistakes and will just continues to do so. I know many who left AOS4 and they will not be coming back from users to developers gone forever. Slow and steady with MorphOS won the OS race for NG Amiga and they kept the band together and got the job done for over 25 years with major amounts of updates and point releases. Where AOS4 had very very little OS/Toolset/productive app development and has chased creating hardware without finishing drivers, started multiprocessor for dead hardware and not finishing it, releasing hardware 4 times the estimate price many years later and again likely not to have all the drivers worked out, start yet another reboot of just the core OS to get away from a lawsuit, we can go on and on with the madness and pretend it will get there... But to what end? Reality is reality at the end of the day and seeing it what it is and making decisions on that as apposed to feelings from the 80's makes sense and will be the best outcome we can hope for. Even it they get multiprocessor working, who will buy the hardware? They really should have seen this and invested in the OS/Tools/Great native apps, now it's too late. Hence why they just see it for what it is and invest in MorphOS and everyone wins.

The AmigaOS situation is a disaster that drives away potential customers for sure. Development is divided between 3 Amiga IP squatting businesses and they practically ignore the true owner of the Amiga IP. On top of that, PPC has long since been dead and buried by ARM with likely no porting work done to another architecture. Then there is the extremely overpriced and underwhelming niche hardware that fewer and fewer AmigaNG people are buying, especially now that they have made it possible to get better AmigaOS 4 performance and value through emulation. I agree than SMP wouldn't solve much as one x86-64 CPU core or RPi 5 CPU core has more than double and up to triple the integer performance of a PPC core (worse for FPU/SIMD performance especially in the case of the A1222 which likely has worse performance for standard PPC executables than software floating point). SMP is still a no go in the latest AmiWest videos too. It's funny how only one developer on the ExecSG team was mentioned as knowing PPC assembler as they are trying to debug SMP. PPC smart compilers were supposed to give PPC the edge without needing easy to read assembler but they were an anchor for PPC instead. In the end, a few thousand SMP AmigaOS PPC machines is not going to make a difference even if they get SMP working and selling a few hundred more antiquated machines every few years is nothing. Trevor marches on oblivious to the Amiga world though.

@matthey

All great points, agree with almost everything you said.

Both the PowerMac 2.7GHz PCIX and 2.5 GHz have altivec and MorphOS makes serious use of that for multimedia and that will impact the results that the on paper results show. The PCIe systems have faster buses, cache, and memory than the PCIX, but for most everything I have tested the 2.7 PCIX for most everything is faster in normal use than the PCIe systems even though it has everything faster (other than cpu clock) including faster video card.

The 2.7 works like a modern systems for almost everything you do. It plays higher density HD content without problems and even lower H.265 content. This system is incredibly fast and even browsing is almost useful on large websites a real testament to just how good the OS and toolset it.

The X5000 running MorphOS runs at a decent speed for sure, but side by side with a older 2.0 GHz PowerMac or a PowerBook 1.67 will be hard to determine a difference. Browsing might still be faster on the PowerMacs since Altivec is used.

I like 2.1 and adopted that and skipped 2.0. I still like 2.1 over 3.1 for some things. Great analogy with MorphOS pioneering and enhancing.

I will say, unless something has changed in the last 3-5 years or so when I last tried 4.1. To me it isn't a complete OS and was painful to even get going. I couldn't see a difference from when I first saw a demo of it many years prior. The OS is stale and stagnant to me and the native apps are missing to put it kindly. It is like 3.x in that it feels like I'm in 1995 in using it and that isn't a good thing.

MorphOS is really on another level. The OS itself has been improved with over 30 or so releases, and I can go back and install an old 1.4 or 2.x version and the evolution of the OS is apparent everywhere. You can tell the team thought about it and methodically worked on it.

You can install the OS in under 5 minutes and you almost have everything you need. The major apps auto update. Accessing SMB is very intiuitve and easy. The native apps like vpdf, jukebox, Iris, Wayfarer, and PolyOrganzer are all trophy apps that all work very well and have very modern features. In AOS4 equivalents just don't exist or if they do, they work like they were coded in the 90's and weren't much fun for me.

AOS4 will never get to where MorphOS is today, the canyon is just too wide now. The secret that MorphOS understood is that the secret is the software and the OS with it toolset and not the hardware and that brings us to today.

It is obvious from the outside looking in. AOS4 is a dead end and doesn't have any applications to make it enjoyable to use. Great thoughts on multicore... I agree if they ever get multicore support working for the x5000 (where 1000 and 1222 will follow - may take a long time with a part time employee as the only resource) who cares anymore? Great I'll load ibrowse with dual core or another app from 1997. As I have said, this is all madness to me at this point. It will be far inferior to MorphOS for about everything and will still probably be slower.

MorphOS is a serious business system for me that I use daily and it does it like a modern system, with AOS (3 and 4) when I tried it just couldn't do it at all. I wish the stakeholders were not oblivious like we mention, there is a real opportunity ready to go that would provide real benefits to the community. I think it also might be pride that they want to make it theirs, and I can understand that but it comes at the expense of others.

The MorphOS Team helped me along the way and I even donated to get a few items where they needed to be. They are just awesome to work with a really seem to care about their OS and users. I hit a roadblock with PolyOrga for work and a MorphOS developer fixed it, not even the original coder (who also improved it afterwards and it onboard to continue).

It is cool that ARM has really offered some nice options and look forward to it, but MorphOS is the way to go to do something productive with it.

Last edited by Matt3k on 16-Oct-2023 at 06:04 AM.

@NutsAboutAmiga

>I also agree that writing software for emulators make little sense, but then again lots of 680x0 >software has been developed on PC’s for years using cross compilers.

You do not need an Emulator to use cross compilers.

The cross compiler outputs AmigaOS 4 Binaries, no Emulation involved. You then transfer the compiled binaries to a physical Amiga (for example using an USB Stick or using ftp or other transfer locally) and test it there.

It also has the advantage that if the test exe crashes it does not annoy your build machine.

I actually have been using cross compile already 20 years ago.

Currently I am using Cross Compilation for AmigaOS targets, but do not have an Emulator installed. I test with a physical Amiga (an x1000).

MagicSN

@geen_naam

The problem I see with emulation is that ports will be inattractive as they suddenly compete with the native versions on these systems. Of course the worst this is for games. But the problem exists for all kind of software.

I have no problem with my programs running on Emulation (and am happy to hear if they work, maybe at one point might even test if they work - currently I have no Amiga emulation on my PC, I do not need it, as with the x1000 I have the real thing) but I have severe doubts that Emulation can carry an Amiga Software market (asides from OpenSource projects).

You mention on "Classics and Vampires" as "for those who want the real thing". I have my severe doubts these things are fast enough for what I am doing at least. Though keeping an open mind for them.

Currently the only thing which can run my ports is OS4 running on a real PowerPC based Amiga. It might work on QEmu, I do not know, I never tested it. But have thought about testing it for some while. Not because being interested in it, but because others might be interested in it. Still I doubt this is really a market.

As a longterm perspective I only see three ways: 1) new PowerPC based Amigas on a performance level beyond x5000 or at a very attractive price point (or possibly that Notebook PPC Amiga project if it takes off) 2) a high power redone-68k Amiga - no, not Vampire. I said "high power". Which means at performance level of x5000 or at least x1000 and with a state-of-the-art 3D Chipset. And no, I do not mean PiStorm either. PiStorm still relies on really old hardware. 3) ARM Amiga (and I do not mean Emulation here). Basically an "ArmAmigaOne" (pronably a dumb name, in my native language the word ARM means "poor").

I personally do not really see a perspective beyond these three. Might even be 2 of them or all 3 at the same time.

Last edited by MagicSN on 16-Oct-2023 at 09:28 AM.

@Matt3k

i have all systems OS3.2 & Aros for 4st Vampire + morphos on macmini + OS4 on sam460ex.
also had many other machines for morph OS and 0S4 like x1000 for many years..

About OS system function difference.. i admit blue camp had steady progress and better structure and planning.. and it's a OS that looks great and maintained with care and love. maybe a bit bloated..anyway..the whole presentation is excellent at it's home page.

while both camps have many useful features they don't share with each other. and thats something i wish both just had, best of both worlds..

i agree OS4 looks like OS3 and works like it.... and here is the real deal for me, i want the control i am used too. and some configuring you get os4 very nice.

and while Morph OS is OS3 like system, the is too much different in MorpOS that i never quite fell home..

For me it's a system that need me to challenge my computer skills.. not like my 7900x 64gb 6800xt gpu 3 x 2tb m2 7000/7000 mb/s, and i only use youtube + mp3 + films

nox

_________________
nox's in the house!

@MagicSN

Quote:

MagicSN wrote: @geen_naam As a longterm perspective I only see three ways: 1) new PowerPC based Amigas on a performance level beyond x5000 or at a very attractive price point (or possibly that Notebook PPC Amiga project if it takes off) 2) a high power redone-68k Amiga - no, not Vampire. I said "high power". Which means at performance level of x5000 or at least x1000 and with a state-of-the-art 3D Chipset. And no, I do not mean PiStorm either. PiStorm still relies on really old hardware. 3) ARM Amiga (and I do not mean Emulation here). Basically an "ArmAmigaOne" (pronably a dumb name, in my native language the word ARM means "poor"). I personally do not really see a perspective beyond these three. Might even be 2 of them or all 3 at the same time.

I think most reasonable people can agree that option 1) is out.
That said, I agree that the other two options are the "longer" term options. Though I'm not sure why x86-64 would be excluded, but I guess as it would not be very different to option 3) it can be considered a variation of the option.

So while hopefully someone is working on the longer term options (I'm looking at @michalsc for 2) and MorphOS for 3)), QEMU + Virtio can be an interim solution. One that keeps the wheels spinning until the other things are ready.
It's not for me as I have no interest in AmigaOS 4, but I can see how if executed right, it can provide some life-support for that ecosystem.

_________________
All the way, with 68k

@MagicSN

Quote:

The problem I see with emulation is that ports will be inattractive as they suddenly compete with the native versions on these systems.

22 years later, and it's still the same non-sensical argument.

How has that been working out for the last 22 years?

Don't conflate whole system emulation with CPU-only emulation/virtualization.

_________________
Doing stupid things for fun...

@MagicSN

Quote:

The problem I see with emulation is that ports will be inattractive as they suddenly compete with the native versions on these systems

Now there's a demonstration of confidence in your product if there was ever one.

If people like a system they'll use it and its software, even if that software exists elsewhere.
The exception being when there's benefit to using said software elsewhere. If a person has a pc theyre going to use software that takes advantage of the hardware resources they provide on the pc, not any form of AmigaOS.

Even the absolute fastest of PPC systems that can be used in the Amiga-verse are, in comparison to modern hardware, completely and utterly retro.
To argue that Amiga versions of software running vastly inferior under emulation to how theyd run on a pc is a reason not to use it is an arguement that people shouldnt be using ppc systems at all.

Emulation and virtualisation is just part of the modern computing world. Why on Earth would one see it as bad to have AmigaOS also included in this?

@Matt3k

MOS is far, far, far from behaving like a modern system.
Every time I see you say this I wonder if youve used a modern system.

Hell,.. youre restricted to 1.5GB RAM... an amount that was waaaay too low 10 years ago to the average computer user.
How about virtualisation? 3d rendering? (and dont say blender... it runs about 5% of the speed as it does a cheap pc and is very old). Running things like av1? Streaming? Concurrently running a 4k youtube video in the background while browsing with even just 20 tabs open in the background? Video editting? Code compiling? (yes, it *can* be done, but its literally less that %1 of the speed of doing it on a half decent modern system and you can run into RAM related issues too). How about playing a game on par with 15 year old games for mainstream systems? Surely a machine that can be used as a modern system has something better than games of 4 generation old consoles?

If a person has very, very basic needs then sure, MOS does a decent job, but to suggest its *anything* resembling a modern system if a person actually uses a computer in any sort of modern way is, quite frankly, somewhere between hilarious and naive.

@noXLar

I guess to each their own, and that is a good thing, using what you want at the end of the day .

I tried a bunch of options and just kept my 3000's for tinkering and fun. That is where my computer skills come in. Although, I will be honest the older I get the less I want to tinker. Some days I just want everything just to work and no put a day into a simple item. It does provide nostalgia for sure.

If MorphOS was adopted like it was suppose to be in the 90's, I would imagine they would be further along. But the user experience is really quite good and doing simple functions are simple. I like to use my 3000 as a tool after I set it up, now with my G5 I do the same. The difference is that it's 25 years later and my expectations have changed.

Many items just can't be ported to AOS4. From what I have read AOS4 hasn't developed their toolset for the OS much. MorphOS invested heavily in it over the years to make it possible to make modern apps. This further demonstrates why I think they should just move over to MorphOS.

Once in a while I do hit a spot where my computer skills are required and I reach out to the team and we can get it done.

If others see value with 4.x and it's trajectory, I guess we can all disagree on that.

To be fair, I will admit it did take me a bit to get the "home" feeling like you mentioned with it. Even the Apple hardware was annoying to me early on. But when I decided to use an Amiga flavor for my daily driver and work. 3.x/4.x/AROS/Others all couldn't do the job or I had to rabbithole to a non native apps for almost everything. I kept an open mind and tried them all for months. MorphOS was in a league of it's own here. I truly didn't care which one I choose but the truth is the truth and I embraced it.

Now that I'm used to Apple hardware and the OS, the brilliance of MorphOS team is evident. Using it day to day is a great experience with the OS and the appset being extremely good. That is unique to MorphOS, using in a modern day world.

Last edited by Matt3k on 16-Oct-2023 at 05:21 PM.

@fishy_fis

I certainly don't have basic needs and thanks for letting me clarify this:

Iris - Does email and renders perfect with vcard support. Have 5 different accounts and Iris works better than other email clients even on other platforms. Yes, calendar invite integration is not done yet, but should be in a few months. Go watch Dan Woods MorphOS video where he thinks Iris is the best email on any platform. He mostly uses modern Macs these days.

Wayfarer - Wayfarer does all the browsing I need for work, renders perfectly and runs well. Will the hardware have trouble with a real busy website? Sure, that will require my PC or NAS. The NAS is fully accessable via containers to I can run Libre Office just fine in a browser in MorphOS to look at a file. I haven't found one website it couldn't display, although speed may become an issue.

PolyOrganizer - I have used Swiftpage Act!, Salesforce, and Telemagic (some time ago) on windows and TimeTracker, OnTheBall, and Organizer on the Amiga. PolyOrganizer is better in many regards. The search function is much better than any of the other systems. The linking is better than say Swiftpage Acts! Company/Client construct. Since Act was designed around the client, company is an afterthought and has a lot less functionality if you work at that level. PolyOrga can build any relationship by linking and keeps the client detail at that level. Sure, Act! now uses Microsoft SQL and PO uses it's own native database and you will lose all those benefits. But for SMB, PolyOrga works fine. Now on to tasks, PolyOrga provides task you can link in relationships and meetings. If you want to do that in Act you really need to purchase 3rd party database items to get it done. It is bulky and has a much larger learning curve, where PolyOrga does all that most will need right out of the box, Fred wrote really good software and it's being updated (even the MorphOS team fixed code in it for me), this is a recurring theme with MorphOS that a lot of thought is put into something before they code it. PolyOrga is way faster than Act and much more productive.

Pagestream - even thought it was updated heavily, it still can be simplified... But it does all the content creation and generates PDF.

For all the stuff I mostly need to do, it works very fast and provides a fun experience. I have the NAS to use for other items.

Sure, I fully agree I'm not going to purchase a brand new game and even think about it. I don't pay attention to 4k in youtube, I just watch it and the 2.7 does a good job here. So can I purchase really high end media and play it? No, of course you would be right. My media hodgepodge all plays just fine though.

So I see you point on those items, and agree of course. But for my scope of modern uses, it really does a sweet job. Sure if you use a Macmini or Efika, your mileage will vary. But if you use say a 2.7 powermac, it will make a big difference.

Now lets dream a bit, and say they do port all this goodness to newer hardware. Now many of the other arguments will just drop away. The apps and tools are all there, they just need the hardware to run it on... Will it ever be on the same level as winblows for OS ability, of course not and that wasn't the reason for my adventure. MorphOS works very well, you really have to give it to them. The business uses I have are handled superior in ways identified above or work in a way you would expect modern apps to be. Jacacaps is a freaking coding animal with enormous talent and his contributions alone differentiate MorphOS from all the other flavors. I never feel slow or like I'm missing something as I do my daily work, it is fast and a joy.

If you think what I'm saying sounds as outlandish as your suggesting, than go ahead and try it for months and compare like I did.

Last edited by Matt3k on 17-Oct-2023 at 12:33 AM.
Last edited by Matt3k on 17-Oct-2023 at 12:30 AM.
Last edited by Matt3k on 16-Oct-2023 at 07:26 PM.
Last edited by Matt3k on 16-Oct-2023 at 04:35 PM.
Last edited by Matt3k on 16-Oct-2023 at 04:26 PM.

fishy_fis Quote:

MOS is far, far, far from behaving like a modern system. Every time I see you say this I wonder if youve used a modern system. Hell,.. youre restricted to 1.5GB RAM... an amount that was waaaay too low 10 years ago to the average computer user. How about virtualisation? 3d rendering? (and dont say blender... it runs about 5% of the speed as it does a cheap pc and is very old). Running things like av1? Streaming? Concurrently running a 4k youtube video in the background while browsing with even just 20 tabs open in the background? Video editting? Code compiling? (yes, it *can* be done, but its literally less that %1 of the speed of doing it on a half decent modern system and you can run into RAM related issues too). How about playing a game on par with 15 year old games for mainstream systems? Surely a machine that can be used as a modern system has something better than games of 4 generation old consoles?

MorphOS PPC handicaps may disappear over night if they introduce a MorphOS x86-64 port with 64 bit support and SMP which is easier to add when porting to a new platform. AmigaNOne hardware with AmigaOS 4 would then look like the performance of the original 68k Amiga hardware in comparison. I doubt AmigaNOne with AmigaOS 4 have anything in the product pipeline to counter with either.

Virtual memory isn't all that. Windows becomes so slow and unstable when it starts paging that I wish I could turn off virtual memory but that isn't safe in Windows. I still get out of memory errors and sometimes I can't read the requestor because there wasn't enough memory to display it. I wish there was a way to limit memory for a process/task like some embedded OSs do. Runaway processes/tasks eating up memory and then causing paging are a security vulnerability that can make the machine unresponsive for minutes and sometimes crash the system (Windows does not handle low memory situations gracefully).

fishy_fis Quote:

If a person has very, very basic needs then sure, MOS does a decent job, but to suggest its *anything* resembling a modern system if a person actually uses a computer in any sort of modern way is, quite frankly, somewhere between hilarious and naive.

MorphOS is ahead of AmigaOS 4 in the software department and they may be using vastly improved x86-64 hardware in the future, perhaps before AmigaOS 4 gets SMP. The MorphOS experience on a G5 Mac is similar to using a X5000. How modern it feels is relative. Even a 68060 Amiga feels modern at times because it is so responsive and programs load fast because they are small. Even fat PPC code can't make AmigaOS feel like Windows.

@pixie

PiStorm32 Lite and Emu68 use case is dependent on the active Amiga 1200 install base. PiStorm mini-ITX can expand this install base.

It's too bad Vampire standalone doesn't have C= A1200 expansion slot for Pistorm32 Lite-Emu68.

_________________
Ryzen 9 7900X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB
Amiga 1200 (Rev 1D1, KS 3.2, PiStorm32lite/RPi 4B 4GB/Emu68)
Amiga 500 (Rev 6A, KS 3.2, PiStorm/RPi 3a/Emu68)

Quote:

The problem I see with emulation is that ports will be inattractive as they suddenly compete with the native versions on these systems

That's right. There is no reason to waste time on 68k when all users use emulator.
Just switch to android and arm.

emulator is emulator and nothing more than emulator.
I prefer to use qemu instead of emu68.
qemu is 20 years old, made by profesionals, stable and faster than emu68.
On my rpi powerpc software emulated by qemu
is up to 40% faster than 68k software emulated by emu68.
On the same hardware.
So no reason to switch to emulated 68k.
Emulated ppc is simply better.

@MagicSN

Quote:

MagicSN wrote: @geen_naam The problem I see with emulation is that ports will be inattractive as they suddenly compete with the native versions on these systems. Of course the worst this is for games. But the problem exists for all kind of software. I have no problem with my programs running on Emulation (and am happy to hear if they work, maybe at one point might even test if they work - currently I have no Amiga emulation on my PC, I do not need it, as with the x1000 I have the real thing) but I have severe doubts that Emulation can carry an Amiga Software market (asides from OpenSource projects). You mention on "Classics and Vampires" as "for those who want the real thing". I have my severe doubts these things are fast enough for what I am doing at least. Though keeping an open mind for them. Currently the only thing which can run my ports is OS4 running on a real PowerPC based Amiga. It might work on QEmu, I do not know, I never tested it. But have thought about testing it for some while. Not because being interested in it, but because others might be interested in it. Still I doubt this is really a market. As a longterm perspective I only see three ways: 1) new PowerPC based Amigas on a performance level beyond x5000 or at a very attractive price point (or possibly that Notebook PPC Amiga project if it takes off) 2) a high power redone-68k Amiga - no, not Vampire. I said "high power". Which means at performance level of x5000 or at least x1000 and with a state-of-the-art 3D Chipset. And no, I do not mean PiStorm either. PiStorm still relies on really old hardware. 3) ARM Amiga (and I do not mean Emulation here). Basically an "ArmAmigaOne" (pronably a dumb name, in my native language the word ARM means "poor"). I personally do not really see a perspective beyond these three. Might even be 2 of them or all 3 at the same time.

AMD X570 and the latest X670E Southbridge chipsets still have legacy PCI to ISA bridge via LPC bridge. There are open-source projects that restore the retro ISA slot from LPC.

AMD IGPs and GPUs, and NVIDIA GPUs still have retro VBIOS. The "IBM compatible PC" is still the "PC" despite Intel's UEFI Class 3 / X86S initiative attempts.

PiStorm respects retro C= Amiga hardware legacy. C= Amiga legacy acts like the south bridge while the high-performance PiStorm-RPi is the northbridge side.

PiStorm-Emu68 works on new FPGA A500Plus ECS clones such as Minimig VER 1.97itx.
There's a market for retro Amiga.

Ideally, Amiga should include a classic Amiga-On-A-Chip southbridge (in place of the PC's legacy) and a high-performance ARM CPU with a high-performance PCIe 3D GPU.

I have tested SAM460's AmigaOS 4.1 FE on Qemu8 and I prefer WinUAE's classic AmigaOS 4.1 FE performance.

I could expose a spare PCIe GPU card to Qemu8, but the problem is my Gigabyte's RTX 4080 Gaming OC or ASUS TUF 4090 OC is about 3.5 slots wide which leaves very little room for a spare single-slot AMD RX card. High-performance hardware accelerated raytracing has a price.

My Ryzen 9 7900X CPU has RDNA 2 IGP, but AmigaOS 4.1 or System 54 doesn't have RDNA 2 driver support for Qemu8.

My old flagship AMD gaming GPU card is a 2.5-slot wide MSI Gaming X Trio Radeon R9 290X card which my office PC's GTX 1660 Super can beat. Most of my office PC GPUs are NVIDIA GeForce types e.g. GTX 1050 and passive GT 710.

My 2021-era gaming/raytracing box PC has RTX 3080 Ti.

"Enhancer 2.3, Virtio GPU driver for Qemu8 could be useful for PCs with NVIDIA GPUs since NVIDIA dominates the PC's discrete GPU market. Sorry, AMD GPUs are not winning.

Raspberry CM4 has PCIe support, hence PiStorm mini-ITX could combine low-cost FPGA C= Amiga retro chipset clone and PCIe 2.0 slot access.

Last edited by Hammer on 17-Oct-2023 at 06:08 AM.
Last edited by Hammer on 17-Oct-2023 at 06:03 AM.

_________________
Ryzen 9 7900X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB
Amiga 1200 (Rev 1D1, KS 3.2, PiStorm32lite/RPi 4B 4GB/Emu68)
Amiga 500 (Rev 6A, KS 3.2, PiStorm/RPi 3a/Emu68)

@ppcamiga1

Quote:

ppcamiga1 wrote: emulator is emulator and nothing more than emulator. I prefer to use qemu instead of emu68. qemu is 20 years old, made by profesionals, stable and faster than emu68. On my rpi powerpc software emulated by qemu is up to 40% faster than 68k software emulated by emu68. On the same hardware. So no reason to switch to emulated 68k. Emulated ppc is simply better.

Post your benchmarks.

A1200's PiStorm32-PRi 4B-Emu68 (running Devilution X 68K) rivals SAM460's Devilution X PPC build performance.

The main reason for PiStorm32-PRi 4B-Emu68 with A1200 is running WHDload Amiga 68K games and 68K apps (e.g. Deluxe Music 2) with real C= AGA hardware while offering SAM460-like NG performance.

Emu68 supports multiple CPU cores when there's software for it.

Deluxe Music 2 is not working on AmigaOS 4.1 FE.

Last edited by Hammer on 17-Oct-2023 at 06:27 AM.
Last edited by Hammer on 17-Oct-2023 at 06:22 AM.
Last edited by Hammer on 17-Oct-2023 at 06:21 AM.
Last edited by Hammer on 17-Oct-2023 at 06:19 AM.

_________________
Ryzen 9 7900X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB
Amiga 1200 (Rev 1D1, KS 3.2, PiStorm32lite/RPi 4B 4GB/Emu68)
Amiga 500 (Rev 6A, KS 3.2, PiStorm/RPi 3a/Emu68)

@ppcamiga1

Quote:

That's right. There is no reason to waste time on 68k when all users use emulator. Just switch to android and arm.

As for Android and ARM, Genshin Impact 4K at max graphics settings runs best on a gaming PC.

_________________
Ryzen 9 7900X, DDR5-6000 64 GB RAM, GeForce RTX 4080 16 GB
Amiga 1200 (Rev 1D1, KS 3.2, PiStorm32lite/RPi 4B 4GB/Emu68)
Amiga 500 (Rev 6A, KS 3.2, PiStorm/RPi 3a/Emu68)

@ppcamiga1

Quote:

I prefer to use qemu instead of emu68. qemu is 20 years old, made by profesionals, stable and faster than emu68. On my rpi powerpc software emulated by qemu is up to 40% faster than 68k software emulated by emu68. On the same hardware. So no reason to switch to emulated 68k. Emulated ppc is simply better.

https://youtu.be/--8dhqNf1nY?si=dVfAViAc8sX6Ikpt
This was the state of QEMU 3 years ago, done by professionals, you should be ashamed of yourself.

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