Virtio GPU, QEMU, KVM, etc.

@All
Some of you take it too pessimistic. Were there any issues when WinUAE got PPC support and ability to run AmigaOS4 on it ? No way! It still faaar to be usable for real. With QEMU it will take years even with new driver for GFX card to sort all the issues which not present on real hardware.

But ok, let it be no issues, and after a year or two everything will be good and same as fast as on x5000. Will it be a problem for anyone ? Of course not! Because there is no new hardware , and after this Tabor fail there will be probably none from AEON. And that no issues as well, it all will take long time till everything dead , and we will (and still) have enough time to enjoy what we enjoy.

Next thing, is about whole AmiWest fiasco: There wasn't anyone who can says about OS4 progress, because there weren't persons involved in. Matthew is against anything Hyperion, and he has no access to the progress of Hyperion to know about. Steven and he's facepalm presentation about nothing, seems just to be a cause that he also can't talk about progress on OS4 from Hyperion's side by any of reasons (I don't know why). And don't ask me why he's presentation were called with words "OS4 updates", he says nothing about, while there are OS4 updates. Just there also wasn't any real OS4 developer involved to talk about something interesting of late progress. Damn, even no one were able to make some “OS4 Gaming” presentation, it has no needs to have anything internal and new to know, to make some tasty presentations showing some of OS4 latest stuff of year, just there wasn't anyone willing to do so.

The only _REAL presentation which worth of watching were one from geennaam : quality, professional, technically interesting and co. And this wasn't requiring him to be os4developer of OS itself, or beta tester. It's all about ppls involved in AmiWest, which just .. dunno .. Don't attract anymore interesting developers and users who want to make something more interesting than those strange "hihi" talks instead of presentation, and walking boing balls over the rooms.


And last thing about OS4 progress itself : yes, whole progress everywhere slowing down. That is true. But is it stopped completely ? Of course not ! For last 2 year there were a lot of things improved and fixed in the OS4 side, that include kernel, kickstart modules , libraries, filesystems, commands, USB, console fixes, NGFS, etc, etc. Matthias right today send me new code which use latest (few days ago) beta kernel where were added profiling support for all platforms which do not have it (x5000, x1000, tabor), but now also pure GNU's gprof can use it ). We also do some good work on the bringing new binutils/etc to us, which mean that new SDK will see day of light sooner or later. And the OS4 update i am sure will happen sooner or later, when things will be ready for. At least now, and new kernel, and new elf.library are need to be released pretty “soon”.

So, it's just AmiWest were mostly crap this year, not the whole situations. There are progress happens, at least from developers working on Hyperion's part. Maybe not that big as we may hope for, but still. Just on AmiWest there wasn't anyone talking about. The only intersting presentation were geennaam's one, and it was only in DevCon days.

@kas1e

I completely agree with what you say.

Sorry, I am not.

The "ordinary user" has not seen any kind of OS4 update for...ages!
It is fine for you, if you can play with updated components but what we need are official updates for all!

That aside completely, that there is not any kind of information or public relations from Hyperion side. This is sad.

@kas1e
+1

@Primax

That’s latterly shame, hope any relationships between developers and Ben can be restored, then I hoper fore better cooperation between different companies, and less legal nonsense.

Most likely hoping for too much.

@LiveForIt

Mainly we do not know, what plans Ben has. And, if he has any future-oriented, positive OS4 plans at all...

Regarding emulation: I think it is a good starting point for any user or developer who just want to get a first impression of AmigaOS 4 in an non-expansive and, this is important, easy way.

Although I am not quite convinced that this "new branch" will give us a big glut of new developers, since: this has been already possible using WinUAE and, yes, also QEMU.

Better colour depths and screen resolutions may be important for gamers. But for developers?

I think first of all, there has to be Hyperion's commitment to further development, or at least any honest assessment of the situation and a (possible) preview to the coming development.

@Primax

+1

Being outside the beta bubble is like feeling the whole platform is already abandoned (even reading the forums is depressing nowadays)

True, there were some update packs (small ones) and third party stuff (enhancer), but neither one really brought impact (except the new gfx drivers)

What about ngfs?
It's already in beta since years now...why not release it, why no news, why no reasons for the delay, why is everything still kept behind closed curtains?

Just talk to your customers already...

edit:
I'm an x1000 user and abandoning (more or less) that hw generation in favour of the x5000 *does* help a lot
/sarcasm off

@Skateman
Quote:
The topic itself isn't depressing, but the conversation has certainly headed that way. It is good to be real about the current situation, but I hope that people will be solutions focused.


@kas1e
Quote:
That's a shame. I didn't present anything this time, because it's been a tough year and I don't have anything ready. Interest in ZitaSync and ZitaFS seems to be weak anyway, so I expect most won't even notice. Some of the work I'm doing for A-EON is taking much longer than anticipated, so it wasn't ready to show.

I didn't know that geennaam presented something at the devcon. I'll try to check that out later.


@LiveForIt
Quote:
IMHO, the best thing Ben could do for AmigaOS 4 would be to sell Hyperion to someone with more time and resources to devote to it (and a decent plan). His day job takes up too much of his time and attention, and yes, the burnt bridges are an obstacle.


@Primax
Quote:
I tried using WinUAE for development and quickly abandoned. It was slow and buggy (as in crashing a lot). Hopefully it's become more stable since then.

Developers working on software that needs higher resolutions and 3D drivers are going to need a system that supports those features. It's very hard to develop software when you don't have the ability to test things directly.

Plus, even developers may want to try out everything the platform has to offer, and emulation still has some gaps (e.g., 3D support).

Hans

@Primax

Quote:

he does, he just does not have lot’s of developers who want to work for him. Developers tend to work for people who pay.

@Primax
You need to take for granted that Ben have no plans. Or, just a plans, which will never be reality. Just those who accept that, can enjoy things more and focus on what they can do instead of waiting.

The lacking of proper promote and news coverting always were the bad side of the things, that true. But this is not different today.

And, also, about "to play with something which ordinary users can't" : as i say, we play with gprof and stuff now, which is open sourced. Same as binutils. Same as SDL2, same as other kind of code. No one is forced to wait anything to be released in terms of CORe for OS4 now, take it like there will be no updates, and do what you can with what you have to have your interest be in :)

@Hans
Quote:

You will like this one for sure, just it recorded in DevCon days.

@All
And btw, you forget one detail about emulation : if emulation of OS4 will be on the x5000 level (without new bugs, without speed issues, without limitations , which will take years surely), you will be able to DEBUG for real your apps while developing, making it easier to create software for os4. So, for developers, it surely will be of big help, same as of big help is cross-compiler allowing us to have the lot of things which if no cross-compiler we didn't.

Quote:

This 100%. While I was developing I pretty much had to wait for my X5K to arrive. There was simply no way to test without the 3D capable GPU. I was able to do "does it run at all" tests with QEMU and GPU passthrough emulation but it ran at a few fps.

The virtio drivers will be a godsend for development.

Will Qemu and virtio bring in a lot of fresh new users? Debatable, I guess the question really is how many OS4 classic licenses were sold because I would think that most of those users can't be happy with it and may want an upgrade. Obviously depends what they thought they were going to use it for before they found out it was slow.

OS4 is no more dead than I am!

@MartinW

Quote:
I don't have direct numbers, but there are some hints on UAE users in two surveys I did:
- 2019 graphics survey. There were over 100 people using OS4 classic via WinUAE in 2018. That dropped by roughly half in 2019
- OS4 hardware vs emulation survey - 175 people said they use/tried UAE. 56 said they use/tried QEMU, with an overlap of 34 who have used/tried both (197 emulation users in total)

One interesting question would be: what would entice more 68K AmigaOS users to upgrade to OS4? And how many might be interested? I get the impression that many are interested in the classic Amiga as a retro machine. The classic chipset is a must to them. Likewise, Vampire users are living their alternative future with an upgraded custom Amiga chipset. There should be some classic users who would be interested in a "nouveu-retro" AmigaOS experience if there were a compelling enough offer.

Hans

@Hans

KVM and Virtio is only for Linux users.
Use It will mean have a good skill in linux.
By the way will be a good thing if we think about cross developing something inside VMs with multiple istances on a single machine.
For sure it will help developers.
In future i can see a good official distro on usb amiga brand stick with AmigaOS4 with all sdk in sell by Hyperion resellers.

@Hans

Quote:

Yes, users, just wants things to work, they are not interested in the technobabble.

I am feeling of the developers missed it and assumed, that once we had good software on AmigaOS4, everyone will adapt it and update their software, well was already too late for that 2005.

the community instead doubled down on hardware upgrades, its not just chipset, its also being authentic, correct for its time-period, and so on, so you never appeal to everyone. Even if stuck some CIAA/B chips in the AmigaONE’s, and added joystick ports.

its bit odd, when think about it, Sam4x0’s and X1000, and X5000 all have gpio ports, can have been mapped to joystick ports. Its actually bit odd no one made addon for that yet. Perhaps GPIO’s can also hosted CIAA/B chips, they are quite simple 8bit circuits.

Now you also need cooper and complete aga to run 68K software native, I think going full out classic won’t make sense because have EUAE so perhaps we don’t need it, we just need to make EUAE experience better, lots of small things to fix, network support, file protection bits.

but at the end its way more expensive then a classic system is, getting price down to a price people are comfortable is also important.

X5000, X1000 is more similar to developers’ system, the gamers might be happy with simple console.

but its not just gamers, there people who are interested in music as well, sadly a lot was promised and not delivered in that apartment as well.

@Hans

Well, I was one of those users that bought OS4 classic when it came out. Used it for a bit and soon realised that without the graphics, I couldn't explore demos, or games, or do anything much that I was interested in that I couldn't do on my physical classic machines (not a huge amount of point in running demos in emulation for me when I can run them natively on the machine next to me).

Had a virtio driver been available (and of course, Qemu - that wasn't OS4 capable back then either) that made the experience include some games, OS4 demos, etc., etc., then I'd have probably stuck with it. That wouldn't have been the end of it mind you, the volume of stuff like demos (as an example) just isn't there compared to classic, but I certainly wouldn't have thought "hmm, that's a bit crap" and shelved it.

@tlosm

Quote:
No. I have tested software running in a virtual machine under Windows. Likewise, I have tested my VirtioGPU.chip driver on Windows too.

Unfortunately, VirGL isn't available on Windows, so that is currently Linux only, and possibly MacOS X.


@LiveForIt

I'm not sure what you're suggesting, but I think reintroducing the classic chipset to newer AmigaOS 4 hardware would be a bad idea. It won't entice hardcore classic Amiga enthuisiasts for whom the Amiga is classic chipset + 68K. They're just not interested. It would actually be easier to move the OS forward if we could finally remove all the old legacy chipset code.

Anyone interested in a nouveau-retro Amiga is unlikely to care too much about copper lists and old joystick ports. More seamless integration with UAE for legacy hardware banging stuff may well have its place. Ultimately, though, I expect those interested in a modern(ish) Amiga to want more modern computer capabilities.

The big question is still: what would it take to get them onboard?

Quote:
Agreed. The price of admission is still too high.

Hans

@noXLar

This sounds like Amithlon 2023 AmigaOS 4.1 version. Could be a nice development for the very development slow OS 4.1 market.

(been ages since I visited and dropped by on this site!)

But any way I think too many think Emulation is the end of a platform, or the end of new hardware. I pointed this out on Amigaworld, but likely the two most successful Amiga hardware project in the last few years have been Emulation base.

A500 Mini and Pistorm.

I also heard they're maybe making a FPGA Pistorm, which sounds interesting.

I keep thinking what if instead of Emu68 the Pi emulated a PPC board, would a PistormPPC be possible?

If it was, Classic AmigaOs4 might have a second shot in the arm for development. an easy low cost upgrade to PPC for classic Amiga would be good to see.

Then also the above FPGA Pistorm would be even more interesting if the software was well developed so you could switch from a excellent emulated classic Amiga to AmigaOS4.1 in QEMU with ease, that would be really interesting hardware.

So I wish all the devs and work going into this best of luck, I know we would all like more new native hardware and faster development of AmigaOS4, but I think just acknowledge the limited resource the Amiga world has (which is not fault of anyone) is not a bad thing, it just what can be done to be most effective to limit work.

I mean another thing I love is Rabbit hole and how well it works on Amikit setup, on the Pi400, it kinda given that dream of kinda running a modern Amiga with access to modern thing like a web browser, to just double click Chrome icon on AmigaOs and have the browser load up and seem like it part of the Os is pretty good. These solutions I feel are clever ways around the problems. We're never going to be main stream without a crazy billionaire coming along (elon musk), but I do love people have developed way around the problems.