Toni, do you have a roadmap to share?

[ilm]

maybe we can start using the sourceforge's bugzilla
so
- we have a general idea about current bugs, feature request in one place
- people outside abime can submit bugs (wider audience)
only finding someone willing to maintain it ...

for starters we can make sticky with a link to it in winuaeprob or smth

[Ultron]

Originally Posted by DamienD

Sorry Ultron, I agree with everything Toni has said because for me the drivers were an issue once

One of my laptops is below the above specification and everything runs just fine and dandy

I would start by making sure you have the correct configuration in WinUAE setup (try Quickstart first) and if that doesn't work then take a look at your hardware drivers...

There's no need to be sorry, i was fully aware even as i posted that, that, issuing criticism, particularly on home grounds, is never easy, that's why so few do it.

I already knew that basically a flurry of "It can't possibly be WinUAE!" would ensue. Of course not, it's the drivers, or the hardware, maybe the config's off, or maybe i just have chubby fingers!

If it were so, then the same chubby fingers or bad drivers would affect, WinFellow, MAME, MAME32, the myriads of emulators fo NES, SNES, GBA, Sega, Model2, NeoGeo, AtariST, and pretty much everything, from modern games to Apache, through MySQL, i use. It doesn't.

I know particularly in a board such as this where most people are Amiga and only Amiga, and rarely try other Emulators, that that's the impression people get: "It's not that bad! It's actually close to perfect!". To be fully honest i already expected the reaction - It's not novel, it isn't the first time it happens either, nor will be the last.

WinUAE was pretty good for what it set out to be. Portable. Feature rich. It is not fast nor efficient. There is a lot of bloat in there. Most of it carried by so many years and different programmers it is, by now, preferable a re-write (new Emulator) than insisting on an Emulator that's a monolith of labirynthine code by now.

WinUAE would gain immensely by getting a programmer who knows the ins and outs of the intensely arcane Amiga UMA\Chipset architechture, someone who knew Windows Multimedia API's left and right - and yes, this would probably end up in a Windows only version, but heck it's what we already have now without the performance benefits, and \ or someone who optimized, removed all the bloat, out of WinUAE's code. That person(s) hasn't appeared, and if it has, it probably wouldn't contribute under current conditions (or lack of there of).

Believe me, it'd help. If someone here wants to download the WinUAE's sources, inspect them, and then tell me that they're the most amazing, optimized, non-bloated efficient C++ code they've ever seen, I'd love hearing from them.

My opinion - and it's just that, it's not really that important - they (sources) aren't. They could be better. A lot better in some cases. A more open development structure would, hopefully, allow it to happen. As it stands, it won't. It hasn't, it isn't, and probably for the aforimented reasons: Even so much as browsing around the sources is a byzantine experience.

Now, we could, conversely, say that WinUAE is perfection on planet earth, and that if it runs fine on a 1.5ghz or a 3.0hz, if you just don't move the chipset slider from 6 to 3, it's more than enough, and move it along, be content, and no upgrades needed.

It is fine by me as well - I don't agree with it - but it is fine by me as well

[ceztko]

Originally Posted by Toni Wilen

Simple. There is no roadmap...

Well Toni, no matter. There are hundreds of died projects with kilometric roadmaps, so it's not a problem if Winuae lacks one

Originally Posted by Toni Wilen

...I have some things I planned to do 2 years ago but some months later it felt not worth the trouble, but I still plan to do it someday...

this is a point where i'd insist: at least a *real* (not the one in winuae.net) know-issue/bugs list would be useful. Because i don't find the source to be annotated with FIXMEs/TODOs somewhere, there are 2 possibilities:
- you profically use your brain memory to store them for later use (even years after);
- you took notes of them during development in personal documents.

Release at least some of these personal development infos is surely useful, because it can transparently tracs where WinUAE is currently lacking. It's not impossible that in this way more people could help in areas of their interest.
Even more, it has the beneficial effect of reducing amount of duplicated bug reports.

Thank you,
Francesco

[jotd]

To Ultron:

I think you're being insulting to Toni's work. WinUAE features rock. JIT rocks. So why don't you stop being an annoyance and step in to optimize WinUAE yourself?

Oh, I'm sorry, you don't know a thing about C++? It's sooo easy to say "I'm sure this is under-optimized and if someone tries very hard he can make it better". I think Toni is a very talented coder, and WinUAE flies compared to my 68060/050 amiga, even on a poor PIII 1.1GHz.

Prove us wrong. Get a sub-contractor who can do better than that. Please keep us posted about the progress.

[DamienD]

Originally Posted by Ultron

I know particularly in a board such as this where most people are Amiga and only Amiga, and rarely try other Emulators, that that's the impression people get:...

A bit off topic but... Hmmm, that's a bit of a funny statement. In my 8 joyful years of emulation I've used loads and loads of different emulators:

-------------------------

Amiga:
WinFellow, WinUAE.

Arcade:
ACE, Calice / Calice 95, Daphne, Final Burn, Impact & Modeler & Zinc, Kawaks, Kiame & U64Emu, MAME, Nebula, NeoRAGE & NeoRAGEX, R.A.G.E, Raine, Sledgehammer, Sparcade, Viva Nonno.

Consoles & Handhelds:
... (too many to name so I'll just list the system)
Dreamcast, FM Towns, Game & Watch, Neo-Geo Pocket, NES, Nintendo64, PC Engine, PC-88, PC-98, PlayStation, Sega MegaDrive, Sega CD, Sega Saturn, Super NES, X68000.

PC:
DosBOX, ScummVM.

-------------------------

In fact I'm playing the Resident Evil series through ePSXe at the moment

I also know many other members here who use a variety of emulators...

Anyway, back on topic... In regards to WinUAE I think this is an absolutely awesome emulator. I first discovered this back in 2001, the first version I used was WinUAE v0.8.17 R3. At that stage it was pretty good and I set about configuring / testing about 250 games...

I then lost interest for a while due to "real" life and pick it up again seriously when WinUAE v1.0.0 came out. Wow, I discovered many features / improvements since WinUAE v0.8.17 R3 and set about recreating / testing 450+ games. I was amazed at the number of games I could now run etc.

For me WinUAE has come along in leaps and bounds since I first discovered it. All thanks to Toni

Yeah sure, you may say it's not perfect (of course that's your opinion) but I think that it would be very, very difficult and take many years of excruciating time / knowledge / dedication to get anything even close to WinUAE in it's current state...

[Retro-Nerd]

Actually there is no other Amiga emulator available besides WinUAE and the ports. WinFellow doesn't count for me, it seems pretty dead and can't offer the compatibility and features of WinUAE. So this speaks for Toni's work itself. Nobody can program a better emulator at the moment or don't want to match with Toni. WinUAE is very compatible and replaces an real Amiga 500, what more could you expect from it?

[Marcuz]

i chime in just to say that while i have asked sometimes features and modifications that toni denied during the last years, still Winuae is a great piece of software. it DOES run well on my old AMD @1Ghz. i have no one complain about the code. it's not perfect, but what's perfection? i think that when one claims for a program that meets his own standards, he does not asks for perfection, but for a program tailored on himself, and that's too much to ask.
maybe the only one critique one can move is that Winuae is a constantly reworked program, and not really a defined project from the start as in toni words. this has made its interface a little patchwork (to me, again).
but that is also a gift, as this fact has made possible for us all to have the program that runs an Amiga on our PCs since a while now, and well. do you think it would have been possible to prompt a staff of people to start from the scratch a long term project, volountary based and not usable for years before the completition and for this reason, detached from a community of users and potential users, for emulate a machine dead since roughly ten years?

[Mangar]

I also, and pardon me the selfishness, think WinUAE should focus more on perfomance and less in portability. I, again, pardon me the selfishness, couldn't give a rats ass if UAE doesn't run on Solaris or ReactOS. I'd rather, at this point, have a fully perfomant, f a s t, emulator on Windows, than an Emulator that stumbles away on every platform out there.

WinUAE is a pretty good emulator for the conditions it has to be done in, but perfect, it ain't. Far from it, unfortunately.[/QUOTE]

Mame is overbloated and not programmed for performance. Any good emulator is programmed for compatibility or else it's junk.

[Doc Mindie]

Ultron, I re-state that the Amiga is likely to be the most advanced and complex computer to have ever been released for consumer market.

AtariST? when compared, it was piece of hastily bundled together hardware. Mac? nothing special there, an Amiga1200/4000 equipped with 68060 emulated a 68k based Mac FASTER than any real 68k based Mac could run.

My A1200 when I had a 68040/40 card, emulated a 486 running at around 20MHz........(with PCTaskTurbo). It took WinUAE to goto Tony and his JITcompiler before the emulated A500 even approached usable speed on P2-350. WinFellow? a tad bit faster back then, but still not usable on the same P2 CPU. "Not Usable" as in it wouldn't run the same amount of games as WinUAE.

WinFellow, back in 1999 or there abouts, outshone WinUAE in terms of speed, but was severaly lacking in compatability. And still is. Whereas WinUAE has proved to be more popular, because of the much higher compatibility level of games.

And you STILL think WinUAE is too bloated? take away the code, and I'm pretty sure that games will start to fail again...... So you have the option, keep WinUAE "bloated" or loose compatibility.

You could try to make an Amiga Emulator yourself, since you seem to think a "from scratch" approach would be the best idea...... And I wish you, from the bottom of my heart, best of luck. (And I'm actually sincere, not sarcastic, though I understand it can seem that way.)

[ceztko]

Originally Posted by Ultron

I also, and pardon me the selfishness, think WinUAE should focus more on perfomance and less in portability.

As Toni told us a lot of time, he already doens't care much about "portability". As you can verify by your own, as today WinUAE is not cross-platform, in the sense that you can't simply ./configure, make, make install and run in every *nix. Winuae is portable because of the language it was written (standard C) and because of the nature of the sw: it's an emulator, so it must be precise and predicible in its behavior. In these situations, endianess, size of INTs and so on are very important factors. If taking care about them for reaching a more accurate emulation gives you automatically a more portable source code, how can you wonder?

Really, you should state that WinUAE and E-UAE are different things. Richard Dummond takes core functionality of WinUAE and maintain the gfx,audio,I/O backends and the UI and deploy a portable package that runs in various platform. This is, for some aspects, unfortunate, because if WinUAE and E-UAE would fully share a source tree (not a so difficult task), Richard and other developers (like the ones of uade) would easily focus on improving the core functionality (from that everyone could benefit), rather than spending time in merging the bits of the core emulation that are platform dependant.

If you think that maintaining portability or cross-platform capability slown down the development rate of the sw, you are simply wrong. See scummvm or dosbox, for examples. Winuae could easily be made a cross platform app, but for the moment Toni doesn't believe that the effort is worth the pain (and this certainly would require some sort of cooperation between Richard and Toni).

[Retro-Nerd]

I also, and pardon me the selfishness, think WinUAE should focus more on perfomance and less in portability

Damn, i can't read this silly whining for more performance anymore. Nobody should use a ten year old PC to emulate an amiga with WinUAE. It is as Toni said, you can play 95 % of all Amiga 500 games on a 1,5 Ghz PC without any errors. This is the roadmap, not a "talking bullshit", thread.

[Marcuz]

Originally Posted by Retro-Nerd

Damn, i can't read this silly whining for more performance anymore. Nobody should use a ten year old PC to emulate an amiga with WinUAE. It is as Toni said, you can play 95 % of all Amiga 500 games on a 1,5 Ghz PC without any errors. This is the roadmap, not a "talking bullshit", thread.

actually one of the reasons for a public roadmap is for people to say his opinion and ask things. that don't qualify as silly whining if it is formulated properly. also, while one should respect requirements stated by the producers of software for the hardware before to complain that the programs are too heavy for his machine, it is also true that Winuae IS heavy, and one is entitled to say about it, i think.
i think Winuae COULD be better, yeah, take a big team of producers who knows all about it, give them money and start from scratch for portability and performance. but that's unlikely. instead you have the dedicated work of toni wilen, who, for the simple reason of being here available for people to talk about his project and betatesting it as he does, he IS a cool programmer.
that doesn't make Winuae perfect, and i don't think Mangar is whining for sayin' so. however i prefer Winuae to be out there as it is than it not being at all. and for that i thank Toni.

[Toni Wilen]

Random "answers" here..

Where did I say (Win)UAE is perfect or the source is great?

It is far from perfect (but we all have different opinions about the meaning of "perfect")

Sources are a mess (and there is no C++ anywhere, btw, it is 100% plain C, not including JIT which is complete mess of C and assembly )

Part of the messiness is optimization, Bernd did great job of optimizing display emulation and CPU emulation but it also makes it very hard to understand. You are clearly not a programmer if you think you can write code that is both optimized and nice looking. (at least not in C)

You are also wrong if you think Windows multimedia APIs are not used or they can give huge speedup in non-3D emulation. You can't accelerate 2D, especially Amiga where you have to have 100% exact timing, you simply have to emulate whole hardware in software. Only the very final step in display rendering (scaling, filtering etc.) can be accelerated (and UAE has a shortcut in non-filter mode there already, writes go directly to display RAM)

There are 3 quite clear steps in display emulation, (first step: emulate single hardware line to internal buffer, this includes basically complete DMA sequency emulation of single line, second step: convert the line to chunky format + merge sprites, handle dualplayfield whatever.., step 3: move data to display RAM. step2+3 are done simultanously when not using filters)

There is nothing that Windows "multimedia" features can help in display emulation. Original display data is in completely wrong format (and in totally incompatible when using HAM or EHB or DP or sprites), Windows can't process it. Learn to program emulators before asking for impossible features.

..

Because I do this just for fun, I don't really care about features I don't personally use, good example is filters, they are surely not the best and I only added them because I got too many request. Any coder should be able to optimize them if they are interested enough (I am not, they work fine enough for me) But I will do try to fix them if I get GOOD BUG REPORTS. (I haven't)

..

Now, we could, conversely, say that WinUAE is perfection on planet earth, and that if it runs fine on a 1.5ghz or a 3.0hz, if you just don't move the chipset slider from 6 to 3, it's more than enough, and move it along, be content, and no upgrades needed.

Why are you playing with that slider? It is compatibility hack only, in normal use you should not enable it. And of course emulation will slow down if you FORCE it to use CPU<>chipset allocation that never can happen in real hardware.

....

I guess I should just delete WinUAE's sources and wait until that great person appears who rewrites perfect emulator from scratch in 1 week

[ceztko]

Originally Posted by Toni Wilen

Random "answers" here..

Can you give me some random answers too?

http://eab.abime.net/showpost.php?p=264242&postcount=23

Thanks,
ceztko

[jabsy]

Originally Posted by Retro-Nerd

Nobody should use a ten year old PC to emulate an amiga with WinUAE.

Exactly. You should use a ten year old Amiga instead, if compatibility is your thing...