I've been lurking on this thread for some time now and it's high time I make a contribution to it:

I spent eight years of my life trying to develop a way to recompile AmosPro software to AmigaOS 4, AROS, and MorphOS in a project codenamed Mattathias.

There were two things we needed: processor independence and chipset equivalence. I'm confident that the first part can be achieved using a statically compiled virtual machine based on modern technolgies like LLVM. Sadly however, the second part can only be accomplished via emulation whether it be hardware or software.

There is almost no equivalence between Amiga chipsets and modern graphics cores. OpenGL is almost no help whatsoever when it comes to Amiga chipset emulation. I know that EUAE release 4 had some graphics acceleration help via OpenGL in much the same way as WinUAE uses DirectX to speed up its emulation marginally but that is the extent of it. 3D APIs are ill-suited to doing what the Amiga did in the same way as Amiga chipsets are ill-suited for texture-mapped 3D like Quake.

While I'm confident that a VM solution for AROS can be concocted out of LLVM or GCC or some other related technology, the old Amiga software will only work on the Classics anyway. This brings up the greatest weakness in all NG Amiga OSs. In all the years that AROS has been under development, only a handful of programming languages have been developed for it and chief among them was C.

Sure there's an Object Pascal implementation in the works for AROS x86. Sure OS 4 has a Modula 2 compiler. AmiLua, being written in C, has been ported hither and yon. Hollywood was written with 24-bit and 32-bit truecolor chipsets in mind and runs on anything with a graphics card.

Even with higher-level implementation of chipset functionality, I think the closest to AmosPro we'd ever hope to find on non-68k AROS versions (or OS 4 or MorphOS, for that matter) would be a BASIC based on the AmiDark engine, probably compatible with DarkBasicPro.

Ultimately, if you look in the Aminet for loads of programming languages the vast majority of them are still 68k and ECS or AGA dependent. ChrisH discovered that his Shadow of the Blitz demo port which he used to test his PortablE runtime couldn't run on a Classic due to radically different architectures of AGA vs RTG.

Therein lies the rub. Old Amiga software can only be emulated on modern hardware and the best hardware to emulate with is an FPGA. Sure you could use a quad-core ARM to do it with, but until there's a quad-threaded version of UAE it's guaranteed to suck the living daylights out of your processor.

Now that I've given up on trying to get Mattathias to run on my MicroA1-c, I've shifted my priorities a bit to helping get the bugs worked out of the now open-source AmosPro and writing graphics drivers and libraries for the common special effects that the Amiga chipsets enabled, along with graphics card equivalent libraries for those same special effects. I also expect to update AROS 68k to be able to do the things that were never properly finished or debugged on AmigaOS 3.1 such as proper Copper-list merging and child-screens.

@Samurai_Crow

I remember Qbasic and TurboBasic, this was the days
also remember I made game in TurboPascal 7, and lots of crap in Amos.

Amos gave people who was not experienced a way to make games and programs, it also stole a lot of time that I might used to learn programing C, or some thing more useful.

Maybe its best forgotten.

Shure its possible to make a AMOS clone, but does not make sense to do it like it was back then, yuu be just thrawing away too many CPU cycles on P2C routines and other silly things. And what about modern features like alpha blending and things like that, no it does not make sense.

I wanted to do the same thing some time ago, it was before AmigaOS4.x, I was trying to make Amos clone for Linux, this was because had made a program I called “Adventure Editor”, it was major program to make adventure games.

I think for anyone that want to learn to program, that does not want to reboot there Amiga for every mistake, there is web development Javascript, and there is Hollywood, personally I don't think that are that many people being interested in that kind of thing now.

AmigaOS it self provided a few legacy draw effects by using graphic library it uses the old-fashioned pen system, so it does not matter if its 32bit or 4 colors, it certainly possible to make programs work from every thing from Amiga 500 to AmigaONE-X1000, if you want to, I'm sure there are some pitfalls.

Actually run "Last Hope" written for AmigaOS 1.2 on AmigaONE-XE / AmigaOS4.0 (no UAE), while porting Catwessel driver from Linux to AmigaOS, I used "Last Hope" for testing .

Last edited by NutsAboutAmiga on 24-Jul-2014 at 06:05 PM.
Last edited by NutsAboutAmiga on 24-Jul-2014 at 06:02 PM.
Last edited by NutsAboutAmiga on 24-Jul-2014 at 05:59 PM.

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

@BigGun

Quote:

What number is needed to give the AMIGA a push again?

David Pleasance, Managing Director of Commodore UK -

Once quoted 50,000 units as the magic number for a small market.

He did amongst other things negotiate with manufacturers and source components and managing lead times for smooth production of Amiga systems.

https://www.facebook.com/david.j.pleasance

When Commodore went bankrupt I said to myself 'Why on earth for ? here in UK we are selling Amigas non stop'.

The Advantage Commodore had was that they also had their own component stores/ warehouses so they knew the component market pretty well and enjoyed good prices.

Commodore did not just 'follow' standards - they ignored them if it increased the cost of production, the PC industry has a monopoly on technology standards - they dictate to everyone what technology one should be using - and its to their advantage as they have already sourced the parts.

We should be saying What is cost for USB2 or USB 3 use cheapest parts, when we need in 2 years time.

Do we really need 10 layer board, what dont we need. If I get this IC increase or decrease layers needed etc.

_________________
Download 499.26 Mbps, 659.94 Mbps Upload :)

@KimmoK

You misunderstood my point. I didn't say that there wasn't some Amiga people still using an Amiga as their main computing device. I said that nobody would be using a $50 68K FPGA machine as their main computing device.

The idea wouldn't be to replace whatever PPC systems people are using now. The idea would be to reach a mass market, reach old Amiga people who don't frequent this forum, and all of the retro gamers out there.

The system would be used for mostly retro gaming and running old software. AROS would be the most modern OS on the thing, so I guess some regular computing could be done, just not enough to be used in lieu of better solutions.

10K-100K+ sales should be doable, if the price is right and proper marketing is used. And at that level it may be possible to attract homebrew/indie developers to the platform.

@Nameless
This small baby is the answer:

www.youtube.com/watch?v=MSi_9Lv4xRs

_________________
AROS Broadway - AEROS - Aminux - AmiCloud - indieGO! Appstore - AmiWallet - VAN lossless video codec - AMC Amiga media Center -KrypUnite - LibertyNet - MinX - amigaNX

@phoenixkonsole

What sort of speed is 68K stuff getting on that?

ARM is fine, but the lack of 68K JIT on ARM tends to be a problem.

Last edited by Nameless on 24-Jul-2014 at 09:43 PM.

@Nameless
I need to test it. The model shown had only a single core and the final version has a dualcore and double ram.
Give me some days to do benchmarks.
As told before there is a ppc jit available. Theoretically you could run the jit enabled ppc build of uae on it.

A amiga500 is no problem since i played through superfrog using plain e-uae from debian repository. I assume original 060 speed should be no problem as well(without jit)

_________________
AROS Broadway - AEROS - Aminux - AmiCloud - indieGO! Appstore - AmiWallet - VAN lossless video codec - AMC Amiga media Center -KrypUnite - LibertyNet - MinX - amigaNX

@phoenixkonsole

If it can get decent speeds with AGA, it may be a solution.

As for the PPC JIT, you'd be running the JIT PPC version of UAE on ARM? I'm a bit confused as to how that would work.

Although one potential issue, no matter how well it runs, is if it would increase developer interest. Sometimes original hardware just has that 'cool factor' that attracts people. A new Amiga using the original chipsets (and spruced up graphics) + fast CPU, in a FPGA, all hardware, probably would be of greater interest than a simple ARM device running an emulator.

Or at the least, the Amiga chipsets in FPGA/ASIC, using a CPU with JIT, could be considered 'original' enough.

Sure, the results may be similar in some ways, but as for marketing and gaining interest out there, I do think a hardware solution would be more popular. Even better if somehow, someway, a cheap $50 FPGA could get Amiga branding.

Last edited by Nameless on 24-Jul-2014 at 10:01 PM.

@Nameless
WiiUAE updated to make use off ppc jit is what i have in mind. Ppc jit is available through dolphin emulator.
AmiQube will come with amigaOS and Aeros is only an option. As soon indieGO! Os is ready it will be available for it as well.

Last edited by phoenixkonsole on 24-Jul-2014 at 10:06 PM.
Last edited by phoenixkonsole on 24-Jul-2014 at 10:05 PM.

_________________
AROS Broadway - AEROS - Aminux - AmiCloud - indieGO! Appstore - AmiWallet - VAN lossless video codec - AMC Amiga media Center -KrypUnite - LibertyNet - MinX - amigaNX

Why bother with ARM JIT, when today there are very cheap and fast low power x86 SoCs?

_________________

@WolfToTheMoon

How cheap? It'd need to be around Raspberry Pi levels of cost.

@phoenixkonsole

So the plan is to run an emulator inside another emulator? Obviously it's hardly ideal, but wondering just what sort of speeds you could get with that.

@NutsAboutAmiga

When I have that device I will inform you

@Nameless

As somewhen else already mentioned people have a oldfashioned interest in real hardware so a FPGA device "simulating" a kind of real hardware will attract people more than a ARM (or X86/X64) device emulating it with the same speed. It is a kind of feeling, not rational. On the other hand other people who do not love Amiga would prefer a ARM or X86 device. Both are basically different groups, I see no real competition there. So both should be done, emulation will help to create a bigger platform in a short time than only with FPGA based devices, on the other hand these devices will attract users and developers that are not interested in emulation. Both are two sides of a coin, helping each other and not harming. My view.

@NutsAboutAmiga

I agree with Samurai there "NG" never was able to become successor regarding codebase. When I was busy with building up I was looking for software all over the web and expecially in aminet. When I looked at the dates and at the ports only a fraction of the 68k software was ported to NG and most 68k software was created 1998 and earlier. 1999 and later activity slowed down visible. The NG platforms (AmigaOS and MorphOS) were created around 2001 but obviously many developers already left before. So most software for amiga is 68k, that has not changed much in recent years. A platform directly building up on that codebase would make much (if not most) sense?

What Samurai mentioned was the lack of compilers on NG platforms. When we want to run browser games we need updated browser. But what is special there? Why needing to use Amiga for it? There are interesting languages that are special to Amiga like Amiga E or PowerD. If we only offer the same as Linux or Windows why not using Linux or Windows? In the old days we had our own innovations, the platform must be interesting. Regarding AMOS, there are a lot of extensions (many of those are free) including those for AGA and Intuition so you can do more than you think. But even if you dislike AMOS you have plenty of alternatives.

@OlafS25

I'm the fellow who just posted above that real hardware would attract more users, so yeah, I understand that mindset. And I never meant to infer that both solutions weren't viable. Just that to reach a mass market, price is key, and the point should be to attract indie developers.

So if we are talking about a $50 FPGA, it could theoretically reach 10K-100K+ new users.

An ARM or x86 device, probably not as many, unless there was some really good marketing and really good emulation. And I'd expect an x86 device to get nowhere near that pricepoint, while ARM could do it ... but that is where the lack of JIT hurts.

You also get into the weird area of developers wondering why not just develop for Android or Win if using those chipsets, rather than code for an emulator.

Last edited by Nameless on 24-Jul-2014 at 11:47 PM.
Last edited by Nameless on 24-Jul-2014 at 11:46 PM.

@NutsAboutAmiga

Quote:

NutsAboutAmiga wrote: @Samurai_Crow I remember Qbasic and TurboBasic, this was the days also remember I made game in TurboPascal 7, and lots of crap in Amos. Amos gave people who was not experienced a way to make games and programs, it also stole a lot of time that I might used to learn programing C, or some thing more useful. Maybe its best forgotten.

If that's best forgotten, then surely the OS that inspired it is also best forgotten! Along with any version of C prior to ISO C 2011/2012 which allows type-safe macros functioning as generics!

Quote:

Shure its possible to make a AMOS clone, but does not make sense to do it like it was back then, yuu be just thrawing away too many CPU cycles on P2C routines and other silly things. And what about modern features like alpha blending and things like that, no it does not make sense.

A SuperAGA chipset core that supports alpha-blending and all of those features is in the works.

Now how many clock cycles do you throw away by running things in EUAE? Petunia? Trance? LOTS! RISC machines are just chock full of wasted clock cycles because they wanted to believe the megahertz myth. Hybrid CISC/RISC machines like the 68060, Phoenix core, and Intel Pentium and up deliver the best performance per clock of any CPU.

Quote:

I wanted to do the same thing some time ago, it was before AmigaOS4.x, I was trying to make Amos clone for Linux, this was because had made a program I called “Adventure Editor”, it was major program to make adventure games. I think for anyone that want to learn to program, that does not want to reboot there Amiga for every mistake, there is web development Javascript, and there is Hollywood, personally I don't think that are that many people being interested in that kind of thing now.

If I only wanted to write programs on JavaScript and Hollywood, I'd buy the Windows version of Hollywood and develop webpages on Dreamweaver instead.

Also, if the Amos interpreter requires a reboot every time you make a mistake you've got one screwed up configuration!

Quote:

AmigaOS it self provided a few legacy draw effects by using graphic library it uses the old-fashioned pen system, so it does not matter if its 32bit or 4 colors, it certainly possible to make programs work from every thing from Amiga 500 to AmigaONE-X1000, if you want to, I'm sure there are some pitfalls. Actually run "Last Hope" written for AmigaOS 1.2 on AmigaONE-XE / AmigaOS4.0 (no UAE), while porting Catwessel driver from Linux to AmigaOS, I used "Last Hope" for testing .

The pen-color thing was a mistake on the Classics. What they should have done to avoid palette collisions was require that every image be on a separate screen with its own palette so that the chipset could fix it for them. Have you ever tried palette-cycling on a screen that's been loaded in from a datatype? It doesn't work due to the pen-mapping!

@Nameless

I contacted former amiga developers and had the inpression that there would be interest in a new updated 68k device. Of course you need hard facts like how much does it cost, how often it is sold, what are your plans for the next 1-2 years. When you deal with people without any amiga background only hard facts are important (how much does it cost, how often it is sold). If we could say that the software (f.e. a game) also runs without changes on other platforms like ARM or X86 it would certainly help. The more potential buyer the better. But hardware is still important (and be it only for emotional reasons).

@Samurai_Crow

Quote:

If that's best forgotten, then surely the OS that inspired it is also best forgotten!

Yes it is, its Atari TOS, its was ported to AmigaOS :-p

Quote:

Along with any version of C prior to ISO C 2011/2012 which allows type-safe macros functioning as generics!

Amos is Basic language, maybe your thinking about Microsoft Basic maybe?

Quote:

Now how many clock cycles do you throw away by running things in EUAE?

Too much, better write native programs, or make the program system friendly so they work with Petunia.

Besides, who want to run things in emulator when it can be run in the OS.

On the other hand it has no problem running at full speed any classic 680x0 program, 2.4 x faster then a 68040 25Mhz is what SysInfo says, It says I have 782 Mhz 68020

Quote:

Petunia? Trance?

This two make use of PowerPC native libraries and devices and so they are pretty efficient. And they also do not need to visualize the memory.

Fake AGA is slower then UAE AGA, so it depends what it is. The best programs or games to run this way is programs that support cyber graphics, Picasso96 or programs that can run on the workbench screen.

Quote:

Hybrid CISC/RISC machines like the 68060, Phoenix core, and Intel Pentium and up deliver the best performance per clock of any CPU.

The fastest I have seen is 105 Mhz 68060, maybe its more efficient then WarpOS (hack) + CyberStrom/BlizzardPPC, even if its fast per clock its slower then Sam440 800Mhz, and I think the Sam440 is too slow for what expect from computer today.

Quote:

Also, if the Amos interpreter requires a reboot every time you make a mistake you've got one screwed up configuration!

I was thinking about writing assembler or C, where you need to handle memory by your self.

Quote:

It doesn't work due to the pen-mapping!

Well that's a limitation of the program you use, not the OS in it self, in AmigaOS you have full control over pen allocation, and you can change the colors if you like.

Last edited by NutsAboutAmiga on 25-Jul-2014 at 12:00 PM.
Last edited by NutsAboutAmiga on 25-Jul-2014 at 03:02 AM.
Last edited by NutsAboutAmiga on 25-Jul-2014 at 02:59 AM.
Last edited by NutsAboutAmiga on 25-Jul-2014 at 02:57 AM.
Last edited by NutsAboutAmiga on 25-Jul-2014 at 02:54 AM.
Last edited by NutsAboutAmiga on 25-Jul-2014 at 01:49 AM.
Last edited by NutsAboutAmiga on 25-Jul-2014 at 01:37 AM.

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

@NutsAboutAmiga
Just to clarify again:
Fpga is nice but Limited. You will only reach amigans.
For Sure the lowest possible number.

The amiqube can Run Android an so it is on par with ouya.
You can Run aeros and so it is a desktop on par with lowend pc's
It has wifi, bluetooth, SATA and Infrared so you can use it as xbmc Media Center.
This are 3 Individual user Profiles.
Indiego-os will bring 30years of gaming to it so there is a 4th user Profile.

Aros developers get it for 50% which means sub 50€.

_________________
AROS Broadway - AEROS - Aminux - AmiCloud - indieGO! Appstore - AmiWallet - VAN lossless video codec - AMC Amiga media Center -KrypUnite - LibertyNet - MinX - amigaNX