FPGA based A500 accelerator

[Mathesar]

Originally Posted by Seiya

i think that these new and open hardwares destroy a little Amiga productivity because it several times that i not read about news games or new apps (ok, that some time someone release a game, but very few things). Why in 2021 i have to buy a card to use the same software that i used happily with A500 or A1200?

Today i think that everyone know that Amiga is requested in very hard way in retrogaming market and not in hardware market.

The only software that i see very very most daily updated are Sysinfo and Aibb that are only software that support new hardwares :P, so i see so many benchmarks discussions and threads and stops.

Amiga needs strong advertise in retrogaming world much more than is today

I somewhat agree with you. As much as I love these hardware projects I agree that what Amiga keeps alive as a retro platform is software and games. However, most of that software targets standard Amiga 500's or 1200's. And that makes sense as well. the whole point of the "Dread" project for example is to prove that the Amiga 500 could have a decent Doom back in the day. Making it run an a Vammpire / Buffee / whatever makes no sense in that regard.

However, I still think that it makes sense to expand the Amiga 500 with things like extra RAM, IDE, scandoubler, SD-card, etc. It just makes it easier to use things like WHDload or exchanging software with a PC. A fast CPU is less useful imho (says the one that build an A3660 for his A3000 )

[Seiya]

Really i know that Doom like games had a strong appeal many years ago and i used and brought classic accelerators cards because really need to play with them. Today is different. If i really want to play with Doom like i not use Amiga. Amiga i love only classic vintage game that run on standard Amiga.

[Mathesar]

Originally Posted by Seiya

Really i know that Doom like games had a strong appeal many years ago and i used and brought classic accelerators cards because really need to play with them. Today is different. If i really want to play with Doom like i not use Amiga. Amiga i love only classic vintage game that run on standard Amiga.

Back in the days it was different yes. Then the more speed the better to run things like word processors, ray tracers, web browsers, email, Spice simulators (in my case), etc, etc. Back then a fast Amiga was also competitive compared to an Intel PC and even faster than a classic Macintosh.
But now even the fastest Amiga is not enough to run a modern HTML5 web browser for example. The only reason I built the A3660 was to restore my computer to its "back in the day" specs.

Still, for me the Amiga hobby is all about fun and projects like this are just that: fun! For running new and exciting games, demos and other software a modestly expanded A500 is all you need though.

[torsti76]

Honestly, I don't see the point.

Of course there's very little going on on the software and game market for the Amiga these days. And that's really sad. For instance, I'd really like to see some Lukas Arts style adventures again.

On the other hand, the Amiga has always been a very expandable system. Combined with todays possibilities from the maker scene, this made and makes a whole lot of things possible that simply didn't exist back in the day. Tinkering with the Amiga's hardware made me understand how computers work internally, how their parts are divided into functional groups and how these groups interact with one another to form a whole.

Creating new hardware for it doesn't take anything away from software development. I don't think that people who develop hardware would instead write great software.
The problem is that a great software usually requires a whole team of people: Programmers, artists, composers, writers. That's really hard to achieve without the infrastructure of a company.

A good piece of hardware, on the other hand, is reasonably easy to develop by a single person.

In fact, I know many good hardware projects driven by one or two people but essentially not one single astonishing game created by someone on his own...

[AJCopland]

That's... all a bit off topic.

@mkstr, yes I do think there is still a place for this accelerator.

I did like your initial approach of using a standard FPGA board, but a custom board for A1200/etc sounds necessary from what you've been saying.

[utri007]

There are demand for this. If price is about 140-150€ this will sell very well.

[zipper]

Originally Posted by torsti76

On the other hand, the Amiga has always been a very expandable system...

Which other system can be accelerated x100 like my A500T with Blizzard 2060@64 MHz? At least pretty hard to PC gear...(crazy...)

[Promilus]

@zipper - acorn archimedes with raspberry pi You can also buy modern pc system on an isa card, insert to old 386 system, hook kb, mouse, vga, storage to it and use mainboard as power supply alone... lol Home amiga never was expandable system. All it had was either trapdoor memory (and side expansion) or trapdoor memory + turbo. It is quite amazing so many products were developed over years but that's mostly because there were no new models. When pc users happily swapped their isa and vlb 486 machines to pci & pentium world amiga users were didn't even have PCI (yet) and were using workarounds to get things like new graphics modes using chipset (graffiti cards for ECS) or soundcards using clockport. That's also the reason why you wouldn't see 486 machine equipped with pentium III while there were plenty of Amiga 1200 using 060+603e duo. It's not because Amiga expandability was better. It's because there was no other choice.

What many ppl have to yet realize - Amiga is slow. It's very, very slow. How turbo card works? Well they do have local memory which allows way faster than original CPU retain it's speed. Any program running inside that local memory is going to execute fast. But. Every access to amiga peripherals is basically just as slow as with original CPU. That's also a reason why - during development of new turbo cards - companies begun to install faster peripherals onboard turbo card to use it's speed to the fullest. That's why mid 90s turbo had SCSI controller (faster than native) and in late 90s even fast local bus for expansions (by default - graphic card but soon enough - pci slots). That's because amiga resources were lagging far behind CPUs available at the time. So there's little benefit of having 040@40MHz (or better) CPU paired up with ECS alone and PIO0 IDE controller. It's just things are. So either super ultra powerful CPU with subsystem allowing super fast peripherals or just dirt cheap expansion with essentials.

[Bruce Abbott]

Originally Posted by Promilus

@zipper - acorn archimedes with raspberry pi

Yuk. I recently bought an Archimedes A3000 with an internal 120MB 2.5" IDE hard drive and fully expanded to 4MB RAM. What a disappointment! The 8MHz ARM processor is supposed to be much faster than a 68000 but doesn't feel like it, and it chews up memory twice as fast. The hard drive is dog slow, the GUI is non-intuitive, the OS is not multitasking, and you have to manually tell it how much memory to allocate for graphics etc. It has a 4096 color palette and up to 256 colors on screen, but only 16 colors can be freely chosen. Apart from one sprite it has no advanced graphics hardware - just a dumb frame buffer. The audio specs look good until you realize that all channels are stuck at the same sample rate. It has a 720k 3.5" floppy (less storage per disk than the Amiga), but no way to add more drives for copying disks etc. The whole thing seems clunky and full of strange quirks and compromises.

I have heard that hardware expansions such as faster CPUs and Ethernet cards are available for the Archimedes, but can't find any for sale (not that I would buy them anyway - no point trying to polish a turd).

Compare that to my A600 with Vampire accelerator. CPU equivalent to a Pentium 120 or higher but running memory efficient 68k code, 128MB of super-fast RAM, high resolution HDMI graphics with 24 million colors as well as the Amiga ECS chpset, superior multitasking OS, PCMCIA port which takes standard Ethernet and CF cards etc.

The A600 was supposed to be the least expandable Amiga, but it didn't take long for hardware developers to get around that. My current machine is more powerful and much cheaper than the Amiga 3000 system with 50MHz 060 and RTG etc. that I used to have, and it's a joy to use.

You can also buy modern pc system on an isa card, insert to old 386 system, hook kb, mouse, vga, storage to it and use mainboard as power supply alone... lol

Totally pointless, and not even possible with many PCs of the day.

Home amiga never was expandable system. All it had was either trapdoor memory (and side expansion) or trapdoor memory + turbo. It is quite amazing so many products were developed over years but that's mostly because there were no new models.

The side expansion on the A1000 and A500 exposed the full bus and did not limit what could be added. Many users had accelerator / hard drive expansions which that port was designed to accept. Also many people had an A2000/A1500 (which were affordable to home users in the base configuration) that they stuffed full of cards.

The A1000 and A500 sensibly have a socketed 68000, with enough room around it to install a large internal expansion card. A Vampired A500 is quite a beast, but just adding a fast CPU, RAM and IDE port is well worthwhile. I am currently building a TF530 with 50MHz 030 that will make my A500 twice as powerful as my A3000 was when I bought it in 1991, for less than NZ$200.

Now that sophisticated expansions can be made cheaply and easily by hobbyists, we are seeing a renaissance of DIY designs that are pushing the Amiga well beyond what was ever thought possible. And this just the beginning. With FPGAs, fast cheap MCUs and dev boards etc., there really is no limit to what we can achieve today - and no limit to the fun we can have doing it!

When pc users happily swapped their isa and vlb 486 machines to pci & pentium world amiga users were didn't even have PCI (yet)

By that time the Amiga was already effectively 'retro' as the manufacturer was long gone. But we didn't have to suffer the pain of doing continual upgrades to keep up. Think about people who bought 386SXs at the same time we were getting A1200s. First they had to move to 486 with VL bus, then Pentium with PCI, different CPU sockets, different Video slots, types of memory etc.... and they are still doing it today! In most cases each upgrade meant buying a whole new machine because little could be used from the old one - either due to compatibility issues or simply because it was no longer enough. Many did not do it 'happily'.

That's also the reason why you wouldn't see 486 machine equipped with pentium III while there were plenty of Amiga 1200 using 060+603e duo. It's not because Amiga expandability was better. It's because there was no other choice.

Not true. Some 486 motherboards could take a Pentium CPU, but it generally wasn't done because just changing the CPU wasn't enough. The ever-increasing hardware specs required to run the latest software meant you also needed more RAM than would fit on the motherboard, a new graphics card slot, bigger hard drive, even a bigger power supply and different sized case. Most of the time it was cheaper and easier to just buy a whole new machine, and that is still true today. So in practice it was PC users who didn't have a choice.

Meanwhile new generation Amigas were also being produced. So Amiga users did have a choice, but many stuck with their existing machines because they didn't need to upgrade. I put a 50MHz accelerator card and 1GB hard drive in my A1200 in the 1990s, and didn't do anything more to it until a few years ago when I upgraded the 16MB SIMM to 32MB. I expect that to be sufficient for a long time to come. I could buy an NG Amiga if I wanted, but why would I want to?

What many ppl have to yet realize - Amiga is slow. It's very, very slow.

I think we do know that.

But strangely I can turn on my A1200 and be downloading the latest files from Aminet and reading posts here before my 2.8GHz dual-core PC has even booted into Windows (and then takes another minute to see the network and run Firefox). I never have to wait for my Amiga to clean up virtual memory (which sometimes takes minutes on my PC), and when I am finished with it I just turn it off! Depending on that you are are doing with it, an Amiga can be plenty fast enough.

[Promilus]

@Bruce - well CISC wasn't only "complex ISA" but also "compact binary". It is quite obvious (or at least was back then) that CISC produced code better suited for low memory systems. That's also a reason why with memory constrained systems (microcontrollers) ARM introduced special ISA Thumb to be a more compact one with minimal performance penalty. But ... no, 2D capabilities of Archimedes weren't exactly perfect but CPU itself - yes, it was fairly powerful. And it shows in those silly pseudo 3D games which run quite nicely on ARM and choked on Amiga.

Compare that to my A600 with Vampire accelerator. CPU equivalent to a Pentium 120 or higher

Oh well, BBC products were rather popular in UK alone, Commodore went to a great lengths to let amiga roaming worldwide. It's thanks to way bigger community it was possible for some expansions to appear. On the other hand since all ARMs are pretty much backward compatible (ISA-wise) to ARM2 used in Archimedes it would take even less effort to produce extremely powerful ACORN Turbo. I wouldn't gloat... that's basically the best you can achieve with fpga implemented 68k design. And Archimedes (hypothetically) can get whatever he wants from ARM world. And since OS has been open for quite some time (unlike AOS) well... You know.

Totally pointless, and not even possible with many PCs of the day

Oh retro is pointless logically speaking. And no, you misunderstand, there are SBC ISA based. So whole computer on ISA or PCI bus. Usually meant for embedded and industrial solutions and running on dedicated busboard (with nothing particular up there). They also can run in A3k ISA bus or similar age 386/486 machines basically as independent system using motherboard for power alone. And that's what sandwitched expansions turned amiga boards into as well.

The A600 was supposed to be the least expandable Amiga, but it didn't take long for hardware developers to get around tha

It still is and every gaming amiga was, is and will be. Thinking that integrated SoC like solution put onto CPU itself is expansibility no other system can achieve is plainly stupid. PC didn't use that solution because there were new models. And Acorn had way smaller community, along with other niche solutions. And Macs - well there were new macs with different architecture but having fair backward-compatibility emulation. So again - you gloat over something which was pointless in different computers but was necessity on Amiga. If you do count new SAM and AmigaOne products as valid successors then all that gloating is silly because that V2 A600 is way behind what G3 500MHz with AGP Radeon can do.

The side expansion on the A1000 and A500 exposed the full bus and did not limit what could be added

Oh well, if you insist... SO IS ISA! And guess what, there is no core 2 turbo on ISA card for PC for a reason I did already mentioned! And 95% of A500 accelerators were using 68000 socket. Only 2 used side expansion BEFORE commodore went bankrupt. That's Supra Turbo and GVP530 - both several years after A500 release. And ACA - well that's quite some tome after Commodore went bankrupt. Commodore designed that for a handful of things, it doesn't matter if one used it to get A590 alone or Micronic Z2 board with Blizzard 1260 and unreleased PPC prototype adapter. Commodore did not intend to use it as such. It was a bizarre and extravagant solution to use it that way. And if popular - only because THERE WAS NO OTHER CHOICE!

With FPGAs, fast cheap MCUs and dev boards etc.

doesn't make them more expansible than before. It only shows how desperate retro gamers are. There is no technical reason why solutions like that couldn't be applied to any other computer popular at that time. It only makes no sense since most popular didn't died out and in fact continued their development. That's why we do have pcie4 and nvme storage with speed levels even v4 is put to the shame - despite it's price.
So new possibilities of Amiga hardware are not through extraordinary capabilities other hw didn't have but rather stubbornness of their owners and better, cheaper development tools for programmable logic. So please do not treat Amiga like it is better because you can put FPGA-based SoC on top of the CPU ... reality check - you could've done that with many more but it was pointless to do so.

By that time the Amiga was already effectively 'retro' as the manufacturer was long gone.

I disagree. Last brand new amiga 1200 was still sold till 1996. It didn't change into retro even till late 90s since there was a strong push in development of a stop-gap ppc-68k hybrid accelerators and sandwitching amiga with expansions. That's right... that's the era where blizzardppc and elbox mediator came out. Nobody at that time considered it a dead end and retro.

and they are still doing it today

And that's why they enjoyed doom right after release, quake right after release, tomb raider, dune 2000, red alert 2, crysis, witcher 3 and so on. That's why you change cars and get better one, more efficient, safer instead of tuning old one. It's healthy situation as long as the change isn't artificially forced through. Problem with amiga community is - we didn't jump to the PPC wagon (and was good to do so since it was dead end anyway, but didn't appear as one at the time). But that said - we're decades behind mainstream. We'll never be anywhere close. Deal with it.

Many did not do it 'happily'.

Oh so you claim you didn't enjoy changing your hardware and playing new games because it did cost you a lot but you do enjoy changing your hardware and playing OLD games because despite the fact it costs you a lot... Ok.. makes perfect sense.
Nope. No 486 motherboard could take Pentium CPU. Some could take PentiumOverdrive CPU which was basically castrated Pentium to match 486 front bus. That's typical "turbo" enhancement of PC since it was pointless to do more. With that regard 386 owners had the ability to get 486DLC/SLC and DxR chips from Cyrix/Texas which were also turbo-like enhancements. And fairly popular at the time (considering units sold) but overall not that popular and considered being pointless attempt to breathe a new life to old units.
Something we do even now and do not consider it pointless - but exactly because we do not have other choice in that matter. There are no new amiga in classic line to jump into. And the best of ppc line are extremely expensive with subpar performance.

So in practice it was PC users who didn't have a choice

well the most work in the last year was towards porting diablo to amiga. That's more than 20 years behind PC. Guess who's the winner then.

But strangely I can turn on my A1200 and be downloading the latest files from Aminet and reading posts here before my 2.8GHz dual-core PC has even booted into Windows

Strangely enough my pc boots up in 10s from scratch, probably 2 from StR and download things with several MB/s due to both fast storage and fast ethernet. I can dump movie to USB3 drive with ~40MB/s write while Vampire fast ide read tops out at 20. We do things and spend money on things which are far, far behind what either x86 or arm is capable of. It doesn't make sense so justifying it with reason is BS. We do things because we love Amiga.

[Bruce Abbott]

Originally Posted by Promilus

Archimedes (hypothetically) can get whatever he wants from ARM world. And since OS has been open for quite some time (unlike AOS) well... You know.

Risc OS is open source because that's all it's worth, like TOS. Amiga OS isn't because it's still worth something. But we do have AROS for those who can't stand the idea of a properly maintained commercial standard.

Oh retro is pointless logically speaking.

Life is pointless logically speaking. In a few years I will be dead for sure, so why bother doing anything? But I like using my Amigas and want to get more enjoyment out them. That includes programming in 68k assembler, playing games that I never had or didn't have to time to finish, and making my machines work better without losing the things I love about them. I gave up on the Amiga for a few years because I had other interests, but now I am getting back to doing what I always wanted to do with them but didn't have the time.

In the 1990's I owned a computer shop and did commercial Amiga development. As with all hobbies, doing it for profit tends to distort your outlook and kill the enjoyment. The fact that Amigas are now officially retro takes the pressure off and allows me to fully explore their character and potential. No having to push stuff out as soon as possible to beat the changing market, no need to compete with PCs, no concerns about whether it is profitable or not - just pure fun! That's what makes life worth living.

And no, you misunderstand, there are SBC ISA based. So whole computer on ISA or PCI bus. Usually meant for embedded and industrial solutions and running on dedicated busboard (with nothing particular up there). They also can run in A3k ISA bus or similar age 386/486 machines basically as independent system using motherboard for power alone.

Yes, I understand that. But these embedded systems are expensive and not intended for the use you are proposing. And no matter how you do it they are still a PC. I have plenty of experience with many kinds of PC, having had an IBM JX, a Sidecar on the A1000, an Amstrad PC1512 and Sinclair PC200, and a plethora of PC clones and machines made from parts. Today I have several PC boxes (two of which get regular use) and laptops of various ages, as well as about a dozen other retro home computers. But when it comes to playing with them the Amiga gets most of the attention because I like it more.

It still is and every gaming amiga was, is and will be. Thinking that integrated SoC like solution put onto CPU itself is expansibility no other system can achieve is plainly stupid. PC didn't use that solution because there were new models. And Acorn had way smaller community, along with other niche solutions. And Macs - well there were new macs with different architecture but having fair backward-compatibility emulation.

There are those who just want to play games and the hardware is merely a tool to do that. Those people should use the latest PC or console that works best with their games.

Then there are those to whom the computer itself is a 'game', with the 'play' being tuning it to their tastes. Those of us who are doing that to our Amigas don't care what some some other architecture may be capable of - it's irrelevant, just like a chess player doesn't care what new hull designs are being used in the latest racing yachts.

So again - you gloat over something which was pointless in different computers but was necessity on Amiga. If you do count new SAM and AmigaOne products as valid successors then all that gloating is silly because that V2 A600 is way behind what G3 500MHz with AGP Radeon can do.

Firstly I am not 'gloating', just presenting a point of view that doesn't follow the PC herd. Just like different lifestyles are valid, so are different tastes in hobby computing. We don't upgrade our Amigas with genuine 68k CPUs or closely equivalent FPGAs because we have to, we do it because we want to. Many of us choose to stick with classic Amiga hardware and software because that's what we like.

95% of A500 accelerators were using 68000 socket. Only 2 used side expansion BEFORE commodore went bankrupt. That's Supra Turbo and GVP530

The GVP A530 was an awesome expansion, but expensive and more than what most A500 users needed in the early days. People who wanted more generally bought an A2000, for which Commodore and others made many accelerators and other cards. There was also an external accelerator called the Golem Turbo 1000 for the A1000, which could take up to 16MB of RAM. But again most users didn't need it.

I had a 2MB expansion and self-built hard drive interface attached to my A1000 and was quite happy with it for many years (until somebody stole it). I bought an A3000 for development to speed up assembling code, then discovered that Hisoft Devpac was 5 times faster than the assembler I was using so I didn't actually need the A3000!

doesn't make them more expansible than before. It only shows how desperate retro gamers are.

Nonsense. The main reason 'retro' gamers stick with the A500 is because that is the machine most 'retro' games were designed for. Having a faster CPU etc. is nice, but sometimes it just gets in the way. PC owners discovered this early on, which is why most PCs of the day had a 'turbo' button. I also discovered this when left with only the A3000, and I stopped playing games for a while because getting them to run was too much of a pain (didn't have WHDload etc. back then). Of course PC owners had similar trouble, only worse due to all the not quite compatible clones out there.

I also have an A500 today that mostly gets used just for games. Being a stock 1MB machine with KS1.3 makes it more compatible and simpler to use.

There is no technical reason why solutions like that couldn't be applied to any other computer popular at that time. It only makes no sense since most popular didn't died out and in fact continued their development. That's why we do have pcie4 and nvme storage with speed levels even v4 is put to the shame - despite it's price.

But it doesn't matter what you do to a PC, it's still a PC. The constant hardware and software development has resulted in a continuum of clones each slightly better than the last but with no character. As soon as you get one set up the way you want, you are forced to 'upgrade'. I gave up a few years ago because I think Windows XP is the best and can't stand later Windows versions. So now my 2 year old web browser is having trouble on some sites for no reason other than the developers can't be bothered (which is probably fair enough considering the monsters that 'modern' web browsers have become).

On the hardware side it's a similar nightmare. My PC started randomly crashing a month back, due to burst caps on the motherboard (which I replaced) and then faulty RAM. I have a bag full of DIMMs which are all different and only work in certain motherboards, not including mine! Luckily I found a compatible pair in another PC that I got for free from a local computer shop.

Pcie4, nvme bla bla - so many acronyms I have lost count and have no interest in learning about. It was bad enough back in the 90's with different slots, incompatible SIMMs, hard drives that wouldn't work in some machines because they had 3.3V logic etc. I'm too old for that crap. PC hardware and software is so complex today there's no way a normal person can fully understand it, so we are forced to just being monkeys using what we are given. More and more hobbyists are moving from the PC to platforms like Arduino and Raspberry Pi where they can get down to the metal like we are used to on the Amiga.

So new possibilities of Amiga hardware are not through extraordinary capabilities other hw didn't have but rather stubbornness of their owners and better, cheaper development tools for programmable logic. So please do not treat Amiga like it is better because you can put FPGA-based SoC on top of the CPU ... reality check - you could've done that with many more but it was pointless to do so.

Just imagine trying to piggyback something on the back of a Pentium CPU. It wasn't just pointless, it was also technically close to impossible - a far cry from what we can do to a classic Amiga. That means PC hobbyists are missing out on a fun activity that we enjoy.

I disagree. Last brand new amiga 1200 was still sold till 1996. It didn't change into retro even till late 90s

In the real world the Amiga died as soon as Commodore did. I waited several years for Escom to revive it and did manage to sell a few of their A1200s, but the market was all but dead by that point. Amiga users took advantage of falling hardware prices especially since vendors were trying to move stock before the market collapsed completely, but everyone new the writing was on the wall. We all loved those cover CDs with full programs for free too, knowing it was the death throws of our hobby.

At the time we were disappointed and demoralized by it, but in the long run it was a good thing. If Commodore had survived we would have been subjected to the same crap as PC and Mac owners, endless 'upgrades' that got further and further way from what we loved. By today the Amiga would have from indistinguishable from a PC clone and just as inaccessible.

Nope. No 486 motherboard could take Pentium CPU. Some could take PentiumOverdrive CPU which was basically castrated Pentium to match 486 front bus.

You contradict yourself. Many 486 PCs were advertised as being upgradeable to Pentium, but by the time that became viable it was better to just replace the whole motherboard. And of course since then a plethora of different Pentium sockets came out, including some that were actually a whole PCB that plugged into a slot on the motherboard.

well the most work in the last year was towards porting diablo to amiga. That's more than 20 years behind PC. Guess who's the winner then.

People do what they enjoy, and sometimes a challenge like that is worth just for the attempt. But I am not at all interested in playing Diablo on any platform, so whether it was the 'most work' or not doesn't concern me. If that's what they enjoy doing then good on them for doing it - perhaps the experience will enable them to port a game I do want to play.

Strangely enough my pc boots up in 10s from scratch, probably 2 from StR and download things with several MB/s due to both fast storage and fast ethernet. I can dump movie to USB3 drive with ~40MB/s write while Vampire fast ide read tops out at 20. We do things and spend money on things which are far, far behind what either x86 or arm is capable of. It doesn't make sense so justifying it with reason is BS.

It's not 'justification' to admit your favorite PC is not as fast in practice as it should be. Perhaps I could get better performance from a new PC, but I don't have several thousand dollars to spend right now (or ever if this Covid thing doesn't get sorted soon). I just spent $230 on a new cellphone which I need for contact tracing. Then there's insurances, rates etc. all sucking out money when I have no income. But my Amigas are still going strong and don't need a lot of money for upkeep.

We do things because we love Amiga.

Finally we agree on something!

[Promilus]

To not spam in original topic much more I think you misunderstood what I was writing about in the first place. A500, A600, A1200 were designed as a short-term computers with just a handful of expansions (being either not-entirely zorro II, not entirely zorro III or not entirely pcmcia). Those computers were supposed to be obsolete, replaced by newer ones which never came. That's the only difference. There were no hardware superiority which allows A600 to get pentium-class processor now. It's not hardware based possibility. It's community based. Atari community forked to Coldfire at some point with only slightly better fate than infamous Dragon. Amiga lives through community alone. Just like Atari, just like BBC, just like C64, Spectrum and so on. There is nothing more special than community right here. Not logo, not hardware, not OS.

More and more hobbyists are moving from the PC to platforms like Arduino and Raspberry Pi where they can get down to the metal like we are used to on the Amiga

YEah, and they do that by writing in Python or C++ without even knowing what lies under it... makes perfect sense to compare it to amiga where most wrote in asm and knew blitter, dma, copper tricks to get the job done, and enjoyed 68k simple and elegant asm.

If Commodore had survived we would have been subjected to the same crap as PC and Mac owners, endless 'upgrades' that got further and further way from what we loved.

Exactly my point. You wouldn't have any more regard towards A600 than to all those crappy PC, Macs, BBC you wrote about. It's only special because it DID END UP DEAD with the release of AGA models and - differently from other platforms - community endured. And it endured because there was a hope for new amiga (ppc based) and/or new expansions for classic amiga. If there was absolutely no hope at that time - most likely community would have scatter much sooner with basically no foundation to things like we have now.

You contradict yourself

Nope. You might think that because you don't seem to understand what Pentium Overdrive was and how that compatibility was achieved.

And of course since then a plethora of different Pentium sockets came out

IF you didn't know 486 used different sockets as well.

and sometimes a challenge like that is worth just for the attempt

I find no reason to assume interfacing modern ARM SoC to Archimedes would be less of a challenge than interfacing modern ARM SoC to Amiga (PiStorm, Buffee). Hardware hacks itself are worthwhile. Should it be SNES with raytracing cartridge, or Amiga with FPGA CPU, or BBC Micro with RPI acting as coprocessors on Tube interface, or C64 with FPGA acting as 1541 drive. And there's pretty big community of ppl buying old PCs as well so it ain't that only amiga community dominates retro.

But my Amigas are still going strong and don't need a lot of money for upkeep.

Well my amigas are in working order as well and I do tend to modify them just because I know how, I want to test their limits etc. Not because it's expansion-friendly. It's not.
If I still had my C64 I'd probably hack it as well. And just because I had several PCs and they shared some kind of continuity I wouldn't go back to the first one to modify it. That being said - I modified plenty of embedded PCs with e.g. SSD IDE drives as a refurbish service with great results. Not many remember SSD aren't that new tech and it was already present in mid 90s as Disk on Module flash drive.

[Bruce Abbott]

Originally Posted by Promilus

To not spam in original topic much more I think you misunderstood what I was writing about in the first place. A500, A600, A1200 were designed as a short-term computers with just a handful of expansions (being either not-entirely zorro II, not entirely zorro III or not entirely pcmcia). Those computers were supposed to be obsolete, replaced by newer ones which never came.

By that criteria the vast majority of personal computers were designed to be obsolete. The PC-XT because it only had 8 bit slots, the AT because it stuck to XT bus speeds, and all 386s that stuck with 16 bit slots. When IBM introduced MCA it made all ISA bus PCs obsolete before they were even designed! The industry responded with EISA which didn't take off, and VL bus which everyone knew had a limited lifespan. Then there were all those laptops that had a proprietary bus connector at the back or nothing but single purpose slots and (later) PCMCIA. Eventually PC manufacturers settled on PCI, but most motherboards only had a few slots. Then when that wasn't fast enough for the latest graphics cards...

That covers the period when the Amiga was produced, but it didn't stop there for PCs. The pace of hardware development meant that no computer could expect to be expandable very far into the future, and eventually manufacturers stopped even trying. The majority of PCs sold today are either laptops or 'all-in-one' machines based on laptop technology, with very limited expansion capabilities.

With the possible exception of the A600, Commodore expected all their Amiga models to have a good lifespan, which is why they included an expansion slot even though they didn't need to because everything you needed was already on the motherboard. Just because the bus expansion on each Amiga wasn't all 'entirely the same' as other models doesn't mean it wasn't meant to be expandable for the foreseeable future (and beyond!).

The Amiga's PCMCIA port wasn't 'entirely PCMCIA' for the simple reason that the standard hadn't been finalized yet. The main problem with PCMCIA was that PC manufacturers attempted to make it into an ISA bus replacement - which made it harder to ensure compatibility on the Amiga - and many PC implementations did not follow the standard correctly.

That's the only difference. There were no hardware superiority which allows A600 to get pentium-class processor now. It's not hardware based possibility.

Nobody said there was a 'hardware superiority', only that having a 'pentium-class' processor is possible because the Amiga was designed from the start to take faster processors (unlike most other 'home' computers). Even the A600 can do it because while a physical bus connection wasn't provided the circuitry still allows it. In fact the circuits of all Amigas share common elements that make adapting an accelerator to different models quite easy - by design.

In comparison most PCs expected the CPU to be installed on the motherboard, and when a new CPU type came out a new motherboard would be designed to take it. So-called 'overdrive' chips were a kludge to get faster CPUs into existing motherboards - a tactic that Commodore didn't need to copy because Amigas were always designed to take a faster CPU via bus expansion.

You wouldn't have any more regard towards A600 than to all those crappy PC, Macs, BBC you wrote about. It's only special because it DID END UP DEAD with the release of AGA models and - differently from other platforms - community endured. And it endured because there was a hope for new amiga (ppc based) and/or new expansions for classic amiga.

I prefer to think of it as a design being frozen at a point where 'improving' it would destroy its essence, and I see modern expansions as realizing that. In the same way the development of OS3.1.4 and beyond is perfecting the OS rather than trying to make it into something else.

I also feel the same about the PC. IMO Windows XP did not need to be 'improved' in the way Microsoft did it. If they had just fixed the bugs and polished it a bit I would have bought the next Windows version - but instead they filled it full of eye candy and silly stuff that just gets in the way and takes up more resources - a bit like Amiga OS 3.5. OS 4 and PPC are even further from what I want, which is to make my wonderful classic Amiga an even better version of itself - not a wannabe PC.

You say the A500, A600, A1200 were designed as 'short-term' computers. I don't know exactly how long Commodore expected to get out of them, but a common criticism is that they tried to make them last longer than they should have. I feel the opposite - it always annoyed me that just as I got to know a platform it became obsolete. As it turns out, that didn't happen to the Amiga. 25 years later we are still enjoying learning more about it and making it work even better!

[Promilus]

The Amiga's PCMCIA port wasn't 'entirely PCMCIA' for the simple reason that the standard hadn't been finalized yet

You are explaining things most ppl here are aware of. You really don't have to. Same with PC evolution, emergence of new standards etc. If I mentioned those I did this to show - in contrast to amiga world - what were we missing. BTW did you know originally there was a movement in C= to have most (big) Amigas with PC capabilities built-in? (which was afterwards only done through bridgeboards).

only that having a 'pentium-class' processor is possible because the Amiga was designed from the start to take faster processors (unlike most other 'home' computers)

Well that's BS. Commodore 64 didn't have any special means to get faster CPU yet it got one with SuperCPU. The very same thing is with Amiga, Atari, Sinclair etc. Every one of those could get pentium class cpu the very same way.

In fact the circuits of all Amigas

It's not amiga feature but 68k feature regarding bus arbitration... Nothing else. And while it is a feature shared with Atari, Macintosh, Sinclair QL - well you see some expansions (like TF-based) for Atari but not so much for either Mac or Sinclair. Why? Those have the same hardware capabilities for expansions. So why not? Why only Amiga? Because of community! That's what I was telling you from the beginning. There is no magic amiga hw capability which allows such expansions. Regardless of what you might think.

So-called 'overdrive' chips were a kludge to get faster CPUs into existing motherboards - a tactic that Commodore didn't need to copy because Amigas were always designed to take a faster CPU via bus expansion.

Oh... and how many faster cpu expansions did commodore produce? The only models with that ideology in mind were A2000, A3000 and A4000 ... They had CPU slot exactly for what you described. A dedicated cpu turbo card. A500 side expansion wasn't used as a turbo until near end of it's designed lifespan and what's most curious - I do not remember titles which required one at that time. From 99% of users back then A500 and A600 were just gaming consoles. They didn't even need a hard drive. And rarely had one. Not to mention turbo which almost nobody had and if someone was interested in faster amiga he bought A1200. Turbo for small amiga was a desperate attempt to make hw attractive to software developers again. To "keep up" with PC. It was failed attempt but at least it gave few fine games, better software and let amiga be competitive slightly longer.

I prefer to think of it as a design being frozen at a point where 'improving' it would destroy its essence

Well now you think about it as such. There was no particular essence back then and Commodore did plan to get away from OCS/ECS/AGA as well. You can check Dave's interviews for confirmation. And despite of that there were plenty of amiga users waiting for new, not entirely compatible, amiga - which never came. Should it come and should it be successful - I think you already know the answer in which row would our classic amiga be sitting. Yes, the same as classic macintosh.

but instead they filled it full of eye candy and silly stuff that just gets in the way and takes up more resources

New games require new hardware, new hardware means new os, new os means new api and new api means new games. That's how it works and that's what keeps companies around it running. Doesn't mean you - as a user - don't get anything in return. At 1990 there probably was typical use of 386DX 25MHz with VGA graphics (256KB) and let's say 2MB of RAM. In 2000 most likely 400MHz Pentium II with 64MB of RAM and 16MB 3D graphics. In 2010 - 4 core 3GHz and 4GB of RAM. In 2020 hexa core/ 12 threads 4GHz and 16GB of RAM with 3D graphics owning at least 4GB. In next decade we'd probably play with chip in our skull connected to all of our senses or at least something like deep dive from Sword Art Online anime... that's real progress. Regardless if it's really what we need - that's something what we can and what we do. Same goes to things like Vampire, PiStorm and Buffee.

[robinsonb5]

Originally Posted by Promilus

Well now you think about it as such. There was no particular essence back then and Commodore did plan to get away from OCS/ECS/AGA as well.

Commodore did indeed plan to move beyond OCS / ECS / AGA - but even back then people were concerned that the new offerings would lose the essence of the Amiga.

I agree that you can only expand an Amiga so far before it ceases to be an Amiga - Graphics cards are nice but now you can't do copper tricks. Sound cards are nice, but now we're into the early-90s-PC-like realms of latency from mixing and pops if the buffer underruns because of system load. Memory protection and virtual memory are nice - but then you enter the realms of responsiveness being destroyed if the machine decides to page out part of a program that's required for its UI to respond.
(A couple of weeks ago I made the mistake of opening a Downloads folder on my Linux box, forgetting that I'd downloaded a *huge* JPEG. The machine decided to make a thumbnail for it; It took me nearly 5 minutes to get that machine back under control. That should not be possible in 2021.)

In a vague attempt to get back on topic... I love that there's so much development happening for these machines: the mega-fast ARM efforts are exciting and interesting in their own right - and I still find myself intrigued by the possibilities of running ARM code natively from AmigaOS - a bit like PowerUP and WarpOS from back in the day. But I also think a fast TG68k-based device will hit the sweet-spot between "Authentic Amiga Experience" and "Why not just use JIT emulation on a PC"?

[Promilus]

@robinsonb5 - ARM code running on AmigaOS 3 is something most ZZ9000 users are trying to get. Either Cyclone V SE or Zynq7k could be used to host 68k softcore along with arm hard processors. Would it be better integration of ARM than in form of RTG card? Basically yes but regardless of how integrated such SoC would be there's problem with caching data from RAM on heterogenous CPU architectures and context switching might kill off performance gains. I do think it's something worth pursing though but that's not something one developer can achieve in a free time, right? Since design is pretty much complete and waiting for manufacturing I think there's no point in creating wishlist now. On the other hand should there be some improvements to TG68 afterwards I'd love them to be implemented to already sold units as well.

[AJCopland]

I'd just like to see _someone_ release one of these TG68 based FPGA systems.
Or using their own core, but from my perspective preferrably TG68 based so that people can modify it.

But just release it. That's where they all seem to fall down.

[dreadnought]

Originally Posted by Bruce Abbott

Most of the time it was cheaper and easier to just buy a whole new machine, and that is still true today. So in practice it was PC users who didn't have a choice.

[Promilus]

@AJCopland - it's "quite simple"
1. Get QMTECH Cyclone V development board.
2. Get Vampire source code (iirc that' version 0.1), available as packed project on majsta.com
3. Buy few buffers with voltage level translations capabilities
4. Install Quartus and import project.
5. Connect unused FPGA pins with 68k pins using voltage level translators.
6. Modify source verilog to match changes (different FPGA, different pins)
7. Synthesize and upload
8. Maybe... you'll get working Vampire V1-like TG68k based turbo If not? try, try, try again.
Or if you are afraid then... wait patiently. Patience is a virtue basically... mandatory in amiga community.

[robinsonb5]

Originally Posted by Promilus

Basically yes but regardless of how integrated such SoC would be there's problem with caching data from RAM on heterogenous CPU architectures and context switching might kill off performance gains.

Indeed - any ARM code would have to do a significant amount of work (decoding a buffer-full of MP3 audio, for instance) before handing back to the 68k for it to be worthwhile.

I do think it's something worth pursing though but that's not something one developer can achieve in a free time, right?

I'm not sure - as I see it there are three problems to solve.
* Establishing a way of sharing memory between the two CPUs.]
* Figuring out synchronisation primitives so that ARM code can be launched and caches can be cleared at the right time.
* Packaging ARM code on the Amiga side so that it can be loaded and relocated if necessary. (Does ARM require any relocations that aren't available within Amiga hunk format, for instance.)

Since design is pretty much complete and waiting for manufacturing I think there's no point in creating wishlist now.

Absolutely - the whole ARM discussion is off-topic for this thread. Nonetheless there are numerous projects now which involve putting an ARM inside an Amiga in one way, shape or form - since they're there anyway, it would be interesting to find a way to use them.

On the other hand should there be some improvements to TG68 afterwards I'd love them to be implemented to already sold units as well.

TG68K is Lesser GPL, which is designed to guarantee exactly that. While it doesn't mean the entire project has to be open source, it does mean that the end-user must be able to obtain enough source / object code to modify the LGPLed portions and build a new binary. (The only way to avoid that obligation is to negotiate a different license with the author.)