Vampire discourse, keep it civil (was: Vampire 1200 V2 waiting times)

[Bruce Abbott]

Originally Posted by TCD

CBM didn't survive its own bad business decisions way before Doom appeared.

And their worst business decision was taking on the Amiga. If only they had stuck to just making the C64 line and PC clones! The world would be a duller place of course, but we would never know what we were were missing...

Doom might have been the final nail in the coffin, but CBM was most likely dead in December 1993 anyway.

And a good thing that was too. CBM died exactly when we needed it to, before they were able to rip out the Amiga's soul and replace it with clone technology (or worse). Unfortunately that didn't stop others from attempting to do it.

Fortunately we now have teams who are getting Amiga development back on track. The Apollo team is one of those. They are the only ones dedicated to extending the Amiga chipset rather than replacing it with clone technology - breathing new life into the Amiga's soul.

[kipper2k]

Originally Posted by bruce abbott

and their worst business decision was taking on the amiga. If only they had stuck to just making the c64 line and pc clones! The world would be a duller place of course, but we would never know what we were were missing...

And a good thing that was too. Cbm died exactly when we needed it to, before they were able to rip out the amiga's soul and replace it with clone technology (or worse). Unfortunately that didn't stop others from attempting to do it.

Fortunately we now have teams who are getting amiga development back on track. The apollo team is one of those. They are the only ones dedicated to extending the amiga chipset rather than replacing it with clone technology - breathing new life into the amiga's soul.

lol

[grelbfarlk]

Yes, Altera is absolutely not a low performing cheap mass produced clone technology. I support this bold decision to salute a bedroom coder level alternative to higher performing silicon.
Have we cancelled Dave Haynie yet for saying he would have adopted the PCI standard if it were finalized during his Zorro-III development?

[Promilus]

It's always funny when someone says things about Amiga which are mostly related to A500, 600, 1200, CD32 forgetting there were expandable models like A2000, A3000 and A4000 which had plenty of expansions going beyond chipset limitations. And that one is amiga heart and soul as well. Buster is part of the chipset as well. And so is Ramsey.

If only they had stuck to just making the C64 line and PC clones! The world would be a duller place of course, but we would never know what we were were missing

We'd have Atari with the very much the same hardware. And Atari just might not make the same mistakes Commodore did while having superior architecture.

[nonarkitten]

Originally Posted by Cego

If somebody decides to write software for AMMX, you really wont lose anything. What are you complaining about?

If and only if the application ALSO supports regular 68K opcodes. If they've made it ONLY for the Vampire, then no one else gets to play. It's vendor lock-in.

Originally Posted by Cego

AMMX makes only sense if AMMX power is needed. You get it? Movie players f.e. are heavily accelerated by AMMX and opens up the possibility to watch videos in higher resolution and bitrate.

AMMX makes no sense, period. The performance gain is minimal and highly selective. AMMX has a specific instruction to make colour table look ups really fast. You'd think it would make something like GNGEO way faster, but cycle-counting the sprite rasterizer, it was only about 10% faster.

And that was Gunnar's highly hand-optimized AMMX code versus C.

So it's not just vendor lock-in, it's mostly useless vendor lock-in.

Originally Posted by Cego

Your AGA programs are also incompatible with OCS. Your FPU programs also wont run on an EC030. These are all exclusive applications through artificial means (whatever that means).

Most authors will compile multiple versions. If someone supports AMMX and non-AMMX systems, then that's fine.

Originally Posted by Cego

You're comparing apples with oranges. AMMX is all about speed, just like Altivec on a G4.

Oh wow are these incomparable technologies.
- AltiVec supports floating-point math making matrix operations really fast
- AltiVec uses massive 128-bit registers, not just 64-bit
- AltiVec has pack and unpack instructions that AMMX is notably missing
- AltiVec has aggregate functions (like pick the max of all fields)
- AltiVec has flags and conditionals and can make branches based on those

More importantly -- and a big reason why AMMX isn't the magic wonder bra all you tits think it is -- is that it's not super scalar. So while you can run two 68K opcodes at once, you can only run one AMMX one, and that makes a big difference because you can do a lot in two 68K instructions. Especially with the 020-and-up bitfield instructions.

Originally Posted by Cego

So AMMX comes from a time where you had to introduce new features and instructions to a CPU, to make it a lot faster at the same clockspeed (68080 runs at 85MHz), because there was no other way.

A lot? No.

And Freescale already HAS a 68K DSP solution. Two in fact. There's the original DSP56K processors which -- fun fact -- are still made. The Quake 2 engine on a DSP56001 + 68030 on the Falcon proves how much of a difference a rather middling little DSP makes. Amiga also got this DSP with the Delphina (sp?) which enabled MP3 decode on a stock 1200.

Modern versions of these chips are dual-core, 250MHz juggernauts in tiny 64-pin QFP packages and could have been added with little thought and zero FPGA resources.

The other is the EMAC that's built into all ColdFire cores. This is more integrated with the CPU and much more limited than the DSP56K, but it's also a tiny amount of silicon to include.

Gunnar HAD the choice to leverage either. Or even both. But he chose neither because it was not about making new opcodes for the 68K to improve performance, it was about...

Say it with me...

VENDOR LOCK-IN.

[Vamplover]

Mods, why it’s okays for others to troll badmouths AMXXX and Gunnar’s here in this threads? I thoughts not allowed and gets peoples banned? Gunnar ships V4 wonder Amiga clones everyday to thousands the customers not vapor.

[jbenam]

Originally Posted by James

You can find a few among the many ports here:

http://eab.abime.net/showpost.php?p=1431788&postcount=2

There are a few missing from that list - like a 3D version of the ESCUtil.

Non 3D version here:

https://aminet.net/package/driver/print/ESCUtil

The 3D version can be found somewhere on archive.org along with a few other things (some 3D and also not on the list) but I have lost the link...

Very nice, thank you. I hope I will finally find something making the time spent in getting the Mediator working worth it

Originally Posted by grelbfarlk

Yes, Altera is absolutely not a low performing cheap mass produced clone technology. I support this bold decision to salute a bedroom coder level alternative to higher performing silicon.
Have we cancelled Dave Haynie yet for saying he would have adopted the PCI standard if it were finalized during his Zorro-III development?

Your posts are quickly becoming my favourite in the entirety of EAB.

Originally Posted by Promilus

It's always funny when someone says things about Amiga which are mostly related to A500, 600, 1200, CD32 forgetting there were expandable models like A2000, A3000 and A4000 which had plenty of expansions going beyond chipset limitations. And that one is amiga heart and soul as well. Buster is part of the chipset as well. And so is Ramsey.

The problem is that there is a shared vision (mostly on the part of Vampire users) that "everything not made by Commodore = not in the original spirit of Amiga". I understand that Picasso96 as a standard was not officially sanctioned, but that was only because it came after Commodore's demise. I have no reason to not believe that it would somehow not officially be adopted by Commodore after seeing the user adoption levels. Later Amiga would've probably adopted PCI (just like Dave Haynie said) and we would've had a similar situation to PCs and Macs.

Truth is, we have to compare OCS and AGA to IBM's original CGA and EGA specifications. Even if they were on add-in cards instead of being built-in (and that wouldn't be the case for the PCjr), clones were 100% compatible and just added additional modes. You had the one true specification and you would just adhere to it. You could also bang the metal if you wanted to get more performance out of it, just like on OCS/AGA.

Amiga's RTG is what SVGA and VESA was to the PC. Nobody in their right mind would call VGA "not true to the spirit of the 486 PC".

Amiga would've probably ended up adopting standard PCI VGA cards anyway. It was the right thing to do in a much more standardized market. That's unless Commodore went full-proprietary hardware without expansion ports, just like a gaming console would do. That would certainly not go down well with big-box Amiga enthusiasts and power users, though, and it would risk alienating Amiga users that would keep their old expandable machines and splitting the Amiga market in the long run. Hey, that sounds familiar...

Originally Posted by nonarkitten

Gunnar HAD the choice to leverage either. Or even both. But he chose neither because it was not about making new opcodes for the 68K to improve performance, it was about...

Say it with me...

VENDOR LOCK-IN.

And remember, if you're in the Apollo Team and you dare to say anything critical against Gunnar's enlightened vision about the future of the Amiga (even if it makes more sense than what originally proposed by him)... You're FIRED!

[OlafSch]

I do not understand what people are complaining about. Software sells hardware, that is known for a long time. I remember people rightfully posting that there is no special software (expecially games) that really use the improved hardware and why buying it then. Now the team behind the project is doing the right thing and trying to get more games in the market that use the hardware. Nobody is forced to buy both hardware and software. Every developer is free to develop adapted software or not. That is called competition. Up to now all software (including PC ports) work on different hardware.

[jbenam]

Originally Posted by OlafSch

I do not understand what people are complaining about. Software sells hardware, that is known for a long time. I remember people rightfully posting that there is no special software (expecially games) that really use the improved hardware and why buying it then.

All of our discussions circle back on the reason *why* there isn't any software written especially for it.

Probably because our concerns (or "complaints" as you call them ) are shared by software developers, which opt to avoid vendor lock-in and instead keep developing for classic platforms without proprietary extensions, which could be enjoyed only by a very small amount of Amiga users.

[OlafSch]

@jbenam

I had some discussions with some of the current game developers. As I understand it most are more attracted by "classic designs" because it is motivation for them to get "magically" most of the current limited hardware. A little like the demo coders. That is a difference to the old days where amiga was one of the target platforms of new games. At that time developers would of course supported it because it makes life much easier and no need to invest lots of time in optimizations to get something running from disc on A500. Today people do development for fun, money is welcome but not priority. That certainly explains partly why there are no new amiga games adapted to Vampire. For real professional developers earning life with it platform is neither powerful enough nor big enough.

The Vampire team tries to overcome this partly now with intensifying own development. I see no problem there. Developers outside are not forced to do that too. I would more understand the discussion if (like apple for example) the apollo team would restrict what software runs on the platform.

[Thomas Richter]

I personally do not see any problem with extending the current hardware with features. That is all fine with me, I do not need to use it if I do not want it, and having more capabilities is certainly ok as it does not block anything.

What is certainly not ok is to cut old features and leave parts of the hardware platform out. Of course, Motorola did something similar with the 040 and 060, leaving out parts of the FPU to name one such omission, but they also took care of having the possibility of a software emulation of the missing parts, and *that* is really missing on the Vampires. So to say, I would not even mind having a restricted MMU all provided I could enhance or extend the current hardware by means of some additional software (for example, by software that does the MMU table walk and ATC filling which may be hard to implement on FPGA for sure, I admit), or say by having a software trap to emulate a full-precision FPU instead of the double-precision logic (which is also costly due to the requirement of needing a deeper or wider pipeline to fetch 12 bytes instead of just 8). This way, full precision might be slow, but it would still be possible.

That is, I am surely aware of *why* Gunnar left certain features out, but what I do not admire is his unwillingness of having at least the possibilities to emulate such features with software aid if there is a need for. This needs of course (as far as the MMU goes) somewhat more work than the current stupid flag-based "MPU", but I still believe that something is possible in this direction provided there is sufficient good will - which, frankly, I do not see. It is more a "take it or leave it" approach to the "customers".

Concerning the speculative development of the Amiga, I agree with OlafSch that it seems quite plausible that CBM would have picked some off-the-shelve VGA chip anyhow, and would have switched the bus to PCI. PCI offers logic that is quite similar to what autoconf on Zorro can do for you, and it even seems doable for me to mirror the PCI entries in expansion and have the PCI bus configuration through an enhanced version of expansion. Instead, we see a lot of proprietary half-working solutions on the market now (mediator, the P5 proprietary PCI bus on the CVisionPPC to name just two). This is certainly a loss. CBM surely discussed internally how to integrate RTG into the system, but external parties were faster and CBM could not get "their ass up" to get the thing going. Instead, they were busy with silly projects like the TIGA graphics card for Amix, which turned out to be a dead end.

At the time CBM died, the "vertical monopoly" of integrating chip manufacturing and software integration in one house stopped making sense, so the approach CBM took of running business broke down, and it would have been a rather radical change anyhow.

OlafSch, if you do not agree with Gunnar, you are not only fired from the team, but also threatened to be beatened up. And no, I am *not* kidding. I am serious.

[Rotzloeffel]

Originally Posted by Vamplover

Mods, why it’s okays for others to troll badmouths AMXXX and Gunnar’s here in this threads? I thoughts not allowed and gets peoples banned? Gunnar ships V4 wonder Amiga clones everyday to thousands the customers not vapor.

Because this thread is not for Fanboys like you ? Do you have any technical Skills to discuss anything in this thread? I do not think so

[OlafSch]

@Thomas Richter

I understand that you feel hurt by Gunnar by his posts, Even if you are emotional about what you do it should never get on such a personal level.

I fortunately do not have comparable experiences but I also do not try to persuade Gunnar to do something different, He is thinking different from how I read his posts... he sees Vampire not just as one amiga project next to others and part of a family of projects but as THE amiga future. So all he (and the team) concentrates on pushing Vampire sales. You can share this or not, I see it different too. For me 68k is niche anyway and Vampire is important but not exclusive there. And to get outside our bubble you need different offerings anayway.

68k in its current concepts cannot become mainstream (even if hardware would be powerful enough) because you have a different world now with different security requirements. And any amiga project is part of the platform and not exclusive project or only future. But that can everybody see different. At the moment compatiblity is very important for Vampire because of all the old software, if that changes in future who knows.

[Thomas Richter]

Originally Posted by OlafSch

I fortunately do not have comparable experiences but I also do not try to persuade Gunnar to do something different, He is thinking different from how I read his posts... he sees Vampire not just as one amiga project next to others and part of a family of projects but as THE amiga future.

And this is where it becomes outright silly. If you think about it, "retro" and "future" are contraction in terms. If you want a mainstream powerful architecture, 68K was and still is a dead end. If you want a lean orthogonal architecture today, "arm" is the answer, and Apple already demonstrated that. If you think "future", then "mainstream GPU solutions" and massively parallel GPU infrastructures are the answer. Not "SAGA".

Originally Posted by OlafSch

At the moment compatiblity is very important for Vampire because of all the old software, if that changes in future who knows.

I know. It will not. Why? Look at *why* people are using Amigas. Not because they want to play the latest blockbuster game. This is a retro market, not an emerging games console market. You need a different architecture for that. Why? Because history tells us. Its not that multiple ideas had been tried.

[jbenam]

Originally Posted by Thomas Richter

[...]
Concerning the speculative development of the Amiga, I agree with OlafSch that it seems quite plausible that CBM would have picked some off-the-shelve VGA chip anyhow, and would have switched the bus to PCI.
[...]
OlafSch, if you do not agree with Gunnar, you are not only fired from the team, but also threatened to be beatened up. And no, I am *not* kidding. I am serious.

I think you meant me, instead of OlafSch? I was the one raising these points

Anyway, I am not against technical improvements either, but they should be decided by a "Great Council of Amiga Wizards", not a single person. The community should decide how to bring the Amiga forward. Since Commodore is gone, and we are just a handful of people still doing this for fun, it makes sense that defining future extensions should be a democracy, not a dictatorship.

Originally Posted by Thomas Richter

And this is where it becomes outright silly. If you think about it, "retro" and "future" are contraction in terms. If you want a mainstream powerful architecture, 68K was and still is a dead end. If you want a lean orthogonal architecture today, "arm" is the answer, and Apple already demonstrated that. If you think "future", then "mainstream GPU solutions" and massively parallel GPU infrastructures are the answer. Not "SAGA".
[...]
I know. It will not. Why? Look at *why* people are using Amigas. Not because they want to play the latest blockbuster game. This is a retro market, not an emerging games console market. You need a different architecture for that. Why? Because history tells us. Its not that multiple ideas had been tried.

Amen to everything you have said.

I am still baffled how Gunnar *really* believes that there is some kind of future in 68k and the Amiga in general.

[TCD]

Originally Posted by jbenam

I am still baffled how Gunnar *really* believes that there is some kind of future in 68k and the Amiga in general.

If we believe the sales numbers there are enough people out there that 'want to believe' too. Of course it would be better if there was a "Great Council of Amiga Wizards", but the reality is that there are a lot of vultures fighting over scraps. I think everyone should be able to decide how to enjoy their Amiga experience. If that means playing games from the mid 90s for the first time and calling it 'NG' so be it

[OlafSch]

@jbenam

You are not the only person saying such things

Regarding future of 68k, it depends how you define it. As a niche and part of the retro platform it can grow of course. Niche can be big too in todays world. If you talk about "revival" and think it would play in same league as Windows or MacOS then it is not realistic

[dreadnought]

Originally Posted by Thomas Richter

If you think about it, "retro" and "future" are contraction in terms.

Not necessarily. You could have a platform which has roots in retro and expands into the future. So, for example, something >95% compatible with "old" Amiga and also capable of doing most of modern day common tasks such as email/web/media, running some sort of Amiga-related OS.


It's not for me, but I totally understand why some people can be interested in this angle. Is Vampire the answer for these folks? Maybe so, but the reason why it's still so divisive is for several reasons:


-the price
-the tech specs (exclusive features, closed source, etc)
-Gunnar/Team's behaviour


Now, we don't live in an ideal world and so some problems with such projects will always exist. If it was one, maybe even two of the above, I guess the whole proposition would be more palatable, but all three combined are the one big reason for these neverending debates.

[Promilus]

@jbenam - the main argument for AMMX is that ppl bough v2 and v4. But that is not a real argument. Ppl bought Phenom processors with SSE4A - did they buy that because SSE4A was so great? Ppl bought Bulldozer with XOP, CVT16 and FMA4 - did they because those ISA enhancements were so important to them? No. Not really. They did it because:
1. price/performance was ok-ish
2. they were fans of AMD products (or just didn't like intel more)
3. they were hoping that over time architecture might improve (which it did rather nicely but only on mobile, also that improvement was still not enough against intel)

So why did ppl buy V2 and V4? Well maybe because those were either the only solutions readily available to their platform (like A600) or only one combining fast ram, fast cpu, ide and rtg? Yeah... Fast and versatile. For some... promise of SAGA (on their ECS machines). AMMX didn't play any major role. But hey it's his playground and he can do whatever he wishes. But there's also a free speech. We can present our opinion as well, especially when arguments used have a merit. Right?

[jbenam]

Originally Posted by dreadnought

Not necessarily. You could have a platform which has roots in retro and expands into the future. So, for example, something >95% compatible with "old" Amiga and also capable of doing most of modern day common tasks such as email/web/media, running some sort of Amiga-related OS.


It's not for me, but I totally understand why some people can be interested in this angle.

Indeed, and AmigaOS4 did reach that lofty goal. It was a modern-ish OS with modern-ish applications. Still, it was adopted only by hardcore Amiga fans (you most certainly didn't see anyone who never had an Amiga before go and purchase an AOS4 system) and even so adoption was a fraction of total Amiga install base.

Anyway, we already had something akin to what the Vampire is trying to achieve and it didn't change much Amiga's position in the market. It still remained a niche system for a certain niche of users. Let's be honest, the ship for the Amiga as a popular system has sailed. In 1994. And nothing will bring it back.

The best part? It's still an awesome system to play with, even if it won't ever be again as popular as it was during its peak. And that's perfectly fine. But thinking that it will somehow make a comeback thanks to some very questionable improvements is just delusional.

Originally Posted by Promilus

So why did ppl buy V2 and V4? Well maybe because those were either the only solutions readily available to their platform (like A600) or only one combining fast ram, fast cpu, ide and rtg? Yeah... Fast and versatile. For some... promise of SAGA (on their ECS machines). AMMX didn't play any major role. But hey it's his playground and he can do whatever he wishes. But there's also a free speech. We can present our opinion as well, especially when arguments used have a merit. Right?

Right Yes, I was very interested in the chance of having full AGA compatibility, a fast CPU, lots of RAM and RTG on my A600. Accelerators for it weren't readily available or very expensive. And yes, I thought AMMX was a stupid idea even when I was in peak hype for the V2.

That's why I put myself in the waiting list for a V2. Somehow I was never contacted and seeing the entire mess with the "illegal" cards and how the V2 series was put aside in favour of the V4, made me very happy indeed that I was never offered the chance to purchase a V2.