@JimIgou

PPC is as dead as 68k was years ago. On desktop 96% X86, 1% ARM. On embedded or smartphones ARM dominates today. Except IBM Servers nonbody uses PPC anymore or invests in it. What are you proposing? ;) A-eon contacts IBM regarding a future cooperation? ;). It is a little funny to discuss that at all. And yes FPGA implementation of 68k is a hobby today, even though relative successful in amiga terms. But of course it is a niche product for amiga fans and perhaps some technic enthusiasts, not for general public. A little like Raspberry f.e., just more expensive

@kolla

Actually, there are real 68000 and 68020 cpus available.
FPGAs work in this application since the processors emulated are so dated that the performance is acceptable in an FPGA implementation.

That is why you don't see X64 or PPC FPGA designs.

I'm glad to see some development on the 68K hobbyist market, but the promises and statements being made by the advocates of this project get a bit outlandish.

Some good things Gunner has done?
First, reducing the number of cycles needed to execute some common instructions.
Second, producing a cpu that has a high degree of legacy compatibility (the '040 and '060 have always presented issues for us).

The outright dumb things?
FPU support, why drag this issue out, it should have been implemented long ago.
Comparisons to more modern cpus.
Simply stated, Vampire will never reach the performance levels that the more powerful X64 or PPC NG systems already have.
Although it should be able to best weak systems like the Efika.

And it WILL be one heck of an Amiga though.
And that's the point, isn't it?

@JimIgou

What are you comparing now? ;)

Gunnar and one or two other developers versus Intel/AMD? ;)

And Intel/AMD wins? Surprise surprise

Be serious now

It is a niche product and will stay it. And even if they managed to create superior hardware there is zero chance to compete on mainstream. But there is trend to retro, there are enthusiasts interested in alternative hardware and/or OS. Most obvious example is raspberry. And regarding amiga, there is still some potential. The amiga facebook group has about 20.000 members. Of course not everyone is willing to buy new amiga hardware but it is still a couple of thousands. More from outside are certainly interested if f.e. standalone hardware is available and perhaps some technic enthusiasts without former amiga experience might be interested too. But we still talk about something in the thousands, not millions like in mainstream world. Looking at the big players is simply silly. At least Gunnar has not only promised but also delivered, in opposite to many others. Some like his ideas, some not. Who not likes it should stay away...

@olaf

No surprise, I'm just glad to see a few, like you, have realistic expectations.

Vampire makes one hell of a nice Amiga enhancement.

But there's no competition here, because we all started out with legacy hardware.
And I still use legacy hardware to run some software (in particular games).

NG and legacy communities serve different needs anyway.

In regard to what's realistic, legacy supporters might have an edge, as NG attempts to resurect and build on something with very limited resources.

My only point has been, I don't have a problem supporting multiple Amiga related projects.

In fact, the availability of multiple products makes me quite happy.
No one approach to rule them all, rather a diversity of approaches.

As to "Who not likes if should stay away..." that's a weak response from those who can't handle criticism.
Almost as weak as the argument to give away copies of OS3.9 because it's not supported anymore.

Then again, considering the current state of affairs with both Hyperion and Cloanto claiming the right to vending OS3.X...
I'm not sure what the right approach is for users of new legacy hardware.

Last edited by JimIgou on 06-Sep-2018 at 07:11 PM.
Last edited by JimIgou on 06-Sep-2018 at 07:01 PM.
Last edited by JimIgou on 06-Sep-2018 at 07:01 PM.

Quote:

JimIgou wrote: There are real 68000 and 68020 cpus if you want them.

The 68020 and 68030 are getting more difficult to source in any kind of quantity. A compatible CPU core in a low end FPGA can simulate them at a similar performance per clock and often at a higher clock speed than the original CPUs.

Quote:

And I appreciate the ability to recreate older cpu designs via FPGA. Of course, the reason this works so well is that these older cpus are limited in performance. That's why the Vampire can provide a performance boost, it's emulating a limited processor.

The Apollo core design most resembles a 68060. The 68060 design resembles how an in-order CISC CPU design would be created today. The biggest limitations in performance compared to today's CPUs are due to the old die size and small area (transistor) requirements.

68060 2,500,000 transistors
ARM Cortex-A9 (32 bit) 26,000,000 transistors
Atom (32 bit) 47,000,000 transistors

You could have many times the number of 68060 cores as a Cortex-A9 or Atom CPU for the same transistor count. It is easy to add multiple cores but what is needed today is strong single core performance without resorting to high clock speeds. The RISC advantages were supposed to be the following.

1) simplify the pipeline to increase clock speeds
+ simplifies CPU design
- produced more heat than expected
a) resulted in multi-core CPUs
- increases the number of instructions and decreases code density
a) resulted in poor single core performance as more instructions needed to be executed
b) resulted in increased instruction fetch requirements, cache requirements and memory traffic
c) resulted in larger and more costly register files to reduce the problems of a and b.
2) move CPU complexity into the compiler
+ simplifies CPU design
- the compiler lacks knowledge of what the CPU and often programmer is doing
a) resulted in poor CPU performance
- compiler design becomes more complex
a) resulted in RISC CPUs needing better compiler support

RISC has been unable to solve the performance problems from a flawed philosophy (it was obvious that #1 above was a mistake when the Alpha put DEC out of business). CISC has superior performance characteristics. If you don't believe me, then it is time to educate yourself. See the comparison of architecture types from a university professor.

http://cs.uccs.edu/~cs520/S99ch2.PDF

Quote:

Which makes the hyperbole a bit painful. Vampire, any FPGA for that matter, will never equal the performance of a later X64 or PPC processor.

It is unlikely that a 68k CPU would ever match the performance of a modern x86_64 CPU. Modern high end high performance CPUs are now in the billions of transistors (some 1000 times the transistor count of the 68060) with huge budgets. It may be possible for the 68k to achieve the PPC level of performance. Current PPC designs do not compare to modern x86_64 CPUs or POWER CPUs in technology, budget or transistors. PPC designs stopped trying to push the performance envelope long ago and use older efficient designs with some modern enhancements. Most ARM CPUs are trying to strike a balance between performance and energy efficiency as well. These are much cheaper CPUs to design yet are the majority of CPUs sold. This is where I believe the 68k would be interesting.

I wouldn't object to using the POWER architecture for the Amiga high end. I do have doubts about POWER Amigas expanding the Amiga user base which is needed for more development and software. POWER prices have come down but they are above PPC hardware prices which have not sold well. Another concern is the POWER focus on efficient multi-threading (POWER RISC lacks the single core performance of x86_64 CISC) where the AmigaOS doesn't have a concept of threads like Linux/BSD. Multi-threading is less efficient when unrelated tasks/processes share the same core. Multi-threading can result in some tasks/processes taking longer to execute which can make the lack of single core performance more visible. I would rather have more cores than multi-threading for the Amiga. Multi-threading is good for servers but the AmigaOS lacks the security to be competitive. The best chance of adding security while keeping compatibility is with a custom hardware design.

Quote:

Hey Greg, this isn't trolling, it's reality calling. Vampire make one he'll of a nice legacy Amiga upgrade. And I for one like some of the things Gunner has added. The reduction in cycles needed to execute some instructions is very nice.

As CISC instruction and addressing mode cycles are reduced, the more of an advantage CISC has over RISC. PPC was supposed to replace x86/x86_64 but the tables turned with it outperforming PPC and even POWER, especially in single core performance. It is not just the economies of scale but also performance advantages. The 68k has some of the same advantages as the x86_64 while likely being more energy efficient.

Quote:

I look forward to using a V4, but it won't be replacing my NG hardware. What it will do, in combination with the older Vampire accelerators is give us a large enough base of users with high power Amiga systems to make developing for the market attractive again. And that is cool enough. And the latest Vampire looks powerful enough to handle a port of OS4. So hey, Herman's, want a market that looks larger than your current one?

You do agree that 68k hardware can expand the Amiga user base. Do you think this amount of expansion is relative to the amount of 68k hardware sold? Do you think the amount of hardware sold is based on supply and demand where a lower price and better value would result in greater demand? Do you think this is a good reason to try to create 68k hardware with a very good price and value?

@JimIgou

Quote:

Actually, there are real 68000 and 68020 cpus available.

68000 variants yes, but 68020?
I wish - do you have a source?
020+FPGA would be nice for a low-end Minimig-AGA clone :)

Quote:

That is why you don't see X64 or PPC FPGA designs.

Really? I am quite confident that I have seen PPC-400 cores used in networking appliances (as well FPGAs with "real" PPC-400 CPU cores attached).

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

@matthey

Quote:

You do agree that 68k hardware can expand the Amiga user base.

I don't - lack of 68k or even price of 68k hardware was never the reason why Amiga user base shrank - the problem was always limitations in the OS and the software, in other words - lack of developers willing to bring the platform forward. As well as a legal nightmare. Other platforms simply offer better, cheaper, safer and faster ways of writing software and "get jobs done", generate revenue etc.

Last edited by kolla on 07-Sep-2018 at 04:28 AM.

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

@matthey

Quote:

...AmigaOS doesn't have a concept of threads like Linux/BSD.

Looking at it another way, all AmigaOS really has is threads, since every process shares the same address space. What it really lacks is the Linux/BSD concept of isolated processes.

_________________
"Unix is supposed to fix that." -- Jay Miner

@kolla

Yes, there where limitations in the OS but so was there in other OS that time. If they had continiued development they for sure would be as advanced as other OS.

The biggest problem was that the hardware never got updated.

Yes, Phase 5 and others made accellerator cards others did 16 bit soundcards etc. but by then it was more for the people with special interest.

Normal people don't care and don't even know what memory protection is. It is about having a mashine that you can buy ready from the shop and has a easy userinterface with software, games that people would like to have.

Normal people don't even know what a cpu is. At least not if it is 68k, ppc or whatever.

The standalone Vampire or the Raspberry Pi AROS can for sure get a growing user base done right.
You marked it as a gaming mashine for people that like pixel gfx and retro gaming.
Looking at the new great game "Worthy" you could even maybe make a deal with them and bundle the hardware with it and maybe other games. Just put the game on a usb stick and you boot right from that just like in the old days with disk, cd, cartridge or whatever. If you are more advanced you could install it to the internal drive too :D
PS4 has no web-browser. It is not like the Vampire or PI should nessesary have that either. Web-browser you have on your telephone that you have with you everywhere.

I promise you bring one great game to that platform and if the price is right it will sell massive numbers.
It should be possible with Raspberry PI and AROS to sell for under 100 Euro.
Yes, I know. Pi is not 68k but maybe with mass production and some heavy investors the vampire could be quite cheap too.

Last edited by nikosidis on 07-Sep-2018 at 09:51 AM.
Last edited by nikosidis on 07-Sep-2018 at 09:49 AM.
Last edited by nikosidis on 07-Sep-2018 at 09:47 AM.
Last edited by nikosidis on 07-Sep-2018 at 09:45 AM.
Last edited by nikosidis on 07-Sep-2018 at 09:41 AM.

@bison

Quote:

Quote: ...AmigaOS doesn't have a concept of threads like Linux/BSD. Looking at it another way, all AmigaOS really has is threads, since every process shares the same address space. What it really lacks is the Linux/BSD concept of isolated processes.

Exactly - AmigaOS an operating system implemented as one single process - exec.

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

@matthey

[QUOTE]68060 2,500,000 transistors
ARM Cortex-A9 (32 bit) 26,000,000 transistors
Atom (32 bit) 47,000,000 transistors[/QUOTE]

Yes, but I think that the biggest difference in transistor count between these CPUs is in far, far larger cache memory on newer CPUs.
The first Atom was basically an updated Pentium which had a comparable transistor count to the 060 - which is a good indication of the hypothethical "modern" 060 revival and how it would perform(minus the non pipelined FPU on the 060)

Last edited by WolfToTheMoon on 07-Sep-2018 at 11:50 AM.

_________________

@nikosidis

Quote:

Yes, there where limitations in the OS but so was there in other OS that time. If they had continiued development they for sure would be as advanced as other OS.

All the other comparable operating systems of that time are gone, or replaced with (for lack of a better word) "real" operating systems. AmigaOS would have had to be replaced too, and that was also Commodore developers' plan forward, sadly with WindowsNT, on Hombre/PA-RISC.

Quote:

The biggest problem was that the hardware never got updated Yes, Phase 5 and others made accellerator cards others did 16 bit soundcards etc. but by then it was more for the people with special interest.

So your argument is that hardware never got updated, but then that it was updated, but only people with special interest bought into it.

Maybe because Amiga is special interest in the first place?

Quote:

Normal people don't care and don't even know what memory protection is. It is about having a mashine that you can buy ready from the shop and has a easy userinterface with software, games that people would like to have. Normal people don't even know what a cpu is. At least not if it is 68k, ppc or whatever.

Which is why normal people use main-stream operating systems and software from experienced vendors who are capable of providing secure systems and services.

Quote:

The standalone Vampire or the Raspberry Pi AROS can for sure get a growing user base done right.

The sheer number of users is meaningless (unless your business is "support"-)
What counts is number of developers. You cannot have a growing user base without a growing developer base, that can support the demands of the users.

Quote:

You marked it as a gaming mashine for people that like pixel gfx and retro gaming. Looking at the new great game "Worthy" you could even maybe make a deal with them and bundle the hardware with it and maybe other games. Just put the game on a usb stick and you boot right from that just like in the old days with disk, cd, cartridge or whatever. If you are more advanced you could install it to the internal drive too :D

Normal people do not really have time to sit down in front of a TV or monitor to play games, that is a "special interest". Normal people play games on their phones, and are more than happy to keep it that way.

Quote:

PS4 has no web-browser. It is not like the Vampire or PI should nessesary have that either. Web-browser you have on your telephone that you have with you everywhere.

If there is one thing I have learnt over the last two decades, it is that Amiga users are helpless without two thing - a web browser, and an office suite. :)

Quote:

I promise you bring one great game to that platform and if the price is right it will sell massive numbers.

Then you have to come up with something new and ground breaking, good luck.

Quote:

It should be possible with Raspberry PI and AROS to sell for under 100 Euro.

Huh, you can have Raspberry Pi and AROS for 10 euros already, what are you talking about?

Quote:

Yes, I know. Pi is not 68k but maybe with mass production and some heavy investors the vampire could be quite cheap too.

Not really, since the main component (the FPGA) is already priced quite high.

No-one is willing to invest into an architecture that is tied to a platform, which main activities over the last decades, have taken place in court rooms.

Apollo Core could have become interesting if they (he) had chosen to _not_ focus on Amiga, but instead focus on offering a pathway for all legacy m68k platforms. The way the project has progressed, it not only has lost the interest of developers of just about all other 68k platforms, it has even alienated half of the Amiga developers as well.

Last edited by kolla on 07-Sep-2018 at 12:27 PM.

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

@kolla

regarding Aros on Raspberry... no you cannot

Aros only runs hosted up to now, there was some work on a native version that stopped and Michael is now starting with it again

And regarding developers and users, it is the chicken and egg problem, you need both to be successful

But you cannot create interest at developers if there is no (potential) userbase and hope for growth. So a growing userbase offers the chance to attract new developers too

Amiga developers alienated? How many? One? Two? Three? ;)

what software are those developing? Where are the updates for "normal" 68k? I can remember one writing Vampire not for him and his fellows that are active in demo programming. Vampire (and 68k in general) needs new applications, not demos. If a game developer makes something special is up to him or her but you can develop a game that makes use of the fast processor without making a special vampire version. Or you offer higher resolutions if vampire is detected. Or a real dedicated game but then again it is a question of the userbase and thus potential buyers. That is not a question of liking or not liking gunnar but professional view...

Either you invest energy because you see potential for your software or you leave it. How many people developing software for windows hating microsoft?

So if Gunnar would have been nice to you and others (who were not very nice too sometimes...) you and others would deliver plenty of new vampire adapted software? Somehow I do not believe that...

Last edited by OlafS25 on 07-Sep-2018 at 12:54 PM.
Last edited by OlafS25 on 07-Sep-2018 at 12:47 PM.
Last edited by OlafS25 on 07-Sep-2018 at 12:46 PM.
Last edited by OlafS25 on 07-Sep-2018 at 12:46 PM.

@kolla

You are a funny guy. You quote everything people are saying and disagree about everything said.

I'm not going to reply to all this. I don't understand what is your interest here at amigaworld anyway. You are so negative about everything amiga releated. People here have a great hobby and we like Amiga and what it is and are very happy and exited about future development and what is going on.
Vampire already have over 1000 followers on facebook. That is without releasing any standalone system. Just like the accellerator cards for the original amiga the vampire for A600 with all the problems for monting is for people with special interest. The vampire standalone version is something that will sell a lot. It will be something working out of the box. Something you could develop for. It is so much more easy for users and developers if something is made for specific hardware like lets say Super Nintendo or Amiga1200. Maybe you don't understand the difference but it is a big difference.
There is a companies doing games for A500 and A1200 in 2018 like Pixelglass. With the Vampire there will be a market for game or software developers. I'm pretty sure 1 year after the release it will be at least 5000 units sold.
You don't have time you say to plug it into your TV, monitor, but you have time sitting here all day with negative comments. Funny!

Last edited by nikosidis on 07-Sep-2018 at 12:52 PM.

@OlafS25

Quote:

regarding Aros on Raspberry... no you cannot Aros only runs hosted up to now

Which is how it should be, and the cost for a Raspberry Pi zero is....?

Quote:

And regarding developers and users, it is the chicken and egg problem, you need both to be successful

Not really, lots of hardware, software and operating systems are successful without any users.

Quote:

But you cannot create interest at developers if there is no (potential) userbase and hope for growth.

Yes you can.

Developers in general do not care about the number of users of their software, they care about solving problems, keeping customers happy, and having a steady income from their work. As a developer it is more lucrative to have one big customer who pays grandly, than to have tens of thousands of (ungrateful) users.

Quote:

So a growing userbase offers the chance to attract new developers too

In theory, given that the platform is open, documentation is easily at hand, developer tools are easily available etc. In case of Amiga though, we are again talking about a special interest thing, software developers in general shun everything that Amiga systems are about :)

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

@kolla

Haha again.
I stop reading your posts.
You have an answer to everything don't you.

Last edited by nikosidis on 07-Sep-2018 at 12:55 PM.

@kolla

one big customer on amiga? ;) stay seriously

developers do it for their own fun but also to get positive reactions by users and of course in normal world to get income. I would say the last one is the most important one, positive reactions are incentives if you do something without asking for money like freeware or open source. Doing something only for fun not really works longterm because somewhen priorities change and development stops. Web is full of projects of that kind. The biggest chance to attract developers is a growing userbase with high future potential, nothing else.

Not if you would drink a beer with the main hardware developers or not. That is just the thinking of hobbyists... and unprofessional

Last edited by OlafS25 on 07-Sep-2018 at 01:01 PM.

@OlafS25

Should we make up our minds about whether we are talking about development within the realms of Amiga, or within the realms of what is the topic - 68k development?

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

@kolla

this is a amiga forum

you want to talk about development on atari? ;)

The Vampire will have a huge following audience of people from 40 - 70.
In that age most people have money.
Whatever released most will buy.

This platform is in no way interesting as a NG system.
It is a retro platform to play games and have fun with.
Be creative do some chip tunes make cool pixel gfx.
The demosceners will pick this up and do something never
seen before on 68k. I'm sure about it.

I think it is cool with everything Amiga 68k related. I totaly love to see sceners productions for the Amiga platform that is still very much alive.

Last edited by nikosidis on 07-Sep-2018 at 01:09 PM.
Last edited by nikosidis on 07-Sep-2018 at 01:08 PM.
Last edited by nikosidis on 07-Sep-2018 at 01:06 PM.