ADOOM on A600 running 22-35 FPS

[kolla]

I bet at least 10% of Pi sales have gone to people that would prefer to have an Amiga computer or 68k CPU but only Raspberry Pi was offered instead of Cherry Pi. Let's say only 3% would have payed up for the Amiga or 68k which would be 150,000 potential Amiga users.

I would say this is a huge over estimate, huge huge HUGE over estimate!!!

It would surprise me a lot if a 68k Cherry pi would sell much at all, probably less than 5000. At least without a full fledge capable CPU with MMU, so that it easily can boot a relevant operating system.

I think you have misunderstood why people buy the raspberry pi, it's not just because it is cheap, it is because it is cheap and fully supported bu Linux. Extremely few are interested in a lamed down EC 68k that only can run AmigaOS and maybe uCLinux if you are lucky.

[Lurch]

I could give you this for $200 today easily.

Sold, you make it I'll buy it.

[QuikSanz]

If it can replace my CS3 060 on my A4Kt with all that it has, I'm sold. Bring it on!

[matthey]

you compare apples and oranges.
The current FPGA boards for AMIGA have quite  high prices for what they offer.
That these board have so high prices DOES NOT mean that an FPGA is more expensive than an ASIC.

An FPGA board with 256 MB fastmem, and 68020 CPU with 600 Mhz performance and SAGA could today be sold for less than $100. If a few thousand would be produced.

How much could the Natami MX board have sold for if a few thousand would have been produced?

How much would the cost increase to have 1GB of memory?

Your specs above and estimated per unit board price may win the apples vs oranges battle using an FPGA and economies of scale but you may only win a few hundred new Amiga users than the existing FPGA boards. That's not bad and it is cheap enough to give as Christmas presents at that point but does it win the bigger battle. We could win thousands and maybe tens of thousands of Amiga users with a product that could grab a potion of the Raspberry Pi pie. There are people who like Cherry Pie but if all they get is a little piece when they could have a big piece of Raspberry Pi for less cost, which are they going to choose? Your specs above are great for the Amiga market but not very competitive against the Raspberry Pi. I'm not saying that your board with an FPGA would be wrong for a first round of production but if gaining lots of Amiga users is the goal then planning, cooperating and setting standards toward a possible ASIC design is a worthwhile strategy IMO.

why not make a crowfunding campaign in kickstarter?

We all wondered why this didn't happen with the Natami where it would have been huge with all the interest and momentum. The down sides are that you are obligated and have limited room to manuever from the advertised plan with kickstarter. I still like it as a grass roots way to raise production money from the little investors and customers. Less invested money is lost to fees if medium size to large investors would come together and do the planning ahead of time.

Of course not, ASICs, even on an older process (90/65nm) are horrendously expensive for the potential market.

ASICs are the cheapest option with high enough production quantity. Making a few thousand FPGA boards could cost $100,000-$200,000. For double that, maybe less with the right partners, you could be in the range of an ASIC that could lower per unit board costs by maybe 25%-50% and increase performance by several times. Add 1 GB of memory and you are delivering a full slice of Cherry Pi. It's risky but it is a better plan to add Amiga users than Hyperion ever had. I bet they spent millions on software development only to gain a few hundred Amiga users.

FPGAs are more than enough - a ~600Mhz 68020 equivalent is more than enough for any 68k Amiga software that's ever been written.

Is 640kB of memory enough too? The 68k has huge potential that was never developed in performance, code density and ease of use. It can blow in-order Thumb2 ARM out of the water in performance while using less memory and being easier to develop software for. I would love to throw 1GB of memory on board just because it is cheap and so overkill for a 68k Amiga. Such a board wouldn't be a Raspberry Pi killer but it could be a competitor for people who prefer Cherry Pi.

I would say this is a huge over estimate, huge huge HUGE over estimate!!!

It would surprise me a lot if a 68k Cherry pi would sell much at all, probably less than 5000. At least without a full fledge capable CPU with MMU, so that it easily can boot a relevant operating system.

I think you have misunderstood why people buy the raspberry pi, it's not just because it is cheap, it is because it is cheap and fully supported by Linux. Extremely few are interested in a lamed down EC 68k that only can run AmigaOS and maybe uCLinux if you are lucky.

What I'm guessing is that at least 10% of Raspberry Pi purchasers know and have a favorable opinion of either the Amiga or 68k (the 68k may be more popular than the Amiga). These are potential converts with a competive product (even with a somewhat inferior product overall). Such a product would need to be open, compatible and use standards where possible. Improvements need to be made in the CPU, with AmigaOS/AROS and with marketing/image to bring everything up to snuff. There are already developers willing to do the labor of love but some inflow of money could buy more of their time thereby accelerating development.

[biggun]

an ASIC that could lower per unit board costs by maybe 25%-50%

What makes you think this?
What do you think that an FPGA to build something comparable to ARCADE or MIST does cost?
What do you think would an ASIC cost?

and increase performance by several times.  

Yes this is true.
An ASIC would in theory allow huge performance gain.
But to be able to build an ASIC the design needs to be 100% tested and you need to be 100% sure that you have all good ideas implemented. Also it needs to be clear that doing an ASIC release includes porting to ASIC libraries. This wil also be some work and take weeks if not month.


Before looking at an ASIC I would first try to get the max out of FPGA options.

With todays Vampire600 you get 200 MHz 68020 speed.
With the new ApolloCard we think to get around 600 MHz 68020 speed
With todays better FPGA you could get over double of this. E.g 1200 MHz 68020 Speed.

The new FPGA generation which is on the horizon looks impressive.
My first impression is that performance could again be double t´with them....

I would like to see how good the next gen low cost FPGA will be.
If we could get out of a $15 FPGA 1200 MHz 68020 Speed - then I do not think that an ASIC is really needed for World domination ...

[TheDaddy]

Awesome!

[wawrzon]

If we could get out of a $15 FPGA 1200 MHz 68020 Speed - then I do not think that an ASIC is really needed for World domination ...

especially an asic once done would age quickly. there would be a window where you can do something with it but it might be narrow. fpga core as i understand it is to an extent reusable in newer generation fpgas.

[kolla]

What I'm guessing is that at least 10% of Raspberry Pi purchasers know and have a favorable opinion of either the Amiga or 68k (the 68k may be more popular than the Amiga).

Who are these 10%? You are pulling numbers out of nowhere. Again, people buy R-pi because it has HDMI, is cheap, and runs Linux, fully supported. The company I work for bought a whole bunch of r-pi to plug into big monitors that we use as info displays, the R-pi is perfect as it is tiny, powered by the TV and has HDMI, and because it runs full fledge Linux.

[kolla]

May be of interest:
http://www.fleasystems.com/

[matthey]

What makes you think this?
What do you think that an FPGA to build something comparable to ARCADE or MIST does cost?
What do you think would an ASIC cost?

It is difficult to come up with good estimates because of hidden costs. An ASIC has more variables and hidden costs. The actual unit cost is based on quantity and demand is difficult to estimate. There may not be that much difference in per unit cost between an FPGA board and an ASIC board but the performance/price goes way up with the ASIC and it's this which is probably needed to win many new Amiga customers and compete with a Raspberry Pi (but not necessary for the niche retro Amiga market). A properly enhanced hard 68k processor would help the 68k to be taken seriously again opening up markets and improving development support. IMO, a more compatible and conservative integer and FPU ISA with a modernized MMU/MPU and probably SIMD unit would be ideal. Modern and compatible sounds easy but it can be more difficult than creating something new.

Yes this is true.
An ASIC would in theory allow huge performance gain.
But to be able to build an ASIC the design needs to be 100% tested and you need to be 100% sure that you have all good ideas implemented. Also it needs to be clear that doing an ASIC release includes porting to ASIC libraries. This wil also be some work and take weeks if not month.

Before looking at an ASIC I would first try to get the max out of FPGA options.

I expect an ASIC needs early planning and experience to organize properly, debug and verify in preparation for an eventual ASIC. That is why I approached Dave@Innovasic who not only understands the embedded market but has experience with a 68k ASIC and may have tools and test cases which could accelerate development in general and ease the work needed to create an ASIC later. He is very creative and I believe would have some good ideas. I think he is somewhat interested as Phoenix shows that you guys are not amateurs with the hardware but there are a lot of loose ends in the business and financing side which are also important. I like Innovasic and Dave but there are also other potential partners which could help with development and marketing resources.

With todays Vampire600 you get 200 MHz 68020 speed.
With the new ApolloCard we think to get around 600 MHz 68020 speed
With todays better FPGA you could get over double of this. E.g 1200 MHz 68020 Speed.

The new FPGA generation which is on the horizon looks impressive.
My first impression is that performance could again be double t´with them....

I would like to see how good the next gen low cost FPGA will be.
If we could get out of a $15 FPGA 1200 MHz 68020 Speed - then I do not think that an ASIC is really needed for World domination ...

This is more of a philosophical quandary. Will people be satisfied with a slower but adequate performance CPU and more efficient OS? I think history shows the answer so far is clearly no. The majority of people would like to have the highest performance processor possible or the most performance/price as affordability becomes a factor.

Who are these 10%? You are pulling numbers out of nowhere. Again, people buy R-pi because it has HDMI, is cheap, and runs Linux, fully supported. The company I work for bought a whole bunch of r-pi to plug into big monitors that we use as info displays, the R-pi is perfect as it is tiny, powered by the TV and has HDMI, and because it runs full fledge Linux.

Haven't you heard that 97.3% of statistics were pulled out of thin air or inaccurate? Of course it's a guess. Please poll and calculate unbiased statistics and get back with me. Once we know the exact demand we can calculate the number of boards to make and investors will trust your results so much that they will pull their money out of no risk negative interest rate investments and invest in a sure thing like the Amiga .

May be of interest:
http://www.fleasystems.com/

I wandered across that site several years ago as I was search for something Amiga related. I was surprised when my search brought up some Dave Haynie quote like this:

NEVER EVER mess with a PCB jumper you don't understand, even if it's labelled "SEX AND FREE BEER".

I couldn't find it anymore so maybe they took it down.

[kolla]

You must first tell me what the poll is supposed to be about, without MMU it will have extremely limited use. With a proper implementation that is compatible and has MMU you could get it to run NeXTStep, Plan9, MiNT, Linux, NetBSD, OpenBSD, SunOS etc which people would actually be interested in, especially if it was mindbogglingly fast. Without MMU, there are already plenty of hardware alternatives, and none of them very popular.

So what is your idea?

EDIT: I see now that Gunnar says Apollo has MMU, which makes it a lot more interesting.

[matthey]

You must first tell me what the poll is supposed to be about, without MMU it will have extremely limited use. With a proper implementation that is compatible and has MMU you could get it to run NeXTStep, Plan9, MiNT, Linux, NetBSD, OpenBSD, SunOS etc which people would actually be interested in, especially if it was mindbogglingly fast. Without MMU, there are already plenty of hardware alternatives, and none of them very popular.

So what is your idea?

EDIT: I see now that Gunnar says Apollo has MMU, which makes it a lot more interesting.

The MMU is one of the areas we need to talk about and work out. Gunnar's MMU is very simple and incompatible with the 68k MMU in the 68030-68060. The 68030 MMU is very flexible and has some interesting ideas but is far from modern MMUs. The 68040 and 68060 MMUs are similar and very capable but the page sizes only go up to 8kB which is small and not optimum for modern large memory sizes. Enlarging the page sizes would break compatibility with the 68040 and 68060 MMU and there may be other changes necessary to modernize it. ThoR or Gunnar could probably give more details as the MMU is not my strong point. The 68060 MMU allows virtualization (virtual addresses and physical to virtual translation) which is a great modern feature but it has a cost. All cached addresses which may have a virtual address need to be flushed on context switches which has overhead. The 68060 may require this already with it's branch cache. Memory performance can degrade with many page misses from accessing memory in a dispersed or random way. More hardware can make up for this problem but at a cost. The CPU pipeline can increase by 1 stage depending on how the MMU is designed also. Gunnar and Jens know the best way to make a modern MMU and ThoR could help with the design. Gunnar is resistent to a more advanced MMU with virtual addresses in an FPGA because of the cost of performance and because the Amiga doesn't need it. The overhead of such an MMU as a percentage of performance is much smaller in an ASIC and the advantages for an open platform 68k board which would support multiple OSs is more compelling. This still leaves the question of how compatible an updated and modernized 68k MMU could be. Some changes to existing 68060 MMU support in the OSs which support the 68k would be necessary for optimal performance at least and maybe even to be usable. Personally, I think an MMU which supports virtual addresses is an important modern feature which would be important if competing against the Raspberry Pi with an ASIC. Embedded applications use memory protection more and more all the time but I expect the extra performance to be more important than virtual addressing.

As for a poll, the questions might be something like this:

Do you know about the AmigaOS, AROS or MorphOS?
If yes, what is your opinion of AmigaOS, AROS or MorphOS?
 1) Very favorable
 2) Somewhat favorable
 3) Indifferent or no opinion
 4) Somewhat unfavorable
 5) Very unfavorable
Do you know about the 68k CPU?
If yes, what is your opinion of a more modern 68k CPU with MMU, FPU and SIMD unit?
 1) Very favorable
 2) Somewhat favorable
 3) Indifferent or no opinion
 4) Somewhat unfavorable
 5) Very unfavorable

Only Raspberry Pi owners could be asked and even the wording of the questions could introduce bias. Data from international regions would probably vary (Asia likely would not be good for the Amiga or 68k). The results would be interesting although I wouldn't put to much confidence in them. My point was that a fraction of the Raspberry Pi market is a huge number of users (1%=50,000 users currently and growing). Hyperion likely spent over a million and maybe millions of U.S. dollars to gain a few hundred users which are not enough to sustain them with software or hardware purchases. It's likely that we could bake ourselves a cherry Pi and grab a slice of a bigger pie for less cost.

[Lurch]

Being an 020 CPU do you need a special 060 library to fool some programs that will only run when an 060 is detected?

[biggun]

The 68030 MMU is very flexible and has some interesting ideas but is far from modern MMUs. The 68040 and 68060 MMUs are similar and very capable but the page sizes only go up to 8kB which is small and not optimum for modern large memory sizes. Enlarging the page sizes would break compatibility with the 68040 and 68060 MMU and there may be other changes necessary to modernize it.

Matt is 100% correct here.

The 030 and 040 MMUs are first of all - incompatible.
From a todays view point both are not perfect and you would want to improve them with a new MMU design.

And of course your "ideal" MMU very much depends on what you want to do.
The "ideal" MMU for Linux looks different than the "ideal" MMU for AMIGA OS would look like.

I can not image that someone will seriously want to run Linux on 68k
But of course the beauty of the FPGA is that - one could create in theory 2 MMU designs -
one tuned for AMIGA OS and one tuned for LINUX And the people can pick their Core version for runnig in their FPGA.

[kolla]

There are plenty of us running Linux on m68k, and m68k is well supported by the Linux kernel and the GNU tool chain, in parallel with ColdFire. If you manage to provide a modern and fast, and well documented 68k CPU, you will quickly find Linux running on it, as well as NetBSD.

[kolla]

For sake of compatibility for those who want hardware to run old m68k OSes (NeXTStep, Plan9, A\UX, MacOS...) I think compatibility with 68040 is favourable. It may be worthwhile to check with people over at http://www.nextcomputers.org/ if they would be interested.