Open sourced Pegasus re-implementation?

Assumptions:

1) Open Sourced Pegasos design usable by non-Genesi entity for re-implementation.
2) People would buy it.

Rough Ideas:
MPC7447a (on motherboard, no CPU card)
Upgrade to Marvell Discover VI controller.
DDR2 RAM (533mhz). 2 Slots.
PCI-e implementation
PCI implementation ???
SATA II
USB 2
Gigabit Ethernet
AC97 Audio
PS2 keys/mouse
RS232 ???

Key factors of design are:
retain performance aspects of original Peg,

reduce cost by elimination of CPU card feature and older legacy device support (parallel, joystick, firewire, floppy controller),

and modernization through PCI-e (even if only X4) - SATA - and USB2.

I know this isn't the greatest idea but thought I would throw it out for comment anyway.

better not the greatest idea, instead of no idea at all ... at the moment I am still wondering what the future of PPC is ?

What you are describing is a completely new motherboard.

Why are you talking about the Pegasos?!

Yes a new board "based" on Pegasos, but not actual clone. Such board would use as many design aspects as possible from Peg but deviate in the ways I described.

The critical thing with original Pegasos, in my opinion, is an already optimized MC7447 - Marvell 2 interface.

I loosely spec the Marvell 6 because it is the easiest way to obtain modern device support (like SATA, PCI-e and USB2) I would not settle for anything less when going through this kind of new board design effort.

What I am getting at is - a decent board can be built using the original Peg as a "reference". The two best performing Morph platforms happen to be 7447 based (Peg and potential Mac Mini) It seems very likely that Morph would be ported to other Mac G4 variations. All of this is very cool, because in my opinion, the more machines Morph runs on, the better. Porting to Mac obviously easier than inventing new hardware as I am outlining here.

This new board idea I outline is a stab at hardware independence. Maybe the 7447 isn't the greatest anymore but it is still capable and is cheap. I do not think plunging directly into 86xx based CPU's is a wise idea. Easier to update Peg design in my opinion. Other than getting potentially screwed by upcoming USB3, the board would have a feature base that should enable it to have an OK lifespan.

All of this is of course just my humble opinion. I thought it wise to let Morph community have input in the matter, hence my posting.

without doubt there will be USB3 controllers for the PCI slot, so really not a problem.

Starting such a board based on the Peg makes no sense, as it would be a completly new design....

Lets say you have all specs and diagramms for the A2000, someone sends you samples for the AGA-chips and the 040, do you really think the A2000 schematics would be any help in designing an A4000 ?

I think a new board should use the 8640 CPU (preferably IMHO) or the 8610.

In order to produce (design, test, build, distribute) such board you would need a lot of money. In order to get your money back you'd need to sell a lot... so the question is: who would you sell this board to ?

I don't think producing ~1000 boards for potential MorphOS users (and I'm optimistic) would be sufficient to get your money back...

Hi, here are the schematics:

http://www.powerdeveloper.org/platforms/pegasos

The BOM does not really help as the last production was not RoHS compliant.

We may do the same with the EFIKA, eventually.

Producing 1000 boards is not enough to get your money back. We would need to rely on a new graphics package, which requires quite a bit more effort and expense.

We are trying to find a good way ahead with the i.MX515. We are still very interested in the 8610 eventhough Freescale is not very excited about PowerPC in the consumer space presently. There are some indications that this may not be the situation much longer...

R&B

Quote:

So maybe someone is actually going to be able how to add extra ram to the Efika sometime ?

Quote:

Might be nice. I wonder how large a new charge of efikas would needto be to get an attractive end user price, given one slight board modification:
bigger on board ram

Quote:

I still think ARM is quite risky. The competition is huge, but the market is huge, too. Personally I am not too interested in ARM, butif it brings some cash to your company (which then could be burned by throwing out a new ppc desigin )

Quote:

That would be gorgeous, but I guess Freescale's shipping difficult waters these days and I cannot avoid my impression that ideas what to do and which path to follow are getting changed on a daily basis (how to satisfy your shareholders and listen more to the 2 week consultant contractors instead of the in house tech department enough).

These are all very good comments.

I am definitely concerned about the Freescale end of things. It doesn't appear as if their long term strategy is clearly known at this time but obvious ARM development on their part makes at least that part of their strategy certain. It is difficult to deny ARM's many advantages. It is fully understood why Genesi is also looking in that direction. It makes good business sense.

The Freescale 8610 does show obvious promise due to its variety of integrated items. We are starting to see initial dev systems built around the 8610. Perhaps a future consumer system utilizing this processor will be released by someone and be Morphable. I am glad Genesi showing some interest in this one. Despite the ARM craze, I would expect Freecell to support devices like the 8610 for a long while yet. It's new, decent, and not everyone will want to dump PPC in hurry just to start completely over on a very different platform.

So my real question is where do all of these changes leave Morph? Mac porting will definitely keep it alive for extended periods of time. Morph gaining momentum in my opinion (at least in its development). Sure there is stuff to work out yet but the major foundation aspects of it are in place. It is a fresh approach that runs efficiently on non-complicated hardware. A better approach than expensive multicore cpu systems that get so hot that the next step in trying to cool them will most likely involve liquid nitrogen. These are a few reasons I inquire about future hardware aspects as they relate to Morph.

Thanks everyone for the comments!

I was really sad Genesi didn't make any upgrade for the Pegasos 2.
We Amigans used to buy accelerators, and expand our computers.
It seems this is not business any more, we are a small amount of users.

Making a computer from scratch costs a lot of money. Ok.
What about a PowerPC PCI-e card?
maybe you remember the inside-out amiga card.

What has ARM to do with MorphOS?

Regards

My original thoughts were on a new board but very good sources indicate economically impossible below 1000 unit quantities. They are expert in this area so I believe them. This is too bad because Morph's only hardware future seems to be a limitied quantity of Genesi boards or Mac machines.

So setting the hardware topic aside....What is the roadmap for MorphOS? As stated in my last post, it is a well designed OS that has distinct advantages and in my opinion a great candidate for mobile computing devices. Why go to the hassles of running complicated operating systems on small devices? Morph becoming a great featured OS requiring very little hardware overhead. Big mystery to me...

My A3000, despite its age, surprises most non-Amiga people I show it to, including hardcore intel-microsoft lovers. Their belief is that it is too old to be any good. They always end up surprised when I turn it on and take them for a spin. These PC people can't seem to believe that a 25mhz-16mb machine can do the things it does. Impossible they say! That was the genius behind the original and a genius recaptured and perfected by Morph.

Well I now have the Amiga Blues again so I will stop. Thanks for the discussion!

Rodomoc;

The design you suggested implies removing every component on the Pegasos and reimplementing it with another set of chipsets. Apart from connector locations and the layout of the backplane - if you even kept those, I doubt it since you would have to make a lot of room for the CPU, and add far more than the Pegasos had - you would be about as far removed from Pegasos as the Mac Mini is. That is to say, it would not even be worth using the Pegasos as a reference.

You may as well start designing a brand new board from scratch, and if you are putting the CPU on the motherboard, why not just use the MPC8610 or MPC8641D and save yourself some PCB space?

Quote:

It's not a business. It is also not economical for users. The proposed 7447A and 7448 upgrades would have given very little performance difference. Since the bus speed of the Discovery II was 133MHz, all you get is clock speed - 1.4Ghz or 1.6Ghz at the top end for these chips.

Therefore, given the pricing of the 7447A and 7448, making your system 50% faster would cost more than the Pegasos itself to end-users. They would also have generated more than 50% extra heat and required heavier cooling solutions (bplan did have a passive 7447A cooler that would have worked though, tested up to 1.4Ghz).

Quote:

It would cost about the same to make as a new computer from scratch, since it would also be.. from scratch.

The best options for putting a co-processor on a PCI-e card would be something of the MPC83xx class (8349E, 8378E) which tops out at 600Mhz. After that you have to cool it like a graphics card. It would also have to be a very complicated software driver to enable it to work - especially when thinking of displays etc. and transferring data between PC host and target PCIe.

Quote:

Nothing, but why does everything Genesi does have to have anything to do with MorphOS?

You might note the "how many registrations" thread and extrapolate how much money MorphOS has made in the last 8 months and decide if working with MorphOS can fund everything that you want to use MorphOS.

[ Edited by Neko on 2009/1/25 18:50 ]

I wonder what the feature for PPC is now ? and more what is the feature for MorphOS (and OS4) if it sticks to PPC ?

Is it doomed to stick with used apples, and underpowered hardware that is designed to use in a router or a decoder or something like that ?

And what happened with the Cell ? it's used in a PS3 and that's it ? What a shame.



[ Edited by Oepabakkes on 2009/1/25 22:20 ]

Quote:

Genesi suggested that idea not me..

Quote:

Maybe as complicated on software side, but hardware would use much less
components.

Quote:

I bought my Pegasos computer because Genesi announced it in conjunction
with MorphOS. There was ODW to run Linux at a first glance.
Not everything Genesi does might be related to MorphOS, but what kind
of business Genesi has apart from having sold and given away Pegasos
and Efikas? I hope you have success with any device you produce,
even more if it's able to run morphos.

Quote:

I might then bet my head that MorphOS or Genesi are on the wrong market, wrong platform, and wrong devices if they are thinking into
serious business.
Here the only reality is that we have MorphOS2.2 running on our
Pegasos G3/G4 and Efikas, myself still not registered.
1.4.5 was free to me except for my donation to the USB stack programmer (C. Hodges)
That it's terribly expensive to support.
Small comunity, high prices, no hardware replacement
in case of hardware disaster.

The registrations number must be compared, to the total number
of Pegasos and Efikas sold, data which I don't know.
And the number of registrations seems to be near 500, too much
in my modest opinion.

One simple Question. Mac took over Intel hardware.
It was so hard to translate this platform to an X86 model, or maybe
yo thought it was no business at all?

Regards

> And what happened with the Cell ? it's used in a PS3 and that's it ?

No, that's not it:

http://www-03.ibm.com/systems/bladecenter/hardware/servers/qs21/
http://www-03.ibm.com/systems/bladecenter/hardware/servers/qs22/
http://www.lanl.gov/roadrunner/ (using the above)
http://www.mc.com/microsites/cell/products.aspx (several boards and systems)
http://us.fixstars.com/products/gigaaccel/
http://www.fixstars.com/en/products/gigaaccel180/ (same as above)
http://www.matrix-vision.com/products/cell/mvxcell8i.php?lang=en
http://www.toshiba.co.jp/about/press/2009_01/pr0701.htm (upcoming "Cell TVs")
http://pro.sony.com/bbsccms/ext/ZEGO/ZEGO.shtml (upcoming, press release, brochure)

You are right.

This too:

http://www.powerdeveloper.org/platforms/cell

R&B

Quote:

Find a market where they need a high performance desktop-style system with PC-peripheral support based on MPC8610 or MPC8640D and we'll jump on it and go with it.

Right now the buzz is around ARM so we'll do ARM. Our MPC8610 Netbook spec is not abandoned.. it's just looking for a partner. Unfortunately the advantages in an MPC8610 Netbook are basically only in better performance and lower power consumption than Atom, however it will not run Windows and "but it can run MorphOS!" is really, really not a selling point to anyone.

While this is true, an MPC8610 Netbook may appear but you will still have to pay again for MorphOS. Unless there is a way MorphOS can be subsidised through sales of a MUCH higher selling product or OS, it needs to be that way.

The other problem - which strikes ARM too - is that Linux distributions available today DO NOT come up to scratch as a viable alternative to Windows. No, not Ubuntu, not Fedora, not SUSE. Linus Torvalds just switched back from KDE to GNOME, even though he thinks GNOME sucks, just because KDE 4.x is really not ready yet and won't be for another 6-9 months.

What are you meant to do here? Ship something that sucks, according to the creator of Linux, that's what. It's the only solution anyone has. All you can do after that is hope that you can ship a viable upgrade later.

Good things are on the horizon though; while KDE 4.3 or 4.4 appears you may be given a Qt-based desktop written by the Ubuntu guys, since it went LGPL this makes it a complete, open-source compliant platform which you do not need to play with the KDE exception to enjoy. However, this "Netbook Qt Desktop" will probably be 12 months or 18 months away..

Quote:

There are two points here:

1. Eee shipped with Linux/QT didn't it ? Has that stopped shipping now ? It seems to me that the market is there if you have the right product.

2. If mainstream consumers accepted this "weird Linux" thing rather than their beloved Windows then MorphOS potentially could be a suitable netbook OS. Although it's far from being ready. Us MorphOS users can dream about that.

Quote:

Mr Torvalds says many things. Perhaps he had corn flakes for breakfast today instead of toast and jam.

If you are stuck for a fast Linux desktop for low power devices then there are a number of available projects, some of which are quite mature. For example, you could easily go for QT embedded/QT Palmtop Environment as in Eee or Sharp Zaurus; then there is UME (is that what you are referring to ?) and also LXDE. Linux users are not waiting for "another 6-9 months" just to get a usable desktop.

By the way, I appreciate you dropping in to enlighten us on the present situation. It's just a shame that Genesi is still so far from a new product that we might enjoy.

all great links about cell .. but hardly mainstream hardware with exception of the PS3. The PS3 is to closed, so unless the unlikely event occurs that the MorphOSteam cuts a deal with them .. I am still not convinced what future a PPCOS has on the (niche) desktopmarket ?

Quote:

Ppl accept Linux on the Eee because it does the things that aere expected. *If* MorphOS came with OOo and a *really* decent browser with Flash support then ppl would accept MorphOS, too. Maybe some folks would even fall in love with MorphOS.
But from the three cruicial issues a netbook must do (multimedia, internet, office) MorphOS is only in the multimedia domain up to date (MPlayer, nice audio players). In the Internet domain, there are good parts: good enough mailers, nice IRCs, ftp is okay, but for browsing Sputnik is just not enough (yet!!). Plus there is no Flash support (no, this old flash player on Aminet is not an option). For office situation is compareable weird as for internet, with the exception that there is no development in this area. ButI'd say Office stuf is the least important, once a browser is advanced you can use google docs or the likes as an option.
Thus, as long as there is no proper www support forget about selling a MorphOS driven maschine to common ppl, once there is, I'd day everything is possible...

koan:
Asus is silently dropping Linux in favour of Windows.

Just look at their latest models - 100% Windows.
Acer does the same, and the rest mostly use Windows.