Market research for new PowerPC system

[markos]

I was pointed to this forum, I didn't know there existed a separate discussion here. Anyway, as noted in:

http://amigaworld.net/modules/newbb/viewtopic.php?mode=viewtopic&topic_id=29594&forum=33&start=40&viewmode=flat&order=0#509376

I appreciate the noise, but there are a few points I'd like to mention:

* I am not a Mythical Hero.
* This is not vapourware, I do indeed plan to invest on this.
* MorphOS and AmigaOS were not target OSes in the beginning, Linux and Haiku were. Whatever you say, Linux is a huge market for embedded AND PowerPC, and Haiku has a great potential. I would like to think that this effort is not a one-off. That is, I hope to sell enough to fund the next batch and then the next and then an upgrade, and another, etc. Whether AmigaOS/MorphOS are supported that will depend heavily on the number of users that ask this and the dev teams involved. If only 10 people ask for AmigaOS support, then it's not going to happen.
* For those who haven't seen it yet, Haiku is going to do on the desktop what Linux failed to do (not now, in a few years).
* Variety is good. PPC is not dead, and unless something really bad happens -eg. Freescale/IBM close down- it won't really die.

In any case, I'll blog about the mails I receive so far, which are surprisingly enough, given that I posted this thing <24h ago.

So, I understand the sceptiscism, I really appreciate the feedback, I note everything and I will take into consideration all important comments.

About the last comment. I'm sorry, I don't see how investing money into software for the AmigaOS, will be successful. That's for the Amiga Software companies, and IMHO, that would be even more suicidal than investing in hardware, but that's just MHO.

Regards

Konstantinos

PS. Yes it's really me.

[markos]

Apart from the variety factor, Atom power consumption is in fact higher than the MPC8610 (at least the earlier models, don't know about the 330 to be honest). Also since the speed is similar, and the Atom is not really the fastest Intel chip -nothing to do with C2D or Nehalems, etc- , performance-wise, the PPC chips will be very competitive. The price is another matter, but for that reason I'm looking to include something extra for added value (eg. USB 3.0 host chip). That still remains to be seen.

Anyway, as I said elsewhere, the question is not if there is a market. The market is there. The question is "is the market big enough to sustain itself?"
If I get enough interest I will go forward. If I get 90% mails to deter me from going forward with this, well chances are that I will abide and just invest my money on ARM instead.

Regards

Konstantinos

[markos]

ARM would be a much better option as an alternative to x86. One big question for Hyperion and MOS crew, will there be a follow on or will they be porting to an orphan mobo? Unless I miss my guess, this is going to be a VERY limited number production run of mobos. Is it likely to be a follow on, or a cold dead end?

The first production run will be about 500 units. This will be used to fund the next and so on. Well, that's going to be the way, unless I manage to get an investor with huge pockets and order 10k units at once... (unlikely given the current situation).

If you think that porting AmigaOS and MorphOS to ARM would be of benefit just for a single motherboard, then I think you got this on a wrong basis. It would be the start for many nice things, eg. eventually, one might just sell AmigaOS-loaded Beagleboards, or sth like that. The potential is big. Anyway, you know, if the PowerPC thing doesn't gather enough interest, I'd just as well try my luck on ARM.

Konstantinos

[markos]

@Dammy

You know it's posts like this that make me want to start this sooner than later. So far, in just one day of a post, I've received ~40 mails of interest. I've set a milestone, end of October to see if I'll manage to surpass 500. And I don't think you know what "profitable" business means. If I manage to make even 1 euro out of every board, that will be a profitable business. Stupid, but profitable. In truth I plan to have a sane profit, if only to be able to fund the next production batch. This is too much hassle to be a one-off production. And it's not that I would need it anyway, I'm doing quite ok doing enterprise and embedded applications so far.

But enough words from me. If you don't believe me, that's respected. I'd love to prove you wrong though and I can only do it with actions (building the board), and not words.

Konstantinos

[markos]

500 minus hold back for warranty replacement of say 50 at a bare minimum would be 450. Don't think I would risk that much money on either porting a OS to it or actual manufacturing of it unless there was a signed contract for the rest. Worst of all, what if there is a major flaw or bug in the chips like we have seen in the past PPC systems? And they do have issues as we have seen posted over the years.

That's a reasonable argument. Well, As far as the MPC8610 and 8640D go, they're not really "new" chips, and most bugs have been fixed. But it does hold as an argument with the P1022, which is not even out yet (Q3/Q4 2010). Well, I can't say much in this case. I hope the producer does not withhold any hidden info -like the cache coherency in the 5121e chip) and the prototypes are hopefully to find any error seen in the chips early enough -yes, there will be prototypes

[markos]

@persia

I don't think that's going to happen, I'm sorry. The scales are just not the same. Even if I wanted to I couldn't sell a PowerPC board at the same price levels as eg. an Asus motherboard -and neither could anyone except Freescale and IBM themselves who will not do that. Go over at mouser.com for example or some other parts supplier and check the prices of the PowerPC cpus there. It's just prohibitive, and there are only a few options to choose from to build a decent board at prices affordable by most people and not just Telcos/Defense companies.

[markos]

Which reference design? MPC8610's developer platform? MPC8640D? P1022? No, it's going to be a from-scratch design -of course following the guidelines set by the reference design and specs. Refabricating a reference design is not the cheapest solution, as to many components are used, which are not needed by the end user.

[markos]

Many of you seem to fail to understand what I'm asking. This is not about a new PowerPC Amiga or a new system running MorphOS. This is about a new and more modern PowerPC platform, which will *primarily* run Linux -and probably Haiku- and which IF there is enough demand, it will support your preferred OSes (AmigaOS/MorphOS). Again. I'm NOT building a new Amiga. The platform will most likely happen. If you want support for it, you might as well ask and if there are enough requests, I will try to follow that route. If only 20 of you ask now, then 99% the board will not support anything than Linux/Haiku. I hope this has been made more clear now.

[markos]

@the_leander:

A games console is much more than just hardware, it's an complete SDK, it's lots of cash to give out to developers and publishers to develop games platform. It costs way much more than just building a board. And when there are already XBOX 360, PS3, and Wii, what chances does a small company have to compete? I haven't seen a niche games console so far.

[markos]

Nice. Interesting, thanks for the link. I think a PPC-based one would be much more expensive though and not directly competitive, but it makes room for new ideas. Thanks a lot.

[markos]

@tone007

Well, I think you're misunderstand the target group. I'm not after people who need the fastest gaming rig, nor professionals who need the absolute 3d power to do their renders or their DTP. I'm not after the Core 2 Duo or Core i7 CPUs, nor the Power7 market, I'm after people who like different things, the same group of geeks who prefer to get a custom DIY helicopter model kit, even though it's not top spec, instead of buying the fastest helicopter model available -which is also cheaper than the one you'll be building. Also, I'm targeting people who really want a power efficient CPU. ARM is great on the power efficiency, but just won't hold a candle to one of the CPUs I mentioned (at least not yet). Atom is more or less at the same performance ratio -I believe it should perform quite less than the MPC8610, let's not even compare it to the MPC8640D, and it's power usage has been ridiculed since it's release. ION saves the situation slightly, but that's a GPU/Chipset, and it has nothing to do with the CPU. I'm pretty sure a MPC8610+radeon mobile gpu could give as much performance with as similar power usage to the Atom+ION. Yes the price is higher, but price isn't everything.

[markos]

@persia

Fine, no one is forcing you to buy it, do I at least have your permission of trying to make myself look ridiculous by failing to sell PowerPC boards? Btw, geeks bother all the time for old fashined, overpriced hardware, I once collected old hardware like SGI O2s, PPC macs, HP/PA workstations, etc. I actually paid MUCH more than the 500€ you find hard to justify for this board. Not referring specifically to you, you're in an Amiga forum after all, but I feel sorry for those people whose computer experience consists solely of a BIOS and a Windows screen. There are many geeks willing to try something different. Just as in the car world, there are many people willing to spend lots of money into a old Jaguar, where obviously a newer Audi would be a much better way to spend their money.

[markos]

I guess there is no interest/technical expertise that is needed. In the case of MacOS X there are literally hundreds of thousands of users/developers that run it on plain x86 hardware. How many Amiga users/developers know AmigaOS AND ppc hardware so well AND have so much time in their hands to do that? Sure it's possible, but what's the point?

[markos]

Update:

http://www.codex.gr/index.php?pageID=&blogItem=62

[markos]

Who wants to spend $600+ on a system without even trying it out then to find it is limited as to what can be run on it.

I don't remember myseflf asking for preorders. When it's out, I'll send a few boards to reviewers, you can make your decision afterwards.

@tonyyeb:
EDIT: I wrote by mistake "paying without trying". No, I guess I didn't want to write that. But I will offer a full refund -minus shipping costs of course- to anyone that feels the board doesn't do what he wants. But it *is* too early for this talk.