As far as I know, Freescale has no immediate plans to discontinue the 8640, but they did state that the 8610 was going to remain available for several years (making it a better bet for long term production).
Personally, I like the 8640 (as you know Andreas). But its a relatively high cost processor and the only Southbridge chips I could find that would work with it are produced by AMD.
Since you pointed out to me that there have already been posts (on MorphZone) on the recent introduction of the first Titan based Soc, I'm sure you're aware of the advantages of that processor (PCIe 2.0, etc.).

The main problem I see with QorlQ (or Titan or, for that matter, the 8640D) is that they're multicore processor (and, as you've repeatedly pointed out, MorphOS doesn't support SMT). While there are a few single core variants they're still limited in speed (making the multicore variants more attractive - if we could take advantage of them).

Anyway, after spending several months studying the idea of producing a new PPC motherboard, I've just decided to buy an Apple G4. Producing a new board would be quite expensive and except for the ability to use new expansion options (via PCIe), there isn't much of advantage over buying a used Apple (after all, the max clock speed for an 8641 is slower than the max speed I can upgrade a G4 Mac to).

I don't see the point in mortgaging my house to produce something that might not even get supported by MorphOS or AmigaOS.

> they did state that the 8610 was going to remain available for several years
> (making it a better bet for long term production).

My impression was that they were stating this about the 8610 because the PowerDeveloper project is about the 8610. I didn't read it as if the 8610 is supposed to be any longer in production than the 8640 will be.

> Since you pointed out to me that there have already been posts (on MorphZone) on the
> recent introduction of the first Titan based Soc, I'm sure you're aware of the advantages
> of that processor (PCIe 2.0, etc.).

Yes, of course I am. Btw, PCIe 2.0 is also featured in QorIQ P4 series (e500mc based, i.e. the one with "real" FPU).

> The main problem I see with QorlQ (or Titan or, for that matter, the 8640D) is that
> they're multicore processor (and, as you've repeatedly pointed out, MorphOS doesn't
> support SMT).

MorphOS can be programmed to ignore every core except one. That's how MorphOS will run on (one CPU of) dual-CPU G4 PowerMacs. Thus, I wouldn't regard it as real problem, just as a little wastage maybe ;-) As it looks now, e500mc based QorIQ will have four cores at least, which would result in a waste of three cores. 8640(D) is available as single core variant, which would result in no waste at all. Titan based APM 83290 will be available dual-cored, resulting in one wasted core.

> While there are a few single core variants they're still limited in speed

Both 8640 and 8640D max out at 1.25 GHz. Single core QorIQ (of P1 and P2 series, e500v2 based) - while being slower (1.2 GHz) than P4 series (1.5 GHz) - are not suitable for MorphOS anyway.

> making the multicore variants more attractive - if we could take advantage of them

We can, but unfortunately only one single core of them.

Thanks Andreas,
If there is anyone on this site who would have a better command of the facts, I haven't run into them.
But,it still seems like the easiest way to get a computer (for MorphOS) is just to buy a Mac. I've seen 1.25GHz Emacs (which have built-in Radeon 9200 graphics) for as little as $50. These and other G4 macs should be supported in the next release.

I do wish I had more reason the continue on with the 8540D project, but I don't know how much longer the SB600 is going to be available (that's the Southbridge I based the initial work on). The SB700, 710, & 750 could also be used (but I haven't bothered to get the technical info on them yet - AMD has granted me enough access to their technical resources that I could consider them).

Then there's some additional problems I've had with necessary "glue" circuitry. Freescale's evaluation board uses a FPGA. Short of recreating that, I could investigate a different approach, but its a added complication I must admit to being intimidated by.
And, of course there's the added complication of working with BGA Packages. I do know one company that I've worked with in the past that has the necessary equipment to bake the SMTs (but today's designs are so complicated that they make the 68K based board I've worked with in the past look like a viable hand soldering project).
And, of course, what happens if (and this seems inevitable) I make an error in the initial design? It could lead to some very time consuming troubleshooting.

Right now, I'm looking at the idea of placing the processor on an X16 PCIe card and installing it in an X86 motherboard. It would eliminate the need for a Southbridge and would lead to a less complicated circuit design.

Whatever happens with that, at least I should be able to run Morph on a Mac soon (while trying to make some use of the 8640Ds I've got on hand).

I just want a couple programs,
1) the office suite being worked on (Open Office like)
2) Simple dvd authoring program with gui- this is a must
3) printing!

[ Edited by xyphoid on 2010/1/4 22:09 ]

> IBM will have their PPC476 core to be implemented by LSI.

2 weeks ago, LSI announced the addition of the PPC476 to its portfolio of custom silicon IP:

http://www.lsi.com/news/product_news/2010/2010_01_20.html
http://www.lsi.com/storage_home/products_home/custom_silicon_solutions/ip/processors/ppc476fp/
http://www.lsi.com/DistributionSystem/AssetDocument/PPC476FP-PB-v7.pdf

I'd really like to see the Powerbook version out this year...

> 2 weeks ago, LSI announced the addition of the PPC476 to its portfolio
> of custom silicon IP

3 days ago, LSI announced the PPC476FP core based Axxia ACP3448 to be available this month:

http://www.lsi.com/about/newsroom/Pages/20100209bpr.aspx
http://www.lsi.com/products/networkingcomponents/Pages/AxxiaCommunicationProcessor.aspx
http://www.lsi.com/downloads/Public/Communication%20Processors/Axxia%20Communication%20Processor/LSI_PB_2pg_Axxia_1.3c_LR_0411.pdf (also including ACP3421, ACP3423 and ACP3442 models)
http://www.power.org/wp-content/uploads/2012/09/LSI_ACP3400_PB.pdf


Edit: Updated links to reflect new LSI website layout.
Edit2: added power.org link because lsi.com links are dead

[ Edited by Andreas_Wolf 20.05.2016 - 20:19 ]

Quote:

!!!!!!! that's something i've been dreaming about for years, having a whole ppc (or 68k years go) running inside my X86 PC. And something affordable ,not these cell cards : )

Update:

>> I think The PPC path is death [...] Well, aside from PS3 and XBOX360 consoles.

> Don't forget the Wii :-)

And another one (okay, more arcade system than gaming console), announced some days ago with a mysterious PPC SoC ("Y3", developed internally):

http://amigaworld.net/modules/newbb/viewtopic.php?topic_id=32389&forum=17
http://www.sie.co.jp/news/news100907.pdf (English translation by Google)

Hi, I don't know if the new supported mac hardware which will be supported will also use the sungem_eth.device but if so then I would expect that the LAN connextion problems will be solved because I still can't even use my Mac Mini's because of these LAN probelms!

Update:

>> Cell processor designs (haven't heard anything in this direction for a long time)

> "In November 2009, an IBM representative said that it has discontinued
> the development of a Cell processor with 32 SPUs but they have not
> halted development of other future products in the Cell family."
> http://en.wikipedia.org/wiki/Cell_(microprocessor)#History

IBM's Systems and Technology Group CTO says:

"Development around the original Cell processor hasn't stalled and IBM will continue to develop chips and supply hardware for future gaming consoles, a company executive said. IBM is working with gaming machine vendors including Nintendo and Sony, said Jai Menon, CTO of IBM's Systems and Technology Group, during an interview Thursday. "We want to stay in the business, we intend to stay in the business," he said. [...] IBM also will continue to invest in Cell as part of its hybrid and multicore chip strategy, Menon said. "I think you'll see [Cell] integrated into our future Power road map. That's the way to think about it as opposed to a separate line -- it'll just get integrated into the next line of things that we do," Menon said. "But certainly, we're working with all of the game folks to provide our capabilities into those next-generation machines." Menon [...] said its Power processors have capabilities that are similar to those of the Cell chips. "The kinds of things you used to be able to do with the Cell you can do in ... [the] Power line," Menon said."
http://www.pcworld.idg.com.au/article/363760/cell_processor_development_hasn_t_stalled_ibm_cto_says/

To me that sounds like the Cell and POWER lines will merge into a single line, i.e. POWER8 will sport Cell SPE (like) units.

2010 is almost over now. While some things kept beyond expectation (e.g. user number increase), I think it was a rather nice year. I used MorphOS for many things and am especially happy that I was able to use it productively for my university job.

An interesting re-visit of those thoughts.. I read each one including my own and they can be pretty telling.

First, PPC is not dead (although Titan is out of here).

Secondt, we did recieve all the promised hardware support, and look forward to even more great additions (and I found a 1.8 not a 1.6 Mhz processor). The path chosen by the development team seems better than any alternative ideas offered by us.

Third, I spent a LOT of time on too many different projects that went no where. Anybody want a couple MPC8640D processors?

Fourth, Andreas fact based posts and projection still appear to be the most accurate posted here (still waiting for a Power/Cell fusion though).

Have a good New Year's Eve tomorrow folks and remember to celebrate my birthday!
(Of course now remembering you in Europe are already at 12/31)

BTW - Does this thread now morph into a what do you hope for 2011 thread or does that belong elsewhere?

naaa We'll just open a new one after 2.8 is released
happy new year 2 all!

@amigadave
Quote:
Personally, I think you would be much better off not trying to compete for an ever shrinking user base and instead go for new users. I keep saying the same about AROS where fans seem obsessed with "winning" over old customers. Real growth will only happen with a focus outside the ever shrinking Amiga community. This is probably the only thing Commodore-USA has done right so far...