It seems that there has been some (whatever) problems with bringing the 5121e chip to a final consumer level. For sure it has taken longer time than anyone would have expected.

However, now I see that the CherryPal site has been updated: "We are accepting orders again, shipment on 11/4 guaranteed!!!". And this in combination with a comment in bbrv's blog from Peter Czanic ("A rev4 ADS5121 board just arrived to my desk, and it is quite promising. It is fast and stable") makes me wonder if they finally got it right?

I hope so!

Oh BTW, I just spotted this on powerdeveloper.org: "the Freescale mpc5121e board boots openSUSE 11.0 and works from the HDD now".

I wouldn't count on it :D

The 5121E SuSE project is to get a handle on how much better the 5121E could be for an Efika-like board, and to show up a few flaws and foibles in the development model they have at Freescale.

It's going ahead whether we have a 5121E design in the pipes or not.. what you make of that, is what you make of that.

@Neko

OK I hear you.

What I tried to do was simply to add up a few public observations I made to draw a conclusion about the *5121e chip*, and not necessarily a board design from you (Genesi/bplan). Many are the times that you have started doing a board design around some CPU (SoC or not), only to find that there *might* be something better beyond the horizon, so you put what you are working at "on hold" and start waiting for the new chip, and when you get your hands on it, you decide that it still isn't 100% according to your needs, so you start waiting for the next thing, and so on. The 8610 was only the latest thing you put on hold, preceded by many others. Which brings us to an interesting question (regarding your post above) - will there ever be a 5121e/5123 board from you at all?

... can You imagine what Powerpc development
would be like if they backed something similar to the Open Pandora
project? From what I've Read TI was promoting there product
and gave all the support they could... It would be nice to
see that kind of development by Freescale.
Genesi has already shown a desire
to spur low power development... especially with the Efika.
Just doesn't seem as if they've had much help...
Maybe this is the new direction Freescale is seeing??
I really think that a Powerpc core could compete with the Arm architecture?

[ Edited by katos1 on 2008/10/14 21:10 ]

Quote:

I agree.

Quote:

Let's hope so!

Quote:

I think so too, IMHO there are already a couple of CPU's in Freescale's portfolio that could compete, and the 5121e/5123 could be seen as a flag ship. Is there even a corresponding ARM CPU? And this message suggests an interesting future in this regard: "Freescale offers customization across SoC platforms to help customers streamline unique product development". I think this could be exactly what you are talking about!

Quote:

I still think so, too. But it's getting harder. Intel is driving the market with the Atom. Atom + chipset based devices *are* actually cheap (and available *now*) while some ppc SoC based devices *could* be cheaper.
Intel is delivering lots of chips now (time to market is very short) - and isn't it strange that I was able to purchase an Asus Eee with Atom (plus according chipset) for more or less the same amount** the Cherrypal is scheduled for?
How on earth should a cherrypal compete with that:
A slim 400MHz box vs. a 9" 1.6 GHz netbook (btw.: this is a rhetoric question *I* know about the strength and flaws of both chip families)?
The ppc could be cheaper once a critical mass of a production is reached, but how can you reach critical mass when the prices are initially not lower? By risking high loans? Well, see the last weeks to learn a lesson about high risk loans...
Thus, I haven't given up yet, but I am not too confident anymore.
Still, a *soon* available Efika2 for 99EUR (board only) or Efika2 mini box for less than 179 EUR (case and 8-12 GB Flash or *silent* hdd) or a Efika2 based 9" netbook (8Gb flash, 4 hrs battery runtime, 9" 1024x 600 display) for 249 EUR would have a chance to get sold in serious quantities IMHO. Prices incl. typical European VAT of about 20%. Plus a cheap bundled MorphOS (in a better polished form, flawless usb for a start) *could* be a real advanatge***.

Another story of course is when focussing on the emerging markets like China (THTF?) or India where all the above written things might be moot.

--
** Got a brandnew Eee 900a for 260 EUR incl. VAT last week. It's a nice thing, but the SSD is rather slow. On first impression applications are running conveniently fast on the Atom. Haven't benchmarked the thing yet.

*** A lesson I learned from the Eee: the OS dosn't matter much. Whether the Eee runs on Xandros or XP doesn't make much of a difference. There is a simple UI, a set of applications and off you go. Joe Average will keep the installed set of apps and just use the thing. A random lock up of the system isn't too worrying on those devices (it's a gadget). Crashy apps are on Xandros, too (at least my first imported powerpoint file made StarOffice crash). And only MorphOS has the power to make a 400 MHz ppc actaully *feeling* fast (with Layers 3D it could *blast away* the crappy XP UI or the slow Linux stuff).

[ Edited by Zylesea on 2008/10/15 11:10 ]

Quote:

Anything in the TI OMAP range blows the MPC5200B away. You can make your own OpenPandora box with the Beagleboard if you want.

Quote:

Note that Atom is actually a really low performance chip and the TDP figures are an outright lie; Intel actually sells Atom as a *kit* - CPU and Northbridge, as there is only one Northbridge available for it - for about $70 for the low end ones. The CPU may be 2-4W but the Northbridge is somewhere around 21-30W.

If you plug in your Pegasos G4 you will get better performance than a single-core Atom and the same power consumption!

The MPC8610 is all set to blow Atom away, if only it cost $70, it would be perfect. Even more perfect would be products using it. The official MPC8610 power consumption figures should be released soon.. the preliminaries I have seen here are pretty favourable. 12.1W at 1.3GHz.

It should not be hard to compete on a board level for a low-power, high performance Linux box. We know MPEG4 decoding at HDTV resolutions is well within the performance characteristics of the MPC8610 (check out lonelywild's youtube video - this is a shot inside Freescale Japan!) although probably not at the same time as running a full desktop Linux and all the overhead it brings, it's still an impressive feat for a 1.3GHz processor.

We're still gunning for a MPC8610 Netbook... it's still cool and will remain being cool until it happens. The Netbook market is going to explode next year. There could be a million or more devices sold from any one player. It would be a damn shame if Freescale did not see that they could sell at least a million chips, and a slightly bigger shame if we could not surprise them with a million chips a quarter.

Quote:

I wouldn't buy a Netbook that only had a 400MHz processor even if it did have wobbly windows. The Efika's performance problems stem entirely from lack of RAM - MorphOS is far leaner than Linux so it manages to make more of what it has. This is why it feels faster. However you are still going to be stuck with a 400MHz 603e core - editing those large OpenOffice documents is going to crawl on either OS (if MorphOS had OpenOffice, that might even be a valid comparison).

Let's take something that we COULD do - get Flash running on Linux and MorphOS. 400MHz is NOT nearly powerful enough to run desktop-sized flash animations at smooth speeds. The architecture Flash provides and the complexity of vector graphics simply cannot go that fast.

What I have in my head (and in the marketing requirements document I wrote to propose the Netbook) is a 10-12" Netbook (see Dell Mini 9, Dell 12, or the HP MiniNote for examples) with the MPC8610, primarily running SuSE Linux (11.0 or 11.1). With a gigabyte or two of RAM (up to the spec of a MacBook Air) you have plenty of performance to spread around both on the processor and in the rest of the system.

[ Edited by Neko on 2008/10/29 17:12 ]

Quote:

I wasn't talking about the 5200B, but the 5121e. And in the context of competing with ARM I think there are a few others that can compete in various areas as well (not every application needs a display controller for instance).

And beyond the horizon there could be some interesting things in development as well?

THIS roadmap image suggests (at least) two more planned models in the MobileGT line, and unlike me you might actually know something about what those two might be, and it would be kind of pointless for me to speculate. However, THIS old Genesi picture from Oct 2006 clearly describes the 5121e ("We will start with this!!!", "it is not the 4U2, at least not yet") is followed by THIS old Genesi picture ("It is 4U2!!!") that illustrates a new SoC CPU that IMHO would fit very well in a product portfolio if you want to compete with the one of ARM.

And then (to quote myself from a post above) this message suggests an interesting future in this regard: "Freescale offers customization across SoC platforms to help customers streamline unique product development".

IMHO, many interesting possibilities can follow from this!

Quote:

While I think the Beagle board is truly interesting, I also think that its SoC lacks quite a few important "on chip controllers" for it to be called "a true SoC". Its built-in DSP and video accelerator is very cool indeed, but relying on USB for everything else kind of takes the edge of it IMHO.

Quote:

That requires someone actually making the products, a business that *at least I* was thinking Genesi was into. Or at least tried to do, even if you haven't actually done that many real products if you look at it a bit closer. Your 8610 project was only the latest in line to be canceled, and it's always explained by some "...don't worry Karl we have something better underway". A lot of talk and bombastic blogs, a lot of time passes, then the thing is shelved forever, and it all starts over with some new "product". It's always something better underway, right?

Quote:



You mean through the Japanese "Redtail" board?

Well, that's great I suppose. Hope it materializes!

With that board (in "book" or any other shape) *available for MorphOS* I would say bye bye to the entire Mac Mini thought!

Quote:

Now you're talking! Let's do it then!

Quote:

While that's true, it is also true that many things are a question of price and othe advanatges.
The 5121 *is* cheap and has a really low energy consumption.
Taken together this could make a really low-price, low-end netbook. Of course the raw cpu power on a 400MHz 603 is really limited and e.g. loading big pictures (average digicam is 8-10 MP today**) or complex pdfs takes several seconds, processing that stuff really requires some patience.
But usually netbooks are for simple tasks, and there 400MHz may be enough for quite something.
--
** Many of those 8-10 MP digi cams are real crap and most ppl never play around with the settings and just shoot the pictures with the deafult settings. Even worse: most ppl never rescale pictures to the actual need, who's not "happy" to recieve emails of many megabytes to show the baby's first steps or such in 8-10 MP... grmpf!

Mixed up some things.

[ Edited by Zylesea on 2008/10/30 15:56 ]

> http://www.flickr.com/photos/54643129@N00/2952481050/
> look srather like some work by bplan and better
> suited for a netbook (*really* small).

That *is* the Redtail. Or did I get you wrong?

No, I mixed up some things. Forget the post before.