|||

[ Edited by Velcro_SP 22.06.2011 - 17:37 ]

In my opinion Cherrypal and LimePC is dead

Freescale don't fix MPC5121e

More info about problems MPC5121e

http://www.powerdeveloper.org/forums/viewtopic.php?t=1478&postdays=0&postorder=asc&start=30

Since the 5121 has issues (which are severe for Linux, but shouldn't matter too much on MorphOS AFAIK) I doubt we'll see much of the 5121. The threads on powerdeveloper are really inteesting.
IMHO it turns out that the 8610 virtually seems the eventual chance for the ppc. Industry is turning towards ARM.
At least ARM can be run in big endian mode (a transparent m68k and ppc emulation would be possible).

> why can't Genesi field a similar unit, maybe
> using the 5200B instead [...]?

Because the MPC5200B has neither integrated display controller nor integrated 2D/3D GPU. As separate chips these won't fit on a board as small as the LimePC/Cherrypal one.

> how about a business model ala Cherrypal?

MorphOS Team selling hardware? Sounds unlikely at most.

|||

[ Edited by Velcro_SP 22.06.2011 - 17:27 ]

@ VelcroSP

Quote:

The basic "business model" of Cherrypal is as follows:

1. Sell a computer at $249 at a loss or no profit.

2. Bundle system software that requires your customers to be online to use essential features so you can show them paid for advertizements.


The label "green computer" is just means to an end. It is supposed to lure a sizable amount of environmentally conscious consumers into buying a computer that, unlike an Efika Open Client, is not even capable to let you run a 22 inch monitor in its native resolution, which happens to be the second most popular display size today in terms of world-wide sales (in saturated markets, it is number one already).

The choice of the Freescale MPC5121e might also have a lot less to do with its technical specifications than you might assume. When your business model is based on selling advertizement space, you have no interest in seeing your customers install other operating systems that do not include your custom advertizing display software or integrate web services from competing providers.

Of course, if you take the easy road and use an off-the-shelves AMD Geode-based design with very similar power consumption, your customers will be able to install a variety of operating systems whenever they like. But, if you use an exotic chip like the 5121e, end consumers do not have much of a choice besides using the bundled system software.


In any case, the Cherrypal business model is in no way applicable to what the MorphOS team is doing. One might even question if their business model is worth pursuing in the first place. If you look at the large number of ad-financed companies that have shut down or changed directions in the last number of years, you should become rather sceptic.

Quote:

All correct.

Right now, it seems CherryPal will add themselves to the pile of internet based startups that failed. One of the reasons is that the processor chosen for its product (MPC5121) looks brilliant on paper, but doesn't deliver on some of the promises, and requires a substantial effort in software to make it work properly. Details are scarce, but the famuos "lack of cache coherency" is to blame. I don't know exactly why Genesi's software (Linux) requires cache coherency, but that's all I know.
Way before CherryPal's emerge, Genesi was the one with a "contract" signed with Tsinghua-Tongfang (THTF) to realize the "LimePC". The chinese company rejected the claims from Genesi about the "defects" of the CPU, and attempted to take the challenge themselves (well, not really). The prototypes shown at CES were probably fakes.
Now, CherryPal is broke. Their once delayed product still hasn't reached a single customer. Or, at least, no one has written proof about them in internet.

Al these informations come from "powerdeveloper.org", which has a couple of must-read threads about this subject, and much more.

What about our beloved MorphOS? I was convinced that we could buy a bunch of CherryPals and donate them to the Team, but that's no guarantee of anything, Genesi says.

So, even if a MPC5121e based device looks like a perfect match for MorphOS (as the MPC5200 in the Efika is), I don't count on seeing it happen. What a pity. The only positive thing is that, after months, there are MPC521e development boards running Linux fine. That's all. It seems nobody is taking the risk to build a product around them.

Have you checked BeagleBoard yet?

Quote:

If I understood it correctly it is 'only' the Linux kernel and drivers that needs to be adapted. The available Linux kernel seems not to check for coherence itself but relies on a correctly working MMU. Thus, on the 5121 where the chip doesn't warrant a coherent cache it just crashes once you use severall chip devices (e300 + gfx and/or axe core, but I guess also when accessing the pci bus or so). But once there is a linux kernel that offers a coherence check and appriate drivers are written everything should run fine (but at a performance cost I'd assume (context switch like pn the PowerUP? - don't know if it requires a clearance of the cache)).
Exec has a function to check for cache coherence while using DMA (CachePreDMA and CachePostDMA), thus device drivers for the 5121 could be accordingly written without the need to change the core of the OS itself.

What I think is a bit odd is that there is Linux on virtually any device today and I guess there must exist linuxes for devices with buggy MMU and inconsistant chaches already, why is a Linux adaptation that difficult then?

Quote:

Zylesea, you've hit the nail exactly on the spot. Software does work on CPUs without cache coherency since ages. What's the deal with Genesi? Sure, there's something else.
Remember, the chinese behind LimePC/CherryPal didn't believe Genesi's claims about this. But then, they were unable to cope with this. Very interesting.

From what I've learnt recently, you actually don't want cache coherency. The affected software is drivers, kernels don't care at all, because this comes into play when we talk about devices, and that's driver territory.

When a driver maps a device into memory, it simply tells the CPU not to have that memory cacheable. That's all. Of course, when you write data to a device, you want your data to get (and come) accross directly, without anything in between that might alter it.

I guess only Matt "Neko" Sealey can shed a light on this.

|||

[ Edited by Velcro_SP 22.06.2011 - 17:21 ]

> the 8610 [...] uses a lot more power and is bigger than the Efika.

I don't think comparing a processor to a whole board makes any sense :-P All the more if you come to the obscure conclusion that the processor is bigger than the board. In fact, I measured that the MPC8610 "Redtail" board you linked to is even smaller than the Efika. Furthermore, an MPC8610 underclocked to 400 MHz would very well be as power-non-intensive as the MPC5200B. When consuming more power, the MPC8610 is higher clocked and thus better performing than the MPC5200B.

> to me it seems like it is going back to what's been done before.

Yes, hopefully it would seem so. What was done before with the Pegasos series was good but too short-lived IMHO.
Now the MPC8610 is much more integrated than the "old" discrete G4s, which allows for smaller board size, lower costs and higher (memory) performance.

> MorphOS doesn't need that. MorphOS can distinguish itself by still
> being responsive on tiny low-power boards with passively-cooled
> embedded-type processors that are slow with other OSs

Operating systems as self purpose? Certain applications running on MorphOS definitely do need horsepower, things like decoding high-resolution videos for instance.
First of all, I'm a desktop user. No MorphOS for desktop like hardware means no MorphOS for me. MorphOS only for "tiny low-power boards with passively-cooled embedded-type processors" doesn't benefit me. If it's for market penetration allowing MorphOS on desktop style hardware (including laptops) as kind of side project, then fine.

> Tsinghua "announced" "LimeBook."

Fine. So they were serious back then.

> some of the LimePCs [...] were actually fake performances by way of
> standard PCs concealed under the counter.

That would very well explain Cherrypal's iTunes and Adobe Flash running natively claim ;-)

|||

[ Edited by Velcro_SP 22.06.2011 - 17:12 ]

> you're not so silly as to think I meant that

Right, therefore the emoticon.

> so why say something like that.

To remind you to be well reasoned in expressing your thoughts.

> I even linked the board.

That's why I went on taking your "8610" for "8610 board" and even explicitly referring to your link to the board.

> I'm surprised if the 8610 board is smaller than the Efika

I simply took the size of the PATA connector -- as projected on my monitor via your link -- as base to conclude the real size of the board. No rocket science involved ;-)

> it still has the disadvantage of needing a fan and using much more power.

That's why I told you that "an MPC8610 underclocked to 400 MHz would very well be as power-non-intensive as the MPC5200B." Read over? A 667 MHz MPC8610 underclocked to 400 MHz would neither need a fan nor use more power than a 400 MHz MPC5200B.

> Efika can play videos pretty well. Yes, you can say "oh but not
> HIGH resolution videos" and then set the bar at whatever resolution
> you need to to "win" the argument, H264 at 1920 x 1152 perhaps?
> I'd take the positive approach that it can play medium-res videos
> pretty well.

Yes, you can say "it can play medium-res videos pretty well" and then set the bar at whatever resolution you need to to win the argument, 320 x 240 perhaps? :-P
Put another way: The MPC5121e couldn't provide the horsepower *I* need and want for *my* everyday desktop computing, which includes video watching. If it's sufficient for *you*, I can accept that.

> you confuse CherryPal's claims with those of LimePC.

No, I had BBRVs "These false systems are what MTC/THTF used to sell CherryPal on the concept, which then CherryPal used to [...] create a website full of misrepresentation about the chip" in mind. So maybe that way MTC's/THTF's claims (iTunes and Adobe Flash on MPC5121e with Linux when in fact running on "standard PCs" with Linux) became Cherrypal's claims (as to be read on their website) as well. Pure speculation, though. If Cherrypal's device never comes to fruition, we'll probably never know if their claims have been true, false on purpose or false by having been mislead by their partner.

|||


[ Edited by Velcro_SP 22.06.2011 - 17:07 ]

@ jcmarcos

Wouldn't it be great if the bufferfly on the box was blue - for a reason.

The Efika 5200B board has a bit bigger diemnsions than the Japanese 8610 board. But it gets relativated when considering taht the Efika comes with mounting space for the hdd while this 8610 board doesn't.

Thus while actually being bigger, it may be effectivly smaller. No wonder - the 8610 istself is bigger, requieres a NIC and a southbridge while teh 5121 is virtually complete. The 8610 integrates many compounds ofa comuter sytsem in one chip, but the 5121 *is* a computer system on a chip.
I still see some potential for the 5121: ultra low cost, ultra low wattage, ultra low dimension systems are possible. Systems have reasonable horse power.

The 8610 OTOH has quite some horsepower and is where the real computing fun starts. And it is still low cost, low wattage, low dimension - but other than the 5121 w/o the prefix 'ultra', but with a faster core, L2 cache, Altivec and a faster memory interface.

Quote:

It strongly reminds me of the yellow "Eclipsis" butterfly however...

297207387_3da5254eb0_s.jpg

[ Edited by takemehomegrandma on 2008/11/14 15:50 ]

Quote:

Holy crap! I didn't realize there was a butterfly in the box! All I wanted was people to realize that LimePC has, at last, a reworked web site, in which products seem pretty real.
There is the "LimeBook", the "LimePC X-1" (which is a "CherryPal"), the "LimeTV" (in assorted sizes) and a mysterious "LimeSTB" set top box which, as we all know is useless without any kind of attached service... Unless a LimePC with a TV decoder can be considered such. The (supposed fake) palm sized devices that people saw at past CES are in the last slide of a presentation, accompanied by forthcoming products such as a... projector!

Now, back to the butterfly... I'm tempted again to buy and give a couple of these devices to the Team, to see MorphOS running on them by next year. Don't take me seriously, but if these claims on "LimeFree.org" are true, then it IS possible:

"The limefree.org project is an open source project sponsored by limepc,with its mission to develop a complete PC system of software+hardware running on the MPC-5121e PowerPC microprocessor,The result is a completely open source software and hardware platform with no proprietar,components"

I understand that Genesi has withdrawed from building products with the MPC5121e. This LimePC offer is exactly the same as CherryPal has, even with the "cloud" thing.

Hi Folks, please check the dates. That is the cover of a 48 page document that Genesi developed for MTC/THTF. With the exception of including MorphOS, they have done pretty much as we outlined in the marketing requirements document for the product line. We basically explained the situation over the last year on PowerDeveloper. And, now MTC/THTF will try to create version of PowerDeveloper. Their business practices are shameful. The management has absolutely no sense of ethics. We have nothing but contempt for the CEO.

<break>

We have worked with Freescale on the 5121e extensively (and been well compensated). After MTC/THTF disregarded our agreement, they went on to basically screw everyone else involved (that includes Freescale so far). We have LimePC and LimeBook units -- they are not commercially viable. The LimePC is the CherryPal with a different logo. CherryPal does not have a cloud and they don't have money. MTC/THTF is trying to sell the LimeBook to users in China that don't know any better.

<break>

We had a few set backs, but we will have a new product in the market in 2009 if things keep moving along as they are in spite of all the effort we made around the 5121e.

R&B

|||

[ Edited by Velcro_SP 22.06.2011 - 17:02 ]

...in a few words:

1. The hardware is poorly designed.

2. The software is incomplete and does not properly use the three cores.

3. The marketing severely exaggerates the ability of the product.

R&B

Quote:

Isn't that whats marketing about?

Quote:

Not if you want happy and returning customers...

Quote:

Wow. Exactly the same product lineup that the LamePC (pun intended)!
Is there any chance that your document can be made public? Sure it's a very interesting read.

Quote:

It's very clear they've done so. MorphOS could be the way of making the MPC5121e chip shine. It seems "regular" operating systems won't do.

Quote:

Sure. You've been amazingly transparent with this subject.

Quote:

Let them be. A forum is only worth its members' quality. I wouldn't waste a second of my time with such a company.

Quote:

Western companies just can't make business with China. It's the most dangerous place in the world, specially for intellectual property. Because they don't have (let alone respect) the concept of intellectual property. They have the culture of copying everything as a key to success.
Their methods can be described as the most aggresive competition that capitalism can imagine, with complete abscense of ethics on top. They don't care if a western company quits relations with them, because some other company will step in soon. Only based in their impossible to match manufacturing costs. In the end, all their greatness comes out of squeezing out their own poor people.

Quote:

At least you got something. I'm glad you did.

Quote:

I guess they are veeeery scared for losing freescale as a customer. What freescale should do is making this public, in order to save other companies from this screwing.
Well, wait, if those other companies were my competition, I'll very gladly guide them to destruction...

What intrigues me is seeing those freescale bosses apparently glad sharing presentation space with these people. Do you have an opinion on that?

Quote:

And can't be something done about this? But perhaps, when software makes this poor hardware useful, it would be too late. I guess that this sexy triple core MPC5121e is no better than the MPC5200 in the current Efika. But, if so, why did freescale develop it?

Quote:

Another useless toy computer. What a sad end for the MPC5121e.

Quote:

Roll on Efika 8610!