Hi Matt (wow this is confusing :)

Aura is what we call our firmware, yes, it's basically Open Firmware - the same as you got on old Sun boxes, PowerPC Apples, and brand new OLPC laptops. Aura has some features most OF implementations don't, though. I'm not really sure this is the best place to go into the finer details, but its what makes the PC graphics cards work, among other things.

You need to put your kernel on some kind of bootable device, be that hard disk, cdrom, flash drive, connected via ATA or USB, or on a network (TFTP and other methods).
ISO9660 with Rockridge but not Joliet, MSDOS FAT12/16/32, ext2/3 (try not to enable any weird features), BSD UFS (original, not UFS2), Amiga FFS (original, not FFS2 with long filenames) and Amiga SFS.

Partition maps would be MBR (but no extended partitions), Amiga RDB and BSD slices. It will also happily detect what is called a "dangerously dedicated" disk - basically a disk with the filesystem and no partition map, because this is how low-capacity USB keys and CDROMs are.

To the firmware, a partition map is just a filesystem that splits the disk into several other filesystems.
No :)

The firmware does provide, however, a Forth interpreter. The Linux guys have pushed some Forth scripts into the kernel (a bad idea from my perspective) and you can get a decent Forth script which exposes further functionality from this site (Device Tree Supplement, or as we end up calling it, efika.forth).

You can drive any device you like as long as you do not need fine-grained timing or interrupts. With clever coding you can even write new filesystem and partition map support - in Forth of course.
You would be better off using a CompactFlash card or something. If we have you 256MB of NAND flash on the board, you would want 512MB. If we gave you 1GB you'd want 2GB. We do not want to play the "keep up with the size of your root filesystem" game - you can play that yourself, with a $10 CF card, SD card, USB stick or some other flash device.
Well, you could always replace the firmware, but would it actually improve the situation? I don't think so. And this is why we don't actively support people fiddling with our firmware. If you touch the contents of flash with something other than our firmware updates, we reserve the right to charge you for the repair when you return it to be debricked :)

_________________
Matt Sealey

I am looking for the same device for many months now :-( I will probably buy some QNAP device or MPC8349 board directly from Micetek.

thanks for your reply matt, sorry i didnt respond earlier.

is it possible to add hfs, hfs+, and zfs to the supported filesystems?

is it possible to add mac partition and efi partition table support ?

i still think an option for corebios would be nice, or at least enough nvram for the kernel, that way if the partition map or file system is not supported the kernel could be loaded and then find the file system its self.

anyway, even without those features i am still hoping to have a new powerpc for my desktop soon :-D

matt

indeed nwhitehorn, we (or at least i ) have been advocating such wireless intigration, user programable FPGA intigration to distinguish the product from all others, and include a divx/xvid, AND an AVC decoder codec (even an AVC encoder if the chosen FPGAs up to it infact, i advocate it should be) and video in/out for a very long time now.....

while waiting, these other products have in many ways, began catching up on this all inclusive low power yet capable general purpose PPC/Altivec MB proposal, and just today they started selling this mini-itx x86 atom (TDP = 4W), for £42.00 UKP no less, (thats around EU 52, $82)

http://www.tranquilpc-shop.co.uk/acatal ... oards.html

http://arstechnica.com/journals/hardwar ... ard-for-83

stick a cheap 11N wireless card and USB Video Encoder key in it and your set to fly.....

its been clear for a while, as the options became more available elsewere,the posters here have dropped to a trickle, and thats a shame, one day perhaps we might see the mass market MB iv advocated but alas not yet and not today, all that lost potential profit and time :(

im not sure were you might market those, although you obviously think there is one, but for the remaining readership here and the Blog ,dont you think those will be FAR to little FAR to late?

how can you/we possibly compete with the Intel Atom powered mini-itx board at £42.00 UKP and on the scale they will be producing them now?, it was possible 12/24 months ago perhaps , but i cant see how it can today, now they are in production and on sale (soon everywere)... :evil:

about the only growing market i cna think of right now is to make it into some kind of Hardware for a plug in end to end network encyption key system to help combat the likes of the unlawful interception for profit Phorm and the Phormettes NebuAd,sell it into the masses of end users market, but even then any of the PPC/Altivec based SOC can be made to fit that space so much better.

http://www.cableforum.co.uk/board/12/33 ... dated.html

If you want an Intel Atom board, buy one, we're not going to stop you.

There are several architectural tweaks in the Atom chips though, compared to the other x86 microarchitectures and RISC designs like the MPC8610 - there is still a lack of integration (even Via's ARTIGO kit has a seperate northbridge, despite their CoreFusion platform offering it all in one package) and Atom has been designed to run with in-order execution putting some burden on the compiler - however while RISC architectures have had the benefit of this compiler support for a long while now, in-order execution hasn't been in a major x86 design since the days of the original Pentium.

Getting extremely high performance stuff done on an Atom would be difficult. At 1GHz it runs slower than most x86 processors you may have seen - enough for Intel's "Nettop" platform, web, email, cloud computing kind of stuff, but not really extensible to high performance graphics or anything in gaming beyond Peggle. It's not going to be a SETI@home box, or be encoding movies. It is not going to play HD video too well.. it's multithreaded (full multithreading, not hyperthreading) but not multi-core.

There is ample room to compete with a sub-GHz processor based on Power Architecture, with higher integration, similar power envelopes from a holistic point of view (the entire board and not just the CPU), and somewhat higher performance. What I am not sure of is price, but we know Intel sell cheap and in high quantity, this is something you lack the ability to gauge when looking at prices on AVnet for 10 or 20 chips in a tray, and only reach when you order 20,000 of the things. It should hit reasonable prices if there is a device to be made from it, comparable to the Atom.

I would even say, if you have a chip like the MPC5121e or MPC5123, at 400MHz it gives you enough power to do the things Intel's Atom is meant to achieve on Nettops and portables. The 5200B was $23 when we started making the Efika and is far less now. The 5121e has a similar price. You can have a whole board for $50 and in high quantity, have a $100 device with screen. You won't see this with an Atom solution where the chip and support modules would - if that motherboard is anything to go by - be $80 to start.

_________________
Matt Sealey

[quote="Neko"][/quote]

I guess what popper wrote was not just a "song of praise" for the x86 architecture, but more about the difficulty to communicate the benefits about ppc/cf compared to solutions from "marketing monsters" like Intel.
During the past years Intel had nothing to offer in the area the ppc/cf is majorly positioned. Now they have interesting products and their well known brilliant marketing (the do it mostly with plain force I guess. They are so enormous huge, they can drive the market). Thus I really guess the problem is really more about communicating the benefits of e.g. Freescale's current ppc processors.
F. ex. I know what is possible on an Efika, I know that it simply can rock, that it is small and reliable and that a *solution* is more than just counting and comparing MIPS. But exactly that is sometimes difficult to communicate. I just read a thread where ppl were highly excited about the new 500MHz Via PX5000EG board. It surely is a nice one, but does it offer more than the Efika(2)? I doubt it. Plus the Efika is availavle since long time.
Anyway when rolling out the Efika2 you must keep boards like the PX5000EG in mind and put a lot of energy into communicating the benefits. It'll be a difficult job.

Well we have never even tried to encourage people to buy a Power box instead of a Dell or so. I don't think it would be a good idea to even try - so I don't really get where you're coming from. I don't think comparing ColdFire or low-end Power chips to "Intel's latest desktop reference design" is prudent either. It is very easy to look at a store and see the price of Intel's latest desktop reference design and it's supposed performance, and then say "Power is dead because it cannot compete"

Power has advantages where you don't need to care what processor type it is - when you don't need to run Windows Vista on it, for example, who cares if it's x86 or Power? The advantages will be in the design of the device (be it portable, laptop or desktop). Current x86 low power solutions, especially Intel's Atom, have design concessions for that purpose, be that in-order execution or other architectural changes. Power Architecture has never really had to make such concessions to be low power, as evidenced with the Efika (which has a 130nm e300 core and a single chip design) and newer chips which exhibit higher integration than x86 designs (Via are the closest to a real x86 SoC with CoreFusion, but seem to have ditched it for the PX5000EG in favor of a return to a separate CPU and the VX700 northbridge) and lower power by default (the MPC5121E has been tested at less than 3W with graphics, disk and AXE unit being stressed. With power management it is roughly the same as the original Efika - but without the need for a seperate Radeon/GeForce graphics controller.
We do set our sights on comparing with the Via solutions as they target more embedded, far lower power designs, but jostling against Intel on the desktop (such as the "Nettop" design) will never and has never been a goal. There is just no point in trying to be first choice on the desktop.

Rather we should concentrate on devices such as the ultraportables Atom may find it's way into, where running Linux means the CPU is not a design concern, and in fact, you would get a cheaper design out of readily available Freescale technology (as mentioned, something like the MPC5200B or MPC8349E are ~$20 chips, this is not something you will see for the Atom+945 combination any more than you ever saw a sub-$200 Centrino branded laptop). It is far easier to differentiate on price and features on a handheld device - the Atom+945 combination does not do a great deal that an MPC5121E can't do either.

The MPC5123 Starter Kit will be a good comparison against the first Atom UMPCs and MIDs Intel are pushing. The advantage of a single-chip design in board layout and complexity is a TRUE advantage, even if you cannot clock them in the 1GHz range. Do you really need 1GHz to check email and browse the web? I am not sure about that.

_________________
Matt Sealey

what about missing popular technologies in LinuxPPC ?
(ie : flash enabled browsers)
Don't you think it can be quite a disavantage when thinking about such devices (low cost laptops) for the masses ?

I understand that gnash is a usable alternative -though i don't use flash at all, so I don't know myself.

Apart from flash, what other popular technologies are missing from Linux/powerpc? I'm not arguing, just curious.

Oh... that is nice. The atom bites the dust there as that on uses over 20W due to the chipset used. I am really impressed by that 3W.

What i am missing from ppc linux is graphics hardware acceleration.

You got me slightly wrong. I wasn't comparing to Intel's desktop solutions. Also note that I am aware of the the Power benefits. It was more about recognition in the general public.
And there the current hype is Atom and the likes. It is *said* to be super-duper efficient and that's what Joe Average believes.
And currently Joe Average is massivle intrigated by the Intel marketing soldiers. And this public recognition makes selling the Power based designs not easier.
Anyway I am still convinced that with the the 512* and the 8610 there are highly interesting chips available and that there is a market for them.

Full ACK

ACK ?

full ack == full acknowledgment

[ there are network packages that send an "ack package" if something went right. geeks soon began to use this short form of "I agree with you" in chats. it soon spread all over the internet chat rooms ;-) ]