Is ARM Ever Coming To the Desktop?

First time accepted submitter bingbangboom writes "Where are the ARM powered desktops? I finally see some desktop models however they are relegated to "developer" models with USD200+ price tags (trimslice, etc). Raspberry Pi seems to be the only thing that will be priced correctly, have the right amount of features, and may actually be released. Is the software side holding ARM desktops back? Everyone seems to be foaming at the mouth about anything with a touch interface, even on the Linux side. Or are manufacturers not wanting to bring the 'netbook effect' to their desktop sales? Are ARM powered desktops destined to join the mythical smartbook?"

Comments

Look on eBay

[jimicus]

Look on eBay for an Archimedes.

They're rapidly becoming a collector's item, but they were on the desktop in 1987.

Re:Look on eBay

[mountaineer76]

Totally agree, just compare RISC OS 2 or 3.1 to the equivalent Windows version in 1988 or so, Acorn was streets ahead. Still have a 200 mhz RISC PC sitting next to my desk, it's a nippy little beast boots in a few secs or so....

Archimedes

[rve]

It was on the desktop first. I was a kid, not terribly good with money, and it was expensive, so I just missed out on being an early adopter.

EOMA Initiative

[lkcl]

it's a long story, but i've been working to get ARM-powered desktop machines and laptops into the hands of free software developers for some time.

one of the key problems are that the chinese and taiwanese factories have absolutely no software expertise whatsoever. some guy decides he got caught out by the USA and UK Governments placing embargos and tariffs on imported clothes a couple years back: his business was affected, so he goes "i know, i'll diversify, i'll make tablets, those are popular". so off he goes, he gets supplied with a GPL-violating Android OS right from the word "go" by a limited number of Chinese ODMs who are having a really hard time keeping hold of their software engineers, and it just goes downhill from there.

the other problem is, as can be seen from the insane amount of money spent by the openpandora group, that case-work for laptops etc. can well be in excess of $100,000. that means that anything like the "pegatron netbook" has to be bought in volumes of 250,000 and above in order for the R&D costs to be amortised over a reasonable period.

this is where the EOMA initiative comes in: http://elinux.org/Embedded_Open_Modular_Architecture/PCMCIA

by reversing everything on its head, and getting free software developers a modular architecture which _could_ be dropped into a mass-volume product, the tables are turned: those Chinese Factories can be supplied *by us* - Free Software Developers - with a completed ready-to-ship OS.

so, yes there's a board which is available that is similar in size and function to the pandaboard, origen exynos board, beagleboard, IMX53QSB etc., but unlike those boards, by complying to the EOMA/PCMCIA Open Standard it would be possible to literally drop that hardware-software combination straight into a mass-volume product, with the development effort of the required motherboard being nothing more than a low-cost 2 to 4 layer board that even KiCAD, Eagle or gEDA could do.

one key part of this strategy is to leverage arduino-like boards, like the leafpad Maple:
http://elinux.org/Embedded_Open_Modular_Architecture/PCMCIA/MiniEngineeringBoard

anyway i think that's enough for one slashdot post. bit of background and some additional links, here:
http://www.openhardwaresummit.org/forum/viewtopic.php?f=5&t=502

Re:EOMA Initiative

[lkcl]

ok - i'm pleased to see a response here: there are several points that are good, and some are, when you look closer, turn out to be unrealistic for mass-volume so-called "embedded" products.

1) the first is form-factor. this is great! yes, one of the options being considered is to have a standard Mini-ATX/ITX motherboard (into which an EOMA/PCMCIA CPU card can be plugged, at the back). there are several embedded companies that produce Mini ATX motherboards as standard, for their "modules", so it is not a new concept, it is in fact a proven one.

2) the second is connectors / interfaces. if you look right across the board at the very latest ARM processors coming out *right now*, you can count the number of Cortex A8 and Cortex A9 systems (as well as Marvell's "ARM-compatible" range of processors) that have PCI-e on the fingers of one hand.

i'll say that again.

the total number of modern ARM processors with even a 1x PCI-e interface is *below* 5 (five).

now there do exist some Cortex A8s (e.g. the OMAP35xx series) which have a HPI bus, onto which you could put a PCI-e "PHY" chip as it's called, but the total number of companies doing actual PCI-e "PHY" chips is, also, very very limited. typically, any company which has PCI-e PHY interface is a "Fabless Semi" company that gets bought up very very rapidly by the likes of Mentor Graphics, Synopsys and so on.

3) the third is the sheer overwhelming disparity between the ARM CPU's power consumption and the average PCI-e-based GPU's power consumption. the absolute ABSOLUTE lowest power consumption PCI-e-based GPU i could find is one from SiS, it's an older 65nm CMOS process, and if you ramp its speed down to the absolute lowest it will go without keeling over, it uses 6 watts. SIX watts!! you wanna connect a 6 watt GPU up to a 0.5 to 1.0 watt processor be my guest!

4) Multi-layer boards at ATX/ITX form-factor are expensive. if you have the CPU on-board the Motherboard (rather than being on a separate card), you then are forced to have the most complex part - the CPU-to-RAM interface - push up the number of layers required for the *whole* motherboard. by contrast, if you do the CPU-plus-RAM as a separate tiny, tiny board, just presenting its interfaces (SATA, ETH, USB, I2C, RGB/TTL etc.) via a simple connector (e.g. PCMCIA 68-pin) then you've just saved a fortune on the cost of the main motherboard because the main motherboard PCB can be done as an ultra-low-cost 4 or even if you're really lucky or a very good designer as a 2 layer board.

5) The power requirements of standard PCs are 10 to 200x larger than is actually needed! 500 to 1000 watts i mean for fuck's sake that's just insane. these ARM processors, the fastest most powerful one available on the market right now (sampling) is the NuSmart 2816, and that uses _two_ watts (shock horror) at 2ghz. wow big fucking deal. why on god's green earth would you want to match a 2 watt CPU with a 1000 watt Power Supply?? the entire motherboard would probably need a big resistor just to draw enough current in order to convince the PSU that nothing's wrong! i'm not joking about that - i'm dead serious.

6) The level of integration on these so-called ARM "embedded" CPUs is so high that it's really not worth the effort. present a USB bus, present an Ethernet port, present an SATA socket, along with an HDMI out and maybe even VGA, you're done! ship the damn product out the door, it cost you $35 to make! don't believe that price? just look at the cost of the RaspberryPi - it's doable. the irony is that for $35 of the "upgraded" RaspberryPi you can get an 800mhz Cortex A9 with an AML-8726-M (single core) for the same price. and the same size. credit-card-sized. you have to ask yourself: why would you _want_ to fit a credit-card-sized computer into a 12 x 15 x 8in "Desktop" case?? :) why not fit it into a 4in x 5in x 0.75in box, instead?

so, whilst on the face of it, fitting into the "standard" - i'm going to go further than that i'm going to

What's the problem with the TrimSlice?

[earls]

Developer only? What is that non-sense? The TrimSlice ships with Ubuntu ready to use. ~$200 for the feature set is a steal, IMO. Not happy without a Dell logo or something? What's the problem with the TrimSlice?

Search a little more, like the Efika

[jonsmirl]

http://www.genesi-usa.com/products/efika
Smarttop $129 thin client
Smartbook $199 laptop

They run Ubuntu and are based on the Freescale iMX51.
They are far more powerful than a Raspberry PI.

Freescale i.MX515 (ARM Cortex-A8 800MHz)
3D Graphics Processing Unit
WXGA display support (HDMI)
Multi-format HD video decoder and D1 video encoder (currently not supported by the included software)
512MB RAM
8GB Internal SSD
10/100Mbit/s Ethernet
802.11 b/g/n WiFi
SDHC card reader
2 x USB 2.0 ports
Audio jack for headset
Built-in speaker

10.1" TFT-LCD, 16:9 with LED backlight, 1024 x 600 resolution
Freescale i.MX515 (ARM Cortex-A8 800MHz)
3D Graphics Processing Unit
Multi-format High-Definition hardware video decoder
16GB Nand Flash
External MMC / SD card slot (up to SD v2.0 and MMC v4.2)
Internal MicroSD slot
802.11 b/g/n WiFi (with on/off switch)
Bluetooth 2.1 + EDR
2 x USB 2.0 ports
Phone jack for headset
Built-in 1.3MP video camera
Built-in microphone
Built-in stereo speaker

Re:Why?

[Bert64]

A few dollars a month for a desktop...
A few thousand dollars a month for an office full of desktops?

The average office worker doesn't do a lot with their computer, and has been doing much the same thing for years... The only thing stopping them from using 10 year old hardware is modern bloated software which is intentionally incompatible with older versions.

There's no reason that the average user's needs couldn't be fulfilled by a low power machine with equivalent processing power to a system from 10+ years ago, with power hungry x86 systems being relegated to the small niche of power users and certain classes of server.

(in short, watch what x86 did to Sparc/MIPS/Alpha/Power, attacked from below)

Re:Why?

[Bert64]

Exactly, no solution fits all... Your needs are specialised, so you will occupy a niche of people who will continue to buy highend workstations...

For the vast majority of people computers became powerful enough for their requirements many years ago (aside from increasingly bloated software trying to mask that fact), and they are concerned about price, running cost (ie power usage), noise and that the machine is not an eyesore, and even more so are the companies who buy hundreds of desktops for their employees and don't want to buy a noisy, expensive, large and power hungry workstation for someone who's sole business use for it is to write letters.

Re:Tabtop momentum building

[hairyfeet]

I'm sorry but.....why? WTF would you want ARM on the desktop? Are you living in a mud hut in Zambundi and don't have any electricity to spare for a desktop?

Lets be honest folks, the big selling point of ARM is how cheap it is on batteries. Well guess what you do NOT need when you are inside? Why that would be a battery! See that plug on the wall right in front of you?

Cycle for cycle x86 stomps the living shit out of ARM, it just uses more power to do so than most mobiles can afford due to the fact we haven't had a real breakthrough in battery tech in ages. Well that and the fashionistas at Apple have made iSliver batteries the "in" thing in which means you have to power the thing on a battery the width of a tic tac. I don't care if you put 8 cores on the thing, a bottom o' the line AMD quad, even the low power AMD quads, will stomp the living shit out of ARM. drop in an i series and it isn't even funny how badly it gets stomped.

Like everything else it is about using the right tool for the job. ARM royally kicks ass in mobile, embedded, and in places where you need a device that'll take milspec levels of abuse due to the fact you can run it fanless. X86 kicks ass in desktop and laptop where you want more performance and don't mind giving up some battery life for it. But ARM on the desktop makes about as much sense as stuffing an i series into your phone, that is none at all. The majority of code out there is x86, even on Linux x86 outnumbers ARM code by a pretty wide margin. So unless you just really really REALLY want the Droid version of Angry birds on your desktop it just seems more than a little stupid to be running a mobile chip in a place where you are right beside a plug in.

Re:Tabtop momentum building

[jd142]

I'm sorry but.....why? WTF would you want ARM on the desktop? Are you living in a mud hut in Zambundi and don't have any electricity to spare for a desktop?

Lets be honest folks, the big selling point of ARM is how cheap it is on batteries. Well guess what you do NOT need when you are inside? Why that would be a battery! See that plug on the wall right in front of you?

You know, it's just possible some people might want to conserve electricity. Or even shave a couple of bucks off the old electricity bill. Just because you can use a resource, doesn't mean you should. I have running water, but I don't just leave the faucet on all day in case I might want a glass of water.

I don't know, but if you had one of those little portable solar cells, could you just power an arm laptop anywhere?

Re:The OP is talking about red apples, not greenap

[Sancho]

Smudging the screen isn't the problem. The problem is holding your arm up for long periods of time, or the repetitive motion of raising your arm up to touch the screen. That's not something most deskjockeys are going to be doing a lot. It's horrible for ergonomics.

A standalone touch pad doesn't have that problem.

Most phones are held in the hands with lowered arms, hence it's not a problem for those devices.

Hell, laptops were being sold with touch pads as the primary pointing interface. Not much different from a desktop, really.

I don't think any particular feature of touch pads was the perceived problem. But then, you seem prejudiced against Jobs, so my reply is likely pointless.

cost and integration, not just power

[YesIAmAScript]

ARM licenses IP. Intel sells chips.

If you license a core from ARM you can put it down on a chip, then put down your other logic (north/south bridge, interface logic like USB) on the same chip. Then you can end up with your entire system on a chip.

With Intel you have to buy a CPU, buy a north/southbridge. If you want custom interfaces beyond that, that's more chips too.

So the net effect is that the Intel-based system uses more chips and that means it costs more, uses more power and is larger. Using more power means you need to put in a larger power supply, that costs more. If it's battery-powered, that means it needs a larger battery, that costs more. Larger in and of itself makes something more expensive to make as it requires more materials. And then it being larger means it costs more to ship from where it is made to the customer. And then finally every increase in costs also means more increase in on-the-shelf price because you not only have to cover the higher costs, but the OEM and retail margins on the costs.

The next effect is that ARM devices will be cheaper to buy and to run. And in the case of portable devices, more sleek too.

This may not matter to some customers but to other customers lower costs means a lot.

Performance is an issue. We have ARMs already in the pipe (dual-core ARM A15) which have sufficient power for most uses and ARM will certainly have even faster cores later.

I see a strong future for ARM in laptops and in home computers. No, not in tower computers but those make up a shrinking part of the market already.

Finally, as others have said, be careful about agreeing with Steve Jobs. He's a consummate liar. Just because he says he doesn't like touch for the desktop doesn't necessarily mean much. It means Apple doesn't deliver touch on the desktop today, but it doesn't necessarily mean anything more. Apple could flip on this at any time like on the video iPod.

Re:Tabtop momentum building

[m50d]

ARM is admittedly easier than many things - you're still 32-bit and usually don't have endianess issues. So the main problems are sizes/alignment of system data structures - things like (network) host information, or even something as simple as date/time vary. Of course you should be using the types opaquely, but it's very easy to accidentally stuff a time_t into an int, and even if you are being careful you can miss implicit assumptions you're making (one of the problems I saw was code not dealing with an array in the network host data structure being of length 0). In this particular case the biggest source of problems was multibyte character handling differences (it was a Japanese game). When porting old code away from x86 you can also get a whole slew of floating-point related bugs, because double on x86 is really x87 extended double (I only encountered one minor problem in this case, it didn't do much floating point, but for other software it would be a bigger issue). Also see the recent glibc memcpy vs memmove issues for something that can come up even on x86-64 vs x86-64.

Once code's been ported to two or three architectures these problems don't tend to come up any more (because the first couple of changes reveal all your implicit assumptions that could be broken), and that's true of a lot of open source projects. But any code that's only ever run on one platform will have portability issues. You don't have to take my word for it - try it yourself, pick a random project that doesn't release non-x86 builds off sourceforge and try and build it for arm.

ARM was developed as a desktop CPU

[Alioth]

ARM was originally developed as a desktop CPU, and it was on the desktop - it's been and gone.

ARM stood originally for Acorn Risc Machine, it was developed by Acorn because they couldn't find an adequate processor for what they wanted to do to follow on from the 6502. Many of the CISC chips at the time (mid 1980s) had very poor utilization of memory bandwidth and poor interrupt response (Steve Furber in one of his talks recently on the development of the ARM - he's one of the two people who developed the first ARM CPU, pointed out in particular the National Semi 32016 (IIRC) that they were thinking of using, until they found out the multiply instruction took over 100 clock cycles and could not be interrupted).

They also wanted ARM to be low power, not to make their new line of desktop computers energy efficient particularly, but because they needed it to be cheap so the computers could be affordable. If they could get it under 1 watt, they could use plastic packaging instead of ceramic packaging which reduces the cost by an order of magnitude. They had no tools for estimating power, so they just designed *everything* on the chip for low power. When they got the first samples back from the fab, they were blown away when they found the chip consumed 0.1 watts - they had massively overachieved.

We had the Acorn Archimedes in school. IIRC, it had an 8MHz ARM and it could emulate - in software - an IBM PC with VGA graphics faster than the original IBM PC ran. That's how much faster the ARM was at the time compared to anything else around. Without needing to be in a ceramic package.