OLPC Set To Dump x86 For Arm Chips In XO 2
angry tapir writes with this excerpt from Good Gear Guide: "One Laptop Per Child is set to dump x86 processors, instead opting to put low-power Arm-based processors in its next-generation XO-2 laptop with the aim of improving battery life. The nonprofit is 'almost' committed to putting the Arm-based chip in the next-generation XO-2 laptop, which is due for release in 18 months, according to Nicholas Negroponte, chairman of OLPC. The XO-1 laptop currently ships with Advanced Micro Devices' aging Geode chip, which is based on an x86 design."
OLPC Set To Dump x86 For Arm Chips In XO 2
I'm sorry, I thought ARM is an acronym for Advanced RISC Machine (formerly Acorn RISC Machine). Why am I seeing it used as "Arm"?
Or is there something I don't know about the processing power of two of my appendages?
My work here is dung.
This is fine as long as they get the keyboard back. The single greatest Human interface tool (except for the SNES controller) not on a computer is really something sad.
My addiction: Arguing with idiots. AKA Slashdot!
It was angry, but it seems to have calmed down now.
W..w..W - Willy Waterloo washes Warren Wiggins who is washing Waldo Woo.
I'm uninitiated at the arts of ARM, and am too lazy to look it up.
If you don't know what AltaVista is (was), get off my lawn.
From TFA
"Like many, we are urging Microsoft to make Windows -- not Windows Mobile -- available on the Arm. This is a complex question for them," Negroponte said.
OLPC is in talks with Microsoft to develop a version of a full Windows OS for XO-2, Negroponte said. The XO-2 is still 18 months away from release, so "a lot can change with regard to Microsoft and Arm," Negroponte said.
I don't really see this working. Windows has run on Risc before of course, but almost no one ported their applications to any of the Risc platforms. And a top of the line Arm (a Snapdragon or Cortex A8) is still less powerful than a bottom of the line x86 (Intel Atom), so it's not like you can run x86 binaries at an acceptable speed through emulation, like Dec tried with FX!32 on the Alpha.
echo -e 'global _start\n _start:\n mov eax, 2\n int 80h\n jmp _start' > a.asm; nasm a.asm -f elf; ld a.o -o a;
The OLPC project is dying. Four years ago, you didn't have the netbooks. Now you do.
Shifting to ARM will simply ensure the death of the OLPC project, because being able to run real windows is an underappreciated benefit of x86.
Test your net with Netalyzr
I seem to recall seeing something awhile ago that Ubuntu is being ported to the ARM architecture. If the port is ready, using it would be a much better proposition than begging Microsoft to make a custom Windows OS for the XO-2, IMO. What would stop Microsoft from deliberately crippling the OS (and making it practically useless as a result) like they did with the starter editions of XP and Vista? Those were meant for the same type of market demographic as OLPC, after all.
"It is a denial of justice not to stretch out a helping hand to the fallen; that is the common right of humanity."
The guy is so pissed off at the likes of Intel he's driving the platform into a ditch. An ARM based client computer. May as well try and bring back the Amiga.
This is my sig.
I remember clearly that /. reported that Steve Jobs had originally agreed to license OS X to the OLPC project for free (as in beer), but that the offer was refused.
Since it is a well-known fact that Apple has had OS X working on an ARM architecture in the iPhone and iPod Touch for nearly 2 years now, it would seem a no-brainer at this point for OLPC to take Apple up on their offer.
...oh yeah, nevermind.
Damn.
ARM architecture will allow them to dodge harassment from Microsoft goons. They can respond "we'd LOVE to have a derivative of your OS on our machines but unfortunately we use ARM chips!"
Negroponte probably got sick of pigs heads on his doorstep and anonymous phone calls at 4am.
After logging in slashdot still does not take you back to the page you were on. It's been that way for 20 years.
They jettisoned Sugar, and they keep courting Microsoft. So sad. I wish the article would have explored the "open source" hardware concept. No idea what the heck that means from the article or for OLPC:
I doubt that the OLPC project is feeling warm and fuzzy about intel; but I don't think that that is the reason for ARM vs. Atom.
Thing is, to fulfill its objectives, the XO-2 has to be cheap, really cheap, to make. Atom based netbooks, even for the lowest spec models, in a highly competitive free market optimization process, have essentially failed to crack the $200 mark. Most are $300-$400. The OLPC guys really want less than $100. At this point, a $200 Atom netbook has already been cut to the bone, very little left(you might be able to cut out the ethernet jack and VGA; but you'd need to add the wireless mesh chip, and the more rugged case, it'd be a wash). Expecting that branch of development to halve in cost in the near future is pretty implausible.
That, rather than bitterness, is most likely the real reason. ARM is available, from a variety of vendors, at price/performance points that scale relatively smoothly from highish-end microcontrollers to modestly powerful laptop chips. x86 isn't(not yet, anyway).
Amalgamated Regional Militia, with jurisdiction over the Earth-Moon system. http://en.wikipedia.org/wiki/Amalgamated_Regional_Militia
Those who can make you believe absurdities can make you commit atrocities. - Voltaire
The fact that AMD is not planning a successor to the Geode processor used in the XO-1 probably influenced this decision, at least in part. In 18 months, there may not be any Geodes remaining.
Colin Dean Go a year without DRM
How is the parent modded 0 and the grandparent 4, insightful? Maybe I'll be informative if I mention that I think Facebook is being ported to ARM...
This post climbed Mt. Washington.
It work great. I run it on an nslu2.
http://en.wikipedia.org/wiki/NSLU2
How much power saving are we talking about here? It seems to me the LCD panel/backlight are by far the biggest consumer of battery power.
http://www.maxineudall.com/2010/02/should-economists-be-sued-for-malpractice.html
Atom based netbooks, even for the lowest spec models, in a highly competitive free market optimization process, have essentially failed to crack the $200 mark.
I don't think thats necessarily a "We can't make them for less than $200" but rather a "Everyone is going to think our laptops are crap if they are less than $200" idea.
Taxation is legalized theft, no more, no less.
So you'd get all of the disadvantages of Windows, while simultaneously loosing the only real advantage it has, plentiful software. Smart.
I do not think Microsoft will work on an ARM port, even something that translates x86 to ARM because the ARM processor is likely to be way too slow for this.
Nintendo DS runs on an ARM architecture. Maybe now we can run those games at full speed on another device? Certainly now with an emulator on this device there would be less translation and more instruction passing. Great!
Same goes for any other ARM-based processor device and emulation.
Wine will not run on this and neither will Windows. I am so fine with that.
What does such a move mean for backward compatibility? Aren't their applications already written with the existing OLPC in mind? I am afraid, it will not be as easy as "just recompile" to port some of them and those, who have already paid for theirs may have to pay again to be able to use them on the new hardware...
How did this get a score of 2? This just shows lack of knowledge of the OLPC project. The OS/GUI is open source. Sugar is free to download as are the Activities which are written in python.
Linux with ARM is superior to Windows with Intel x86 in this platform and target user group. I have worked with both and ARM is not powerful (comparatively) but is exceedingly power efficient.
The XO was a product of the western media lab -
custom hardware, FOSS and a western - constructivist - philosophy of education bundled into an all or nothing package for the third world education minister.
His alternative was the Classmate - a straight-line path to the higher grades, the trade school or college, the job market -
for the students who had a real shot at making it that far.
I'm a developer who ports Windows CE to devices. All day, every day. Teach classes on it even. Been doing it since CE 3.0. Currently on 6.0.
CE makes a passable embedded/PDA device, but there is no way in the world you'd want it on a laptop.
It just isn't made for that kind of a setup. No native compilers, no swap file. Expensive license restrictions. It's less like a computer and more like a gadget in terms of overall feel.
Use Linux instead.
Weaselmancer
rediculous.
Doesn't quite work like that. The DS has such a significantly different architecture that even though they use an ARM, everything will have to be emulated(with some dynamic recompilation) anyway.
I suspect power consumption has more to do with it (although given that battery cost is a significant cost of a system, reducing power may reduce cost too)
For every problem, there is at least one solution that is simple, neat, and wrong.
What does such a move mean for backward compatibility? Aren't their applications already written with the existing OLPC in mind? I am afraid, it will not be as easy as "just recompile" to port some of them and those, who have already paid for theirs may have to pay again to be able to use them on the new hardware...
Have you ever used, um, open source software before?
It's a piece of cake to port. I regularly run Linux on my x86_64 computer at home, x86 at work, and MIPS on my router.
Porting well-written open source apps is mostly just a matter of recompiling these days. It is really "just that easy" in most cases. That's why Linux has had flawless kernel-drivers-and-apps support for x86_64 for >5 years, while Windows still doesn't.
My bicyles
So, OLPC is switching because x86 costs a leg and an ARM?
Slow down, cowboy! It has been 4 hours since you last posted. You must wait another few hours.
Arch is my favorite distro, but Debian would be a better fit for these computers if you think about what they are designed to do.
... than they have been so far...
IANAL but write like a drunk one.
I don't understand Mr Negroponte. His natural allies would be the geeks moving the wheels of Linux.
The project would save substantial amounts of money and would provide a flexible, extensible machine for children to wander and learn.
In Windows you are straight-jacked to do whatever the licenses you are given allow you to do, and you have to pay for the privilege and enjoy it. It is like paying for a bad tempered dominatrix ...
IANAL but write like a drunk one.
How's your elbonian?
Yes, as a matter of fact, I have. And I not-so-humbly think, I know more than 99.9% of Slashdot users (including yourself) about such things...
Did I say "well written" somewhere? Or "open source"? I did not... Even if we are talking about a vendor porting their own app (which is "open source" to them), it is far from certain, that it is "well written". And even if it were, they are likely to charge a non-trivial sum for a port to a new hardware architecture. Which means, a person, who bought an older OLPC will have to pay for using the same application on the new hardware.
And, boy, is "well written" a high benchmark to clear! A big portion of applications breaks, for example, when people move from one major GCC release to another. If you told their authors, that the apps are not "well written", you would've been flamed to crisp...
In Soviet Washington the swamp drains you.
How the hell can they expect people to take them seriously on the shipping price of a unit when they say how much it is going to cost _before_ they have even chosen the CPU.
-- That which does not kill us has made its last mistake.
Ok I have 2 questions..
Firstly, WHY would Microsoft ever want to port XP to anything? This is an OS they have been desperately trying to kill off so they can get people buying their new ones. If anything is going to be ported its going to be Windows 7, and I personally cant see that going well.
Secondly, even if you HAD a port of windows on ARM, you'll get about the same number of Apps that you did when windows ran of the Alpha, ie, none. So why would you bother? "Being able to run all the normal software people use" is Windows ONLY selling point these days, and that nothing to do with the OS and everything to do with the developers.
Given the whole OLPC Linux to Windows switch fiasco, i'd be surprised if they get anyone seriously interested on helping them with a Linux port and you'd probably find a few people trying to actively hurt them for it.
Absolutely amazing idea (some may say world changing) but the implementation was very pore and badly managed. 2/10 would not shop again.
Technically, moving to ARM is a great choice. These CPUs are far cheaper. They also require far less power meaning that batteries and power circuitry can be smaller, cheaper, lighter and the handcrank give more page loads per crank. Also no need for cooling fans etc.
Given the way OLPC and similar projects move, the skeptic in me wonders whether this will actually happen. OLPC could just be "inviting" Microsoft/Intel to offer some good deal to keep them on x86.
Engineering is the art of compromise.
I've done WinCE diriver/OS work since WinCE1.0 days. In the beginning there was **no** ARM support, just MIPS, x86 and SH3. PowerPC came along a bit later and then ARM.
Some of the PowerPC code was directly lifted off the NT PowerPC porting effort (that got shelved).
WinCE has very little CPU-specific code - no more than Linux - but most of the rewrite was done to fit into a small system (few MB) where the traditional Windows bloatware just would not work.
Engineering is the art of compromise.
of what I believe OLPC is all about. Beiong an effective tool for delivery of education to children.
The success of this hardware in meeting this need rests on every component from the memory to the computing power, the battery technology and the display.
I believe ARM is a smart choice and improves the focus of this device on what the children need and not on what some idiot believes about training everyone to be able to use powerpoint.
Nullius in verba
Obviously, the apps included in the distribution would be ported by the vendor — this is not a concern regardless of whether the OS/GUI/bundled apps are open source or not. I was talking about third-party applications, which, presumably exist by now...
Even those written in Python will require testing. Those written in C or C++ will require some porting effort — and will not "simply work".
In Soviet Washington the swamp drains you.
It is a scam, they are using it as a ploy to sell laptops. It shouldn't even be called OLPC One Laptop Per Child, it should be called TLBOOGTCWWRHFOSI, Two Laptops But Only One Goes To Child Who Would Rather Have Food Or Shelter Instead.
Respect the Constitution
All the code converted/created for this new XO could be used to fuel the ARM ecosystem. Thus, people will stop depending on the x86 architecture (which will hurt both AMD and Intel). Offtopic: Negroponte is an idiot.
Arm is an excellent choice for such an implementation. The power profile is good and the number of units that can be crammed onto a wafer mean that costs of fab can eventually go below $1 per processor, with high volume. The future of computing is a place where general purpose CPUs primarily function as controllers and routers for special purpose signal processors and stream processor anyway and x86 processors are overkill for that.
This netbook and these development boards are based on the ARM architecture. You can also get ARM on a SODIMM form factor. And this little box looks nice.
now we need to go OSS in diesel cars
> ...vendor...
Where did "vendors" come into this?
And Debian already has tens of thousands of packages running on ARM, as well as ten other architectures. An ARM-based OLPC would be a routine port.
Warning: this article may contain humor, sarcasm, parody, and perhaps even irony. Read at your own risk.
Whoever sells (a.k.a. vends) the laptops is the vendor. They will port the OS and whatever software, that's bundled with the device itself. This is not the problem.
The problem is third-party applications...
In Soviet Washington the swamp drains you.
Debian already runs on OLPC, and Debian also already runs on ARM. Porting Debian to an ARM-based OLPC would be routine. Within a few months of the appearance of an ARM-based OLPC it would have official Debian support, with most of the tens of thousands of Debian packages supported.
As to closed-source OLPC software (if there is any): eh.
Warning: this article may contain humor, sarcasm, parody, and perhaps even irony. Read at your own risk.
5 posts later, you finally got my point...
In Soviet Washington the swamp drains you.
I have trouble believing that the XO-2 will be ready in 18 months - or that the generic x86 netbook of 2010-2011 won't be a viable competitor.
The problem with the XO wasn't hardware.
It was sales.
It wasn't a product the education minister was convinced he needed or wanted to buy.
The Windows option was an important - perhaps essential concession, a reminder that need to be talking to these people before you make the big decisions.
Since Windows is not their preferred os it probably does not really matter which processor they take. ARM currently is really the logical choice for netbook like devices if you dont need Windows. Their is absolutely no reason outside of needing some uncompilable legacy applications to go with Windows not even hd videos which can be covered by embedded gpus like nvidias offerings way better than to offload it onto the cpu!
The currently probably best netbook design would be an ultra slim convertible design with ARM processor and Nvidia netbook GPU!
No one has that yet, but I will be one of the first to buy such a thing once it is available!
I would be in favour of free software instead of OSX. It was offered for free for marketing reasons. They couldn't sell it anyways and market share also presents value to a company.
Btw. Apple is the new evil now.
Well... While I am not familiar with ARM-based Unix-like OSs, most of the time yes, it's just a question of recompiling the program.
And, when you use interpreted languages, you don't even need to go that far.
http://www.dieblinkenlights.com
Even when it is this simple, who is going to do this? For the customer — even if they had the source, this wouldn't be trivial for OLPC's target audience (poor people, relying on charity to connect to the Internet — if they had the knowledge to compile things, they would've been gainfully employed).
And so it would have to be the vendor/author of the software...
Yes... Except in cases, when the installer performs ``uname -a'' and tells you: "This platform is not supported." Yes, this is fixable. By whom — see above...
In Soviet Washington the swamp drains you.
Let's do it this way: you search, let's say, the Debian repositories for ARM architecture, and then you let us know if you can't find a given program you need that much.
Usually, package maintainers do a very nice job of porting programs. I would not be surprised if you could run Tuxracer on an IBM zSeries mainframe.
And you don't need to be dismissive on the users of OLPCs. People have the ability to learn, given the incentives.
http://www.dieblinkenlights.com