OLPC is excited to announce that a refresh of the XO-1 laptop is in progress. In our continued effort to maintain a low price point, OLPC is refreshing the hardware to take advantage of the latest component technologies.
This refresh (Generation 1.5) is separate from the Generation 2.0 project, and will continue using the same industrial design and batteries as Gen 1. The design goal is to provide an overall update of the system within the same ID and external appearance.
In order to maximize compatibility with existing software, this refresh will continue with an x86 processor, using a chipset from VIA. The memory will be increased to 1 GB of DDR2 SDRAM, and the built-in storage will be 4 GB of NAND Flash with an option for 8 GB (installed at manufacture).
The processor will be a VIA C7-M, with plans on using one whose clock ranges from 400 MHz (1.5 W) to 1GHz (5 W). The clock may be throttled back automatically if necessary to meet thermal constraints.
The enabling chipset is hot off the fab line, the VX855. This single chip provides the memory interface, a 3D graphics engine, an HD video decoder, USB, SDIO, and other system interface and management functions, in a low power and small footprint package. One change induced by the chipset change is a move from AC'97 to HD Audio. This brings higher sampling rates and allows an upgrade to a stereo external microphone (and DC sensor) input.
The CaFE chip is being retired, and replaced with an external Flash management controller, possibly one of the low cost SSD controllers currently being tested. The camera will now be tied directly to the VX855's video capture port.
The network interface will be upgraded to an 88W8686, which will halve its power dissipation and move it to an SDIO interface (further dropping the power consumption). The current goal is to locate it in a removable module, allowing its replacement for repair.
It will remain powered while the laptop suspends, waking the laptop if a packet addressed to it arrives. It is likely that early production models will not directly support 802.11s (i.e. forwarding mesh packets while the interface is asleep), but we are working with Marvell on several different 802.11s solutions.
Gen 1.5 will continue with the existing display, although OLPC is working with PixelQi to try to improve the brightness and efficiency of the screen. The DCON is retained (even though the VX855 includes much of its functionality) as it provides the low power interface and the timing controller functions for the existing display.
Overall, the target is to match the Gen 1 XO-1 in power consumption while making aggressive suspend easier, and in price (while changing to components which are more likely to decrease in price). It is likely that both goals can be met.
We also expect the Gen 1.5 machines to ship with an OLPC 8.2.x software release, modified to support Gen 1.5's new hardware but otherwise unchanged from the current production software release and compatible with our current software in the field. Gen 1.5 machines will be deployed in environments already populated by Gen 1 machines, so seamless software interoperability is an important goal.
Early versions of the hardware (bare board) should be available for driver development at the end of May. A larger number of prototype laptops (several hundred) for software development and testing will become available around the end of August. The OLPC contributors program will be the preferred way of requesting a Gen 1.5 machine for testing your software for compatibility or development. We hope to use the contributors program to ensure Gen 1.5 support for the wide variety of application and OS solutions created for Gen 1.0.
We're excited to be finally able to make this news public. While members of the technical team have been working on this for several months, it was not until last week that we could with any certainty say that we were going to refresh the hardware and what that refresh was likely to be. We're now committed to this project and look forward to working with you to make it happen.
---John, Ed, and the OLPC Tech team.
Stay on top of the XO-1.5 laptop: subscribe to OLPC News via RSS Feed, Emails or Twitter.
Wow, the 400 MHz to 1GHz speed stepping is amazing! No more performance issues now. And when I see "HD video decoder", should I be thinking the days of XO's jumpy playback are over?
don't worry, the software always has found ways to keep pace with the hardware to keep performance an issue. Remember, nobody ever needed more than 640 Kb...
True, but if it is still running the same OS as the 1.0, it should have no problem chugging along. Personally, I'm hoping for a Windows 7 version. According to the VIA website, it's very capable of running it.
It's a pity its not using ARM arch. It would even ban windows :)
What the...?
Is this for real?
This is not another joke Wayan is pulling off right?
Hm... I am a bit conflicted about this... and not sure what to make of it yet...
Well, if they can do this while maintaining the battery life, it seems like a great idea. But thats a whole 600MHz were talking here. More then 2 times what was used before. But then again, new manufacturing processes have improved a lot since the geode (and, if i'm not mistaken, the geode was aging when the XO1 first came out, so that would make it quite the old little processor). Lookin' forward to another G1G1 lol.
http://blog.laptop.org/2009/04/17/announcing-the-xo-15-updated-hardware-for-the-xo-laptop/
Confirmed.
Please remember this is the start of engineering, not the finish.
Actually, Seth, I pulled this post from the devl list as confirmed, a full 3 hours before blog.laptop.org published the same post.
You gotta get up pretty early to beat OLPC News!
This is great, but I'd rather like to see ways to make a cheap, cheap XO-1 than going for feature creep.
The XO-1 is great hardware, if it just 1) were available for purchase and 2) had curricular content, we could actually sell it retail all over the place, and with what is gained by economy of scale selling to schools in the developed world, we could even lower the price down to benefit the developing nations. What we have with the XO 1.5 might be better hardware, but I am not sure it makes for a better educational model (uh, we're for education, yes?)
This is great, but I'd rather like to see ways to make a cheap, cheap XO-1 than going for feature creep.
The XO-1 is great hardware, if it just 1) were available for purchase and 2) had curricular content, we could actually sell it retail all over the place, and with what is gained by economy of scale selling to schools in the developed world, we could even lower the price down to benefit the developing nations. What we have with the XO 1.5 might be better hardware, but I am not sure it makes for a better educational model (uh, we're for education, yes?)
I remember back when Negroponte promised no to change the XO-1 hardware, but to just drive prices down to $100 or even $50. But it seems like he too fell to bloat.
The only way to beat feature creep is with volume and time. Just time doesn't cut it.
Sure, you could probably build a machine that rivalled the XO-1 in performance today for cheaper. But it would mean using ARM - a complete redesign, and OLPC does not have the people for that today. With the resources they have, this is an impressive refresh. And as the announcement says, these components are more likely than XO-1 to drop in price with time, so even if you cared nothing for performance, this is good news.
hey wayan i got a question...a few weeks ago NN made a little bit of an announcement that the price of the XO 1 was going to drop substancially...any idea what happened with that?
Abe,
No clue what happened to the price drop hint that NN gave. I was expecting this to be it (a "big thing" I heard rumored) but its not the price drop. Not sure if that's another announcement or if this is it.
This is very useful for OLPC to continue an XO1 line now since the AMD Geode fanless processor manufacturing is being discontinued.
What I thought sounded interesting is that VIA might have integrated the DCON system features inside of their new processor? Is that DCON feature on a VIA processor a worlds first? DCON, being that functionality that actually can turn off the main processor completely within milliseconds while it is not needed, for example when nothing is moving on the screen in ebook mode.
Stronger standby mode sounds great. My experience with the XO1 is that standby doesn't really work as envisioned does it yet? The vision was to automatically suspend on closing the lid, automatically forward WiFi Mesh packets for more than 6 days on a battery while in standby mode and waking up from standby mode on incoming packets such as VOIP calls, instant messages, incoming email and other such features.
This could I guess help lower the XO1 series price to a $150 with mass production. While the XO2 series should start at $75 hopefully soon since it's going to be based on the much more optimized ARM processors with something I guess which could be based on a very customized Android Linux.
$150 1kg for XO1.5 X86 with a 7 hour battery, $75 500gr for ARM on XO2 with a smaller 20 hour+ battery. Both should be totally compatible so eventually it won't even matter if some kids have one model and some other kids have the other within the same school or within the same country.
Better embedded video support is really great. I think they are saying that it's 1ghz but it's downclocked to save energy and limit the amount of heat that it will output to keep the whole thing fanless. As long as they can avoid using Intel, everything is going to be fine and OLPC is thus also going to lead and guide the next wave of X86 based netbooks. Using this new VIA instead of Intel Atoms in netbooks gives manufacturers freedom to configure and customize it the way they want to, to optimize the costs and provide the features that the customers are asking for, bluring the lines and making of the netbooks some real laptops.
"Both should be totally compatible so eventually it won't even matter if some kids have one model and some other kids have the other within the same school or within the same country."
I wish you are right, but software bloat follows hardware bloat and gains on it in ways it's not even funny. I can make my 2.5 GHz machine freeze about as well as I was freezing my 486 of yore.
It will require incredible levels of self control not to attempt to "add features that the new hardware can finally support".
But then, as I said, I hope you're right.
now since the AMD Geode fanless processor manufacturing is being discontinued.
AFAIK AMD is just discontinuing design in Geodes, not discontinuing manufacturing. Which given the chip has been virtually untouched for 10 years, isn't anything new. (AMD did some work to relaunch it in 2005... 4 years ago.)
In fact I think it's actually a Cyrix design that AMD bought from National.
I'm still impressed OLPC left AMD and went with VIA. Shows how deep the cracks are in the original OLPC supporting corporations. Those that are left anyway, now that Nortel pulled out
Maybe it's just me and my conspiracy theories, but maybe that's one of the reasons that OLPC surprised us with this news - because they had to do something. They were loosing investors, and they needed something to attract new ones and keep current ones.
Daniel,
OLPC is a marketing machine like no other, and I am sure this was pushed out now because they were looking for a quick hit. Personally, I would have hoped for a mid-week announcement so we could all digest the news over a workday vs. the weekend, but its nice to see there is movement at 1CC.
How flattering. ^_^
Which is clearly why we sent a tech-heavy email to devel@ (and the blog three hours later) instead of using the old and ineffective method of a press release.
Really, we wanted to let the dev community know what's going on, which is why it was sent to them.
"A marketing machine like no other"? It is to laugh. Please, Wayan, turn off your cynicism for half a second when OLPC does something good. I have as many criticisms of NN as the next guy, but this announcement represents a lot of work from some skilled people. Instead of parsing announcement dates, congratulations are in order.
Jameson,
That's not cynicism, I am in utter awe of the OLPC marketing juggernaut. They've had more than 9 lives already and are still a huge force in technology for education. And that's without a real marketing team or strategy that I can see.
Intel doesn't want OLPC, the whole concept is contradictory to Intel's whole philosophy of One Laptop Per Rich Badly Advised Overconsumer.
Intel Atoms are designed for very specific reference designs which Intel has very strict control over. They are built to keep prices inflated as high as possible, even considering that the whole netbook movement is about being cheaper.
You can see how specific 10.2" 1GB RAM, 160HDD, WinXP reference designs are being pushed out by dozens of brands at the same time, they are all basically based on the same Wintel approved reference designs. OLPC cannot fit into that control, OLPC needs very specific customizations made. Intel would be the last one to accept turning off the processor 98% of the time using those DCON systems, it basically demonstrates how the main processor is not supposed to be the main part of the system.
... and the hardware enhancement efforts for my XO-1 start to become orphaned. Damn!
Wayan... if an 1.5 G1G1 is announce OLPCNews should start a program to solicit 1.0 G1G1's from the readers and donate them in bulk to a specific OLPCNews project.
When XO-1.5's come out, I think we'll see XO-1's coming back from deployments. What ministry wants to use yesterday's technology? They'll all want XO 1.5s.
Have you ever been to Peru? Or anywhere in the third world? (Almost nowhere in Mexico counts since soon after 1994.)
Jameson,
I spend my life in the developing world. And what I see is small NGO's taking recycled & donated computers, not Ministries of Education on hundreds of thousands of computers scale. And here, we're talking high Ministry officials as buyers. They're not going to go for end-of-life equipment.
The key word for the education ministries is 'FREE'
What 3rd world country wouldn't want free educational laptops ?
Chris, you should know better - there isn't a such thing as "free computers". No matter the initial hardware cost, there are always significant expenses in computer rollouts.
And I don't think there will be too many XO-1 donations, unless OLPC offers to upgrade all G1G1 donors to XO-1.5's
First, congratulations. This is great news, and to all the engineers involved, good job.
Now. This is could be an opportunity for OLPC to become important again. But just improved/cheaper hardware is not enough to achieve that. They also need better distribution. In particular, any school, anywhere, needs to be able to buy XOs in classroom-sized lots, with under 3 month lead times for firm orders.
I understand that there are real problems with a non-profit selling a commodity. And I know that NN has been very clear that he'd rather keep the non-profit business model than face those problems head on; I was in the room when he announced the killing of the "give many" program, and I can even understand his logic. But at this point, it's life or death. Either these machines become available for grass-roots pilot programs, within a few short months of the assembly line running full strength - or OLPC will continue limping along as it is currently.
Users are life, and pilot users are the best kind. Look at how last year Nepal was making much, much more software with a few hundred machines, than invisible Mexico was with 50K of them. That is the energy we need now, and OLPC needs to find a way to not obstruct that goal.
James,
I'd love for you to explore the line of thought around challenges a nonprofit has in selling a commodity. I see NGOs doing this all the time (donated/recycled computers for example).
Might you also explore this topic in a Guest Post (vs a comment) so we can analyze the subject in a more visible setting.?
"the energy we need now, OLPC needs to find a way to not obstruct that goal"
Jameson, you're right in many points, as in mentioning how much "this announcement represents a lot of work from some skilled people". And you are especially right in that it wouldn't be really that essential that OLPC took the lead in software, content and community (though they should have, and a lot of people hoped they were). In many cases it would be good enough they just were not obstructing. At least with letting Sugar go they did take a step in this direction, and as to how the community got gutted, oh well.
About VX855's HD playback:
http://www.youtube.com/watch?v=kTIUTz4zcJA
Reading through, it looks like the chassis, monitor, and other bits will remain the same. Has anyone considered making the XO-1.5 form-compatable, so that old XO 1.0's can be retrofitted with the new motherboard?
This would be in keeping with OLPC's goal of green computing - otherwise the old XO's would just have to be tossed and replaced with new ones (or foisted off to others - which is essentially the same thing).
I understand that the developed world is not the marketplace for these units, but they are the visible side, and the one that many in the press see. If I could take the 10 or so XO's in my community and upgrade them, it would be a huge boon.
hey, i'm confused about the mention of a 3d graphics engine...will the XO 1.5 finally be able to pump out some basic 3d graphics? I ask this because i thought the XO 1 has some sort of graphics capabilities as well, but apparently, when inquiring about it, this was really just about being able to display basic images or something and not "graphics" as i was defining them.
hey...question...i would like to try and build support for the idea of a butterfly keyboard being used on an XO (at least to show the strengths of such an idea) . Where on the site would I post that?
Geode has accelerated 2D but not 3D graphics -- to display anything three-dimensional software has to render it using CPU, then GPU part of Geode chip can be used to do minimal 2D operations on it. While software is capable of doing it using the same API (OpenGL), it's usually too slow for interactive applications. For example, OpenGL gears demo and 3D screensavers work, Armadillo Run is too slow to be playable.
C7 has no GPU, however its chipset (VX855) includes 3D-capable GPU. I don't know how well it's going to work because VIA has poor reputation when it comes to Linux driver support for its graphics adapters, but apparently is making an effort trying to improve it.
Hola!
I also want to be able to upgrade my original XO-1 to XO-1.5!
Maybe someone from the DC users group could get their hands on a new model and switch the motherboards? Then, we could get the new motherboards installed on our old XO's!
One detail that's interesting about using VIA here is that their latest CPU ("Nano") is pin-compatible with the C7 family of processors which means that the XO-1.5 design could potentially have a long-term upgrade path.