I’m sitting in front of an ExoPC which is now referred to internally as the Exo-1 because they’re looking for new hardware for the next generation ExoPC device Product X
There’s a very open an interesting post up on the ExoPC discussion board about new hardware they’re testing. It’s clearly an Oaktrail-based device but they’re not giving the ID away at this stage because there’s some negative points being aired. It’s a thread that has already prompted me to post something on our own forum about Oaktrail and I’d like to bring it up here. Maybe it’s useful for Exo PC. Maybe it’s useful for readers.
Obviously if you’re going to bring similar looking / acting hardware to the market as a new model, it needs to be better than the existing product, right? If there’s one thing that’s worse than the previous model, it confuses customers and causes rumblings in the owner community. While the battery life on the Oaktrail-based devices might be better, the performance isn’t. An Atom N450-based Win 7 product will outperform a 1.5Ghz Oaktrail equivalent in most, if not all, departments. Even if you bump the CPU up to the new Z615 at 1.9Ghz, the GPU isn’t going to be better than the previous generation. It’s a problem for anyone that’s got a 10 inch Atom N-series in the market. Where do you go?
The problem is, the N450-based tablets aren’t exactly zippy and are not the best at battery life. Fan noise, weight and size are an issue so some progress needs to be made there. Oaktrail is not the solution here and until Cedar Trail or more Ultrabook-like designs (E.g. EP121) come along, there’s no solution.
The way to make for ExoPC to make a second device is to move to a new form factor and / or a new operating system. Changing the OS is a problem for ExoPC because most of their business is wrapped up in their overlay software development, for Windows. Personally, I don’t use it but if it’s their bread and butter, it’s important.
So why not keep the user interface and apply it to a 7 inch model? The 7 inch Windows tablet market is small, yes, but so is the 11.6 inch Windows tablet market. In fact, there’s less competition in the 7 inch space, more importance placed on efficiency and at 7 inch, Windows needs an overlay more than it does at 11.6 inch Using Oaktrail allows you to make a different hardware product for a new market and it also brings a bridge to other operating systems that could boost hardware sales. Meego is an interesting option for an Oaktrail tablet and if your skill is in user interfaces, you could turn something round quickly and make quite a differentiator in the Meego market. Acer turned a lot of heads when they presented their 10 inch Morestown-based Iconia M500 tablet that’s due to be available in Q4. Don’t forget the Windows 8 mobility path either. If you’re working with Oaktrail now, making the move to Clovertrail and Windows 8 should be easier later. It will make an even better 7 inch tablet platform and gives the company a product to focus on should Windows 8 kill the ExoPC software overlay business.
ExoPC should do two things. 1) Look at the 7 inch slate space where user interfaces become even more important. It’s currently a fairly empty space if you consider pro-mobile users and although niche, could allow ExoPC to become the king of it. Better to be king of a small niche than a bit-player in a large one. Asses Oaktrail devices and think about Meego, Windows 8 and even Android which we’ve already seen running on Intel. Look to break out a new product with a new customer base. 2) Take time to find a better product platform for the existing (10-12 inch) ExoPC market. Mid 2012 perhaps? Sandy Bridge-based tablet designs could be cheaper then as Ultrabook-style manufacturing matures.
Of course there’s a risk. You’ll need to investigate why the 7 inch space is empty for a start but the Galaxy Tab 7 sales (millions of them) say something about the form factor. There’s also the risk that someone else has seen this opportunity too. The Viliv X70 Slate design still exists (although the company appears to be going under) and we shouldn’t forget the ECS 7 inch slate design either. WeTab appear to be going through the same process so stay tuned here, ExoPC, for the latest news! (And good luck, of course.)
Unfortunately the initial impressions of Oak Trail is right. It’s very poor. It underperforms Pine Trail devices by a noticeable amount. It’s battery life claims don’t seem to hold up either. I wonder what’s up.
Why would anyone expect Oak Trail to even match Pine Trail?
Oak Trail is intended for very low powered, along with very low thermals, devices and by its nature that means it can’t give as much performance as the higher power consuming Pine Trail.
Oak Trail can go into devices that Pine Trail is simply too power hungry and too hot for.
While we have yet to see proper battery run tests on properly optimized systems to really see how good or bad its run time efficiency is.
Regardless, Intel is stuck with Oak Trail until they can get Cloverfield out, but that should be out just before Windows 8 official release. Though the significant improvements aren’t expected till they get out the 22nm chips.
While we should see Cedar Trail for the 10″ range tablets, for a higher performance solution. Unless Intel makes a tablet version of the Cedar Trail, like AMD did with their Z-01 (modified C-50), as then they could more likely get it into at least 7″ as well.
Why do you keep spreading this BS? Sandy Bridge uses significantly less power than Westmere which both are on the 32nm process technology, and the latter has much better performance. Oak Trail is a “Tablet” oriented part, and paired with the MP20 chipset it was supposed to be even in a Smartphone! Pine Trail came in early 2010, meaning they spent 1 year making Oak Trail with minimum power usage reductions but with slower performance.
Do you know what Atom needs 22nm for? Power reductions? I don’t think so. By 22nm, Atom will need a significant performance overhaul to keep up with ARM enhancements. Meaning, zero power reductions on 22nm. They have to make power reduction catch ups ALL at 32nm, when they get the proper SoC process ready.
BTW, Cloverview IS a Tablet version of Cedar Trail, just like Oak Trail was one for Pine Trail.
No BS, first try to get the ATOM series straight. Oak Trail is a Z-Series ATOM. Pine Trail is N-Series! They are different! Different power optimizations, different GPU, different drivers, different product applications, etc.
The full details of Cloverfield have yet to be released but by all reports it is intended to replace Oak Trail and even maintain the 3W Max TDP, which is also used in tablets, and that means it’ll also be a Z-Series ATOM. Since that line is specifically intended for low power usage and that’s what allows it to be used in more mobile devices than the N-Series, which uses more power and generates more heat that limits what it can be squeezed into.
Even 32nm Cedar Trail starts at 3.5W TDP, about 2W less than equivalent Pine Trail, while Oak Trail has 3W and down to 1.3W TDP offerings. So there’s more than Clock speed separating them!
While 22nm will reduce power requirements, that’s a natural result of going smaller and simple physics. It’s just a question of how much they will increase performance as to whether that will balance out or give us a net gain in run time. But features like Intel 3D Transistors, among other enhancements, will boost both efficiency as well as performance and it’s clearly Intel’s intent to reduce power usage to better compete with ARM power efficiencies.
Like you said Sandy Bridge uses less power, but the 22nm Ivy bridge will use even less and Intel won’t put many of those enhancements in the ATOM line until they go to the 22nm Silvermont.
Right now Cedar Trail is still using the same architecture as Pine Trail and only gets the lower 2W lower TDP between a combination of the new GPU and the 32nm size reduction. Mind also they are boosting performance a little too by increasing average clock speeds and making all chips dual cores.
While Oak Trail is still only offering single core solutions, to further emphasis the difference that reducing the manufacture size can provide and could mean a performance boost when Cloverfield comes out if it maintains the same TDP range but improves efficiency like Cedar Trail does compared to Pine Trail.
Yes, it would be nice to have a 7″ windows device in that space. For Windows users, the Galaxy Tab is just an interim device. It does some things well, but it is not the familiar windows environment. while the wishing continues, using the Tab as the base device, how about half the thickness and half the weight, double or triple the processing power and the battery life. Does that mean Windows 8 on an ARM? Hope Exo Pc follows Chippy…
I agree. 7 inches is the sweet spot. It has always been. 5 inches is nearly too small to be useful and 10 inches is nearly too big to be ultra mobile. Moreover, lose the Atom. Intel is being moronic in bringing all of the driver hassles of embed computing into a PC-targeted product. That already hasn’t worked, and I don’t see what is going to change in order to fix the problem. However, why not AMD instead of Sandy/Ivy Bridge? AMD currently makes a low watt APU which offers the smoothness of two cores, advanced graphics processing, and a redundancy of open and proprietary drivers. A 7 inch slate or clamshell with such processing heft would make a great UMPC.
For 5-7 inch systems that simply not going to happen, at least until AMD goes 28nm next year. Present AMD Fusion’s lowest powered offering still uses too much power and generates too much heat to go into anything smaller than 10″ systems.
Even then they had to create a tablet optimized version of the C-50 with the Z-01, which didn’t reduce the TDP as much as they were originally shooting for.
Meanwhile, we’ll see in hopefully less than 2 years how things goes once Intel brings ATOM to 22nm and revamps the ATOM architecture with Silvermont and AMD’s consequential follow up shortly after that.
ah yes, 2 years.
Intel has been saying “next” year for years now.
remember when the original UMPC’s came out (before Atom) & Intel promised us that Atom was going to be the breakthrough we all wanted? yet everybody just stuck it in netbooks instead.
hey everybody just WAIT, I super promise that next year all the problems will be solved. don’t worry when next year comes we at Intel won’t be telling you to wait again for another year.
lol, I gave up on Intel years ago. it’s NEVER going to happen.
That’s exactly what I was thinking.
Looki.
You gave up on Intel for UMPCs I assume. You think the AMD solutions are better? Would like to hear why. Or is it ARM you’re looking at for solutions?
Chippy.
Versions of the ATOM did get put into UMPC’s, UMPCs just didn’t do well in the market and the performance of the ATOM was only part of the reason. Netbooks were just more successful.
While the eventual 22nm Silvermont update is definitely coming. Intel can’t procrastinate any longer and has to get serious or they will lose out on a big portion of the market.
Part of the problem was the original ATOM line was set to a 5 year cycle, but now it’s on a 2 year cycle like their main offerings. So it’s catching up and is now more a question of how good their offerings will be once it does come out, vs the competitions solutions.
While it’s not like the competition hasn’t run into delays either. AMD Fusion took over 2 years longer than they originally planned for example but it has arguably proved to have been worth the wait.
I’ve been looking for a decent win 7 slate for some time but heat, battery life and processor/graphics mix just isn’t right yet and i am wondering if it ever will be in the foreseeable future.
There are a few areas left untapped. The win 7 slate market is begging for a decent transformer type device. Like the asus transformer in terms of how it works rather than the acer w500. You can pack in a better processor and be content with say 4 hours battery in the slate part considering the keyboard battery can boost the overall battery life to 8-10 hours.
Win 7 slates tend to be most useful and popular when they have a digitizer in them too.
The thinkpad tablet (android) is an interesting option as it builds on the consumer tablet craze and adds many useful productivity elements missing in devices like the ipad especially the digitizer. I’d like to see more devices like this. I’ll probably go for one of these until the aforementioned transformer device is built by someone.
The value of windows surely is in all the regular applications it can run which are best used on a larger screen eg 7″ netbooks failed … even 8.9″ ones faded away. 7″ win 7 slates will therefore be very niche. 7″ android tablets will be more popular due to the touch gaming you can do, better battery life etc. I don’t see 7″ win 7 slates being a worthwhile avenue for any company to exploit.
. . .unless they take cues from the tablet world? Windows 8 could really solve some issues here. Not just with UI, but with cor support for things like always-on, scheduled time, idle power drain.
There still will be the issue with processing/graphics power vs battery/heat issues which will be very difficult to solve on regular windows even with enhancements such as the ones you mention. This will most likely take years if the manufacturers even bother to go this route. EG Video streaming on a low powered HP slate lasts 2-3 hours but can go 10 hours on an ipad… that is a lot of ground to catch up.
Also there has to be massive development with touch applications which hardly seems to be forthcoming for regular windows. Especially important for a 7″ device.
Regular win 8 will be pretty much similar to the present win 7 but with some much needed UI enhancements. It will remain niche/for power users. I can’t see many people switching from android and ios tablets to regular windows even when 8 is launched (maybe the arm version if it offers some decent support for regular windows applications but i’m pretty clueless as to what that is looking like). Little tweaks just aren’t enough to overcome the windows tablet sufficient power vs battery/heat issues.
Issue is significant but isn’t quite as big as you’re thinking. Intel and AMD do have to catch up on making efficient low power devices but ARM has to catch up on making capable higher end power devices.
Continued competition is pushing them towards each other and starting to address the strength and weaknesses of both up till now.
For example, ARM is still 32bit… They’ve only started adopting 64bit memory management… Many are still using 45-40nm manufacturing, even the upcoming Tegra 3… The big companies at least are still following Apple’s pricing, while the cheaper ones give less and many are still using Cortex A8 and haven’t updated to A9, let alone A15.
Remember, the demand for performance isn’t just being driven with the idea of running Windows 8 by the end of next year but all OS platforms as well. Especially with the more recent push for gaming and performance finally starting to rival game consoles. The traditional role of mobile devices will start changing and that will pose challenges and opportunities to all involved.
By the time most of this is addressed Intel should already be updating ATOM to 22nm, with AMD not far behind with their own slightly even smaller die shrink. Technology like Intel’s 3D transistors should help close the power efficiency gap along with the improvement provided by the manufacture size shrink. Especially if ARM continues to be behind the curve on the manufacture size shrink.
Power efficiency also has to do with the OS and Windows up till now isn’t made for maximum power efficiency, so not just a matter of hardware efficiency, but that is part of what they are going to address with Windows 8.
Comparisons to the iPad are also misleading because that doesn’t take into account the performance difference or the fact that devices like the iPad are equipped with much larger batteries, in comparison to the size of the rest of the system, and don’t need to support as much as x86 systems with fewer ports, etc.
The original iPad even used the PowerVR SGX 535, which some might remember Intel also used in the Intel GMA 500/600. It’s still a question of what Intel plans for the 22nm Silvermont but Cedar Trail will be using a little improved version with PowerVR SGX 545 and overall TDP for the whole system will be 2W less than the previous Pine Trail.
So while there is still a long way to go, I wouldn’t say the gap is insurmountable.
I think this is a good way forward for windows slates and arm slates:
http://www.jkkmobile.com/2011/05/punk-this-module-from-cupp-computing.html
Yeah, the Cupp Computing PunkThis is a very interesting solution. I was planning on getting one myself, but they were suppose to come out in July and now we’re in August and no word from them yet.
Hopefully just a delay but doesn’t bode well… While jkk seems to be on a extended vacation and he was suppose to help promote it…
Good points in the original post but I disagree about to 7 inches. It really depends on how you want to use the tablet. There just isn’t enough in the ‘over 10 inch’ space and it remains valuable for pdfs, comfortable html, comics and apps with complete interfaces. It’s the form factor I’m holding out for re: tablets.