Moorestown. Digging a little deeper. | UMPCPortal – Ultra Mobile Personal Computing – http://www.umpcportal.com/2009/09/moorestown

Maybe the Moblin SDK is the place to look for clues.

No, I’m saying that driver info I have seen indicates to me that GMA500 is not being used in pinetrail, Intel is instead using a derivative of its own in-house graphics. I’ have seen two drivers, one of them is specifically for Menlow and Moorestown, indicating to me that Moorestown using the same graphics as Menlow (maybe clocked quickly). The other driver I’ve seen is for Intel’s inhouse graphics, a driver for 3100,4000,4500 etc etc, which includes new definitions for “Pineview -G” and “Pineview -GM”. as an aside it also contains references for Clarksdale and Arrandale. It does not have any reference to Menlow or Moorestown. Conclusion would be that pinetrail is in the same family as Intel’s existing integrated graphics, and Moorestown is in the same family as Menlow.

The 535 can do about twice the polygons/second that the 530 can, and the OMAP3 has its GPU clocked at 110MHz.

Even clocked at 400MHz, that’s still only ~7 times the graphics power of a smartphone.

Drivers also indicate to me that Pinetrail is NOT SGX graphics, but rather a derivation on Intels in-house graphics techonology.

A future SGX may use the DX10.1 hardware available to higher end SGX (54x) which would add very little if Intel choose to keep clock rates the same to reduce power consumption.

SGX with the right drivers should be able to handle 720P, however the trend for MIDs is to use a DSP to do it for less power. 1080P DSPs are starting to appear in cheap Chinese MP4 players, so we should see them soon in in MIDs and UMPCs.

It’s dang troublesome not to be able to delete or edit my previous (or next)posts…

So far it’s not clear if Intel will try to implement turbo boost on CPU and GPU. As you may know Intel’s turbo boost just overclocks a CPU (or GPU) to a certain value that is measured outside the processor (the temperature).
Unlike previous overclocking this type of overclocking is more accurate, and efficient, allowing the processor to be operating within certain thermal values.
This would mean that certain notebooks will work slower than others, depending on how well they are able to substract heat from the CPU/GPU.

I’m not sure if Intel will implement this technology in their upcoming Atom processors, but it would sure do the netbook sector good!

The upcoming graphics chip will playback most 720p video flawlessly. And will be aimed for that resolution. There’s speculation that if Intel will allow these chips to be implemented on notebooks with larger screens and resolutions, that the graphics might display 1080HD video, but that’s not so certain yet.

Apart from reduced latencies between RAM/GPU CPU/GPU and RAM/CPU, the CPU will most likely support 667Mhz DDR2 RAM, or even faster.
The GPU will increase performance with increased RAM speeds.
Also the GPU will be manufactured at a smaller die (45nm instead of 65nm). This will allow higher clockrates, while still having lower powerconsumption.

Lower powerconsumption of the overall system will mean lower powerdraw, longer battery life, but also a cooler system; which in it’s turn can be translated to higher overclock ability (as long as there is a good heat spreader available), but that will be the responsibility of the netbook manufacturers, not intel’s.
Concern has been uttered that the GPU/CPU package will probably get hotter, since both GPU and CPU are on one package (as where before they where separated, and a better heat dissipation was possible.
A valid argument, which we’ll see if this could possibly have negative consequences on the performance of both chips.
Most of the time, either a CPU (office tasks, encoding, program loading…), or GPU (watching flash based websites or video’s) is used. Only in a gaming environment where CPU and GPU are both ballasted, ‘might’ one see a reduction in performance, but that has yet to be determined when the product gets released…

I also highly doubt they will plug a GMA500 in there, it is more likely for them to pair this netbook with a more suitable, slightly modified GPU, which (knowing Intel) despite being on the same package with the CPU, will most likely not outperform regular notebook graphics.

Intel also made a comment when comparing their product to an ION platform, that their product will be focused on battery life, and will not compete with the GeForce ?9300/9400? chip series. Most likely the ION will outperform the upcoming graphics processor for the next gen Intel Atom CPU’s.
But a second meaning could be ‘powerwise’. Intel kind of leaves everyone in the dark what they meant with it, since it’s perfectly possible for them to manufacture a similar performing GPU, just with a way lower TDP thanks to High Metal K manufacturing process, and their 45nm design.

All these are but speculations, but they would make sense from business perspective.
We can only think it would be natural that the upcoming platform will not only overcome the Intel Atom netbooks for sale today in power consumption, but I believe in performance as well.

I think Intel spend a lot of time finetweaking their design to find a balance between powersaving, and the required power.

Then there’s the newer, more aggressive sleep modes which intel is working on. No say if they will be for netbooks, but I surely hope so!
Intel’s working on getting not only newer sleep states that require less battery in a low power state of the CPU, but also they are working on saving power through their GPU’s.
Different power states will lower the CPU even further down to 500Mhz (instead of 800), and in 2D applications will reduce the GPU’s core speed as well.
Also Intel is looking into ways to shut down more area’s in the GPU from power when not in operation, or not needed.

These powersavings will result in an additional battery life, and cooler running system, which in turn could benefit Intel’s Turbo Boost.
This all is good news, but unless we’re sure this information applies to the upcoming platform we’ll be little with this info (at least for now)!

Even if not now, Intel must be working to see how it can apply these improvements in netbooks!

With current 6 and 12 cell batteries, we might be looking at netbooks with a real-life 12-15hours of battery life per charge, not factory time, but in real life (eg: when browsing or listening to MP3)!

The added cost of the processor, and Windows 7 (since XP will be discontinued sooner or later), will probably force manufacturers to go for lower capacity batteries (eg: 2, 3, or 4 cell), with about the same battery life as current notebooks, and an increased pricetag of between $50-150 above current $299 notebooks.

These are my speculations… They’re no final words though…
My 2 ct.

The device, like a mini HTC shift, would flip between Windows and Android/Winmo.You’d get a device with a full desktop whilst having hours of talk time and instant on/off capability – beating ARM at its own game.

Those boring linux/unix/macos/symbian smartphones are selling loads and maybe Intel want to compete on that level, just with the extra codec/features support that x86 brings. Even if linux underplays the potential of Moorestown, maybe Intel are afraid at the extra effort it takes to get the mass-market to accept Windows on 4-5″ 800×480 screens.