Let's also remember this is without ANY power management turned on.. the chip is running at top speed and the backlight is turned on full.. You can expect some very decent power savings across the board once the software is more mature.

With this power usage there is no need for power saving :-)
E.g. take an atom pc (ion platform) that does about 20w when idle.
However does the efika mx have a sleep mode? What is its power usage then?

There are lots of power management options for the processor and other chip components which can be automatically powered up or down. Notice how using USB ports goes up by nearly a whole 1W - features like USB selective suspend and suchlike will bring things down more giving you a better battery life as a whole.

Of course, the battery life should be fairly reasonable to start with, if you have everything on full power - not worse than the best Atom Netbook.

Shouldn't it last much longer with the same battery?

That's pretty much what I said, but of course a thinner, lighter Netbook requires a smaller, lighter battery. Even with the prospect of the power management not being fully enabled until a future software update, it should still get better battery life than the best battery life on any Atom Netbook.

It's funny to see that, again, hardware development is ahead of software. Much like in the days of the original Amiga, when hardware had so many new functions that software barely took advantage of them.
Power management has nothing to do with multimedia abilities, but it's a whole world of its own, and having those functions attended leads to very importante benefits that customers really want: More battery life. Plus the "I'm ecologist" badge that you can stick to your products.

We will meet that simply with backlight control and turning the wireless card off when not in use. We ARE looking for an EnergyStar rating.

Turning audio off when it's not playing audio, USB suspend, and other features which are really experimental in the Linux kernel (and some of which should be in firmware anyway) and high levels of unit and clock control inside the CPU core may come later, and simply improve on this.

Is Ion that power hungry? My meaurements with my Asus Eee 900A (N270 + GSE 945) yields ~ 10 W in low power usage. The 35W battery delivers power for 3.5 hrs operation. Of course that's in downclocked battery mode but the Atom is in daily usage not as bad as one may assume from the technical list specs. On the other hand it is no secret that the ARM blows away the Atom in cpu-power/Watt ratio. And that is quite good, but keep the numbers for the Atom correct (the 20W while idle are just wrong (for N270 + 945 GSE)).

While idle and in full power mode an N270 will draw far more than the 2W "Thermal Design Power" for the chip. The 945GSE is a nice Northbridge but, it is also not as low power as Intel's figures.

Embedded processors tend to be rated on their maximum power consumption - power supplies and regulators are very important to these systems so the amount of current needing to be pushed through each power rail has to be accounted for.

Intel processors are generally marketed to people who have a practically infinite supply of power - a 450W+ PSU or suchlike. They rate their chips on how good the *heatsink* has to be validated for to draw enough heat away for reliable operation. A 65W Intel processor will give off 65W of heat; the heatsink *must* quench this according to the TDP. It may draw closer to 90W though.

Check out Freescale's docs and when they say an MPC8610 draws 11W maximum at 1.3GHz, they mean that. It is also fairly indicative of system ratings as a whole - since the chips are so well integrated they can be placed on a board with little more than RAM to operate perfectly well (unlike Atom + 945GSE).

Intel's battery life is entirely reliant on dynamic voltage scaling and dynamic frequency scaling and clever standby modes, to the point that they have added newer and better processor sleep states in every major revision. ARM and PowerPC processors require much less work, but still implement (in certain revisions) some of these features - higher end ARM chips in particular will allow processor core voltage scaling and frequency changing, plus allow dynamic clocking of nearly every bus attached.

The source for my claim of the atom ion platform taking 20w when idle:
http://www.legitreviews.com/article/934/12/
E.g. the atom itself may be using 2w, but the rest is due to the chipset being used.

Remember the Atom N270 is 2W Thermal Design Power - that's how much the heatsink has to draw away to keep the chip operating at a safe operating temperature :)

i liked the original reply blizzard better ; )

it was also factologically correct re intel's power-quoting practices.

It's still there, just one instead of 10 :D



Indeed. I wish it was better known but it is on Wikipedia so most people probably think it is a lie :D

Measured ~12W while idle on Dell Inspiron Mini 10v, and ~19W peak power usage with 3D and HDD working.

BTW: the screen of this machine is actually worse, than the one bundled with the imx515 devel board. That is the same size, but has better resolution, no need to sit exactly in front of it, but can be read in an angle, nicer colors, etc.

But..

You can run Windows on it.. That is a big selling factor.
Maybe Sweden is alone, but Linux Netbooks are no longer
beeing sold. 90% of all Netbooks run Windows. People
gladly paid an extra 100 EUR for Windows.

Running Citrix over 3G is of course possible, but not
exactly pleasant. Consumers are used to thick clients
and not thin. Thin clients work for enterprise-use.