Intel does fondleslabs with Atom 'Oak Trail'

Eugh...

Now intel are trying to push x86 down our throats in the form of tablets.

It's awful for anything that requires battery life.

"... announcing shipment of Oak Trail and the 45-namometer Z670..."

Intel will never be competitive in the tablet/ultra-mobile/smartphone market until it starts giving its Atom/mobile processors and chipsets the same love Intel currently gives to its desktop and server offerings.

In the tablet/mobile space, it's all about performance per watt.

While Oak Trail does seem to be better suited to this market than the previous Atom kit Intel targeted for the mobile space, Oak Trail still doesn't seem to go "far enough" to be a viable competitor to the latest ARM SoCs out there.

Transistor junction sizes being equal, the x86/x86-64 architecture has **historically** needed to move a lot more electrons around to perform a given operation than an ARM core performing the same task.

I just don't see how Intel can be competitive in the mobile space with a 45nm part; it would have been better if they held off until Cedar Trail was ready at 32nm. And even that might not be enough, given that nVidia's Cortex-A9-based Tegra 2 is already out at 40nm, and will have shrunk to 28nm for the "Wayne" series release in 2012.

x86 vs ARM

One definite advantage of x86 processors is choice. Yes, end-user's choice over what runs on his device and how it runs there. In the past couple of years I've seen a plethora of ARM devices that are forever locked in with their old versions of firmware, unable to be updated. Even the most prominent players on this market, like Motorola and Samsung, "forget" about their older devices, leaving them without newest ROMs, not to mention the smaller companies like Archos. Having to either keep using the defunct Android 1.5 or replace the device, makes the tablet into really nothing more than a game console or a media player of olde. To add insult to injury, patch cycles on those devices are either slow or non-existent, forcing the majority of users to use devices with known and widely exploited security holes.

x86 architecture leaves the choice of OS to the user, while most of available choices for x86 are more mature, easier to upgrade and more frequently updated than their ARM counterparts. Granted, neither Windows7, nor latest Linux flavors are particularly well-suited for solely touch-screen experience, but I dare to hope it's easier to bring touch functionality to a versatile, mature and feature-rich OS, than vice versa (need I to mention lack of cut&paste on certain mobile OS until recently?).

As to the argument that software written for those OSes is not suited for small-screen touch-based form factor - this is true, user-space applications for such devices will have to be rewritten... but same goes for any ARM OS as well - and recent history shows that the road from 0 to 100,000 applications is shorter than a year.

So I, for one, welcome our battery-saving, x86-running, generic OS-supporting overlords... err... fondleslabs.

And please, please don't forget the digital stylus. Taking notes with finger-smudging is not fun.

64 bit ARM

Not needed.

64 bits only needed for multiple VM. Cheaper & better performance to have multiple ARM each with own RAM like Transputer.

32bit Intel CPU can do 16 GByte RAM efficiently via extension scheme supported as long ago as NT 4.0 Enterprise.

64 bit CPU will be less power efficient for normal applications that 32 bit at same speed and geometry.

If intel wants to compete with ARM they need a Single Chip 32 bit Atom that connects direct to RAM, Flash, Radio chip (WiFi/Bluetooth/FM radio etc), ethernet, LCD/Touch screen, camera chip, USB and SD card connector directly.

For 3 years Samsung has had ARM sandwich Soc with all I/O built in and RAM + Flash as two sandwich chips in same BGA module. The Atom is too hot for that.