Intel Atom - RIP?

[Plantje]

I have owned an Asus Vivotab RT. On one hand I loved the device! It was thin and light and did almost everything I wanted! On the other hand I hated the device! With my Surface 3 I can install applications that I need because Windows has no apps for it (that may exist for iOS and Android)

No, with a device limited to apps available in the store Microsoft created a device that has to compete with iPads and Android tablets, but with a store that has way less apps! I will never trust Windows RT again!

Edit: I can imagine Intel making this decision. If you look at the Samsung TabPro S you can see a very thin and light tablet with passive cooling holding a Core M3 CPU. That paves the way to ditch Atom all together, right?

 

[bronsky]

... If you look at the Samsung TabPro S you can see a very thin and light tablet with passive cooling holding a Core M3 CPU. That paves the way to ditch Atom all together, right?

I may be the only human being lamenting the demise of the lowly Atom. In only 5 years, small tablet devices have evolved from using single Atoms like the Z570 in the Slate 500 to the Quad Core Cherry Trial X7-Z8700 in the Surface 3 that was powerful enough to outperform the 10w SU9300 C2D in my Thinkpad X301 laptop.

Still, I agree that Core M and x86 is the way to go with the Surface line. Not ARM and RT. If MS wants to use WP10 for 8" tables and below, more power to it, since I'm sold on my 5.7" GN5 as a mobile note-taker and a 10" tablet for the rest. Samsung created a nearly perfect platform for the M3 and showed MS that it can be done. It had its flaws but overall was a nice effort on Samsung's part.

 

[GreyFox7]

I lament Intel's failure to produce an Atom capable of effectively competing with ARM, but IMO they hamstrung Atom's performance and by doing so prevented it from succeeding. Perhaps they truly couldn't make it perform better per watt but I point to the only slightly higher wattage Core m's performance as proof they could have if they wanted to.

By some measures, not all, the top ARM SoCs already meet or exceed the performance of core m parts although in a broader analysis the core m's still prevail.

The competition however has shifted to core m vs ARM. Core m is still functionally deficient compared to ARM in terms of necessary functions for a Tablet or Phone SoC with built in wifi, LTE. Image DSP etc. and uses more power as best I can tell.

By 2020 when Intel is ready to compete again (according to their statements) Win32 should be quite the ancient legacy app model if not rightfully dead and buried. We will be relying on UWP by then which as I previously stated would work equally well on ARM and does so today. One is here now, the other promises to be here later, should you believe.

TBH I'm not sure UWP is catching on but AFAIK it IS linked to the future of Windows while Win32 is fading away and needs to fade much faster. The sooner people get their heads around that the sooner we can fully embrace the future with all that is available. I don't think there really is a future that relies on Win32 any more than there is a future for Windows XP, it's the two halves of the titanic. Great in it's day but it's day has past.

Note: I use Win32 to refer to Desktop Apps, high DPI unfriendly Apps, power unfriendly Apps. These will not survive on servers of the future either.

In the future Windows on ARM could be a viable platform, it depends on availability of UWP Apps just as the future of Windows on x86 depends on the availability of UWP Apps. Without UWP Apps there can only be a lingering, Legacy Windows environment dwindling down into the sunset. Yes it could live for some years...even decades... just like Unisys and the rest of the bunch (Burroughs, Univac, NCR, CDC, Honeywell) but that's a fate worse than death.