http://liliputing.com/2013/02/nvidia-tegra-4-android-benchmarks-faster-than-a...    (by a factor of 3-4 no less)

I mentioned that Apple gets an extra year to work on ARM stuff before it gets released to everybody else.   Apple used to use this extra lead time this to pack lots of WOW into their yearly product releases.

Since Steve Jobs died, not a lot of WOW has been coming out of APPLE on a regular basis.   Google is now WOW'ing us more than APPLE, and that seems to be a continuing thing now.

Now, the Nvida Tegra 4 chip is out and it was in test products and it was on test boards and the Tech reporter people at the World Mobile Conference at Barcelona Spain were allowed hands on time with the products where they were allowed to run whatever benchmarks they wanted to on these production level Tegra 4 SOC's.    

And believe me, they did.    Tegra 4 outperforms APPLE A6 .... that's bad news for APPLE and it means it's time to trot out the new big gun from APPLE .....  but where is it?     (big silence from the APPLE folks)



and I quote Brad Linder "The Tegra 4 chip, on the other hand, scores about as well as my desktop PC in the SunSpider test."

Folks, I done told you a long time ago this was a coming, but I got hooted at some.   So, it's here (on one benchmark, anyway).   There ARE phone chips that can perform on the PC level now.

So, what about APPLE?  The only way for APPLE to WOW !!! anybody now this year is to actually uncork a next generation chipset (A57 level at 64 bits on a 20 nm process).   They could do it, they have had the chipset specs for nearly a year and a half now and if 'ol Steve Jobs was around to push things you could bet they surely could/would do it.  

But Steve is dead now, and so apparently is APPLE.

If APPLE loses their tech edge and gives up being the WOW company to folks like Nvida and Google then silly Chinese may well be giving us our next set of WOW products in a couple of years time.

Nvida and Google are the tech leaders right now, let's see how long this lasts.

In this 2010-2015 iteration of computer change (the cell phone / tablet wars) APPLE has led the pack the whole way both in screen resolution and in ARM processor power.

No one has ever bested them on a tablet or a phone.   We will see if this is changing now long term.

Yes, in PC space (now shrinking) APPLE never led processor power-wise.  They were always stuck on IBMs Power PC platform at the first and they changed over late to Intel's products.  Because they didn't lead, they suffered greatly as a company during this period.

You'd think they may have learned something over the years.  Leadership in an industry MUST always be protected and that takes work to always be the first to innovate so as to stay on top.

And now that they are a creator/maker of ARM processors, they have to work to stay best at that too.

Salt on Intel's wounds

Nvidia's Tegra 4 was bad enough, but Samsung's Exnos Octa core was worse yet.   Now here come the oriental phone vendors, buying up ARM liscenses and cooperating to build state of the art 20 nanometer fab plants in China.

http://liliputing.com/2013/02/renesas-introduces-ape6-8-core-processor-for-ph...

That's two oriental companies that will be fabbing Octa core A15/A7 big little SOCs starting this year.

The supply level of powerful chips has increased this year compared to all previous years.

Anybody can be their own private Intel now-a-days it seems .....  just buy the license and buy a share in a new fab plant that is being constructed.

What Intel hopes will happen is that the fastest ARM SOCs will start to draw more power and waste more heat.   Then Intel can start to claim "equivalence" or "superiority" to ARM again.

With Big/Little configurations, ARM has beaten this equation as it runs off the small chipsets most all of the time, only revving the big guns when doing something that requires them.

And what sprinkles the salt on Intel's wounds is that when the big guns and the little guns are all firing at the same time (WFO) the current crop of ARM SOCs still draw less current and waste less heat than Intel's most efficient set up.

But Intel claims that will all change once everybody gets below 20 nm.  They have made similar claims at 48, at 32, at 28 nm but it hasn't panned out yet.

==================

What will be the 12 gage loaded with rock salt?

When ARM goes to 64 bits next year at 20 nm (A57) and suddenly Intel has no lithography advantage to base their claims off of at all.   Parity at 64 bits, ARM at 20nm and Intel at 22nm, with Intel costing at least twice as much per SOC, making more heat and costing more battery time.

By then Tegra 4 will be dirt cheap as Tegra 5 (64 bit) will be the hot new SOC.   The oriental companies octa core A-15/A7's will be the bottom of the fishbowl price-wise because "64 bit shift" thinking will have gripped everybody.  

Nobody buys a 32 bit SOC when the 64 bit stuff comes out and the software begins to shift to 64 bit.    If they do, the price has to be compelling enough to get them to do it "against their better judgement".

AMD will be shifting to their new life as a 64 bit ARM SOC builder (assuming they survive the next winter) and HP will still be reinventing itself as an ARM hardware supplier.

Intel and MS will still be pushing out new versions of Windows 9 or 10 but it will be to a still smaller shrinking market.  

It is very hard to compete against "free software" when it is as good or better than what you are offering.   When business begins to accept non-office file structures as normal stuff, it is all over for MS.

Google will be busy playing out Google Fiber (100 times faster than anything internet that is now being piloted in Kansas City) as the GF technology is cooked now and the costs are low enough for Google to just go "do it".  

This key competency falls in line with "ad revenue" and Google's main mission which is to quietly sell ads while providing associated data and services.   And like the Nexus 7, Google will make a small profit while doing this "public service".   But they will quietly rake in a big profit off the little ads that will ride along with the show.



Google has enough money to create a new market or to correct old constipated markets (like cable internet).   And they are quirky enough to just go do it.

Virtual keyboards have been around for decades. They were a gimmick then and they're a gimmick now. The real world still needs to look at the keyboard when typing because we're not all stenographers. Voice typing is prolly the only thing that will be of any use - as long as you're then only one within earshot.

Geez,.. it's not like I don't have enough trouble with typing,... now they expect me to do it upside-down and backwards!...

T!hIIIss wil nowt g0o wlll... ...

Look forward 3-4 years ....

Computer fanboys will be bragging on their custom handled 64 bit "suitcase supercomputers" and will keep track of their little carry case expandable "sea of cores" ego toys in teraflops.   Normal users will be happy with mere gigaflops ....

Hint:  you can buy lots of tablets and phones today that outdo the second generation Cray 2 Supercomputer for straight out performance.

And some person will build a ultra-realistic game that takes MORE, so you have to add some processor cards to your backbone.

(and yes, this graph suggests that the Tegra 4, Snapdragon 800 and the Exnos octa core (and the chinese octa cores) will come out this year at about the Cray 3 level of performance)

And I can remember when I thought a supercomputer was really something unique and different.