So, you have a major player (SONY) going with the first of the sea of cores idea using the very smallest of the ARM chipsets in big multiple numbers giving enough power to kick the asses of the folks who went with the ARM approved big/LITTLE format (while still using less die space, less battery power and at a higher processing yield rate and a lower cost too, let's not forget that little nugget).

ARM needs to rethink their progression path a bit and streamline their 64 bit implementation according to the reality paths the market has chosen.  Then they need to get tight with Linaro to make sure all the necessary bits & pieces are RELEASED in the Linux Kernel before the stuff starts to roll out ....

They also need to come out with a really really good graphics core set to go along with those plans.  Primo phones all seem to run VR graphics at this point in time.    Mali isn't bad, but it isn't the best either.

The third tier suppliers of chipsets are beginning to lead the show now and they will jump abruptly into 64 bit land looking for more speed and better energy consumption and they will do it with larger numbers of the very smallest most energy efficient ARM cores first.

They are leaving out the big in big/LITTLE in other words, as 2 littles more than equals one big as far as energy use and processing power goes.

Go with your sea of the smaller cores stuff right off the bat and plan very well for your littles performance boosts first off.   Your bigs won't really have anything to do until you go into desktop space and go head to head with Intel.   Indeed, your bigs aren't even hitting silicone before the next generation of littles does the "overcome by events" thing to them.

PS   What will beat Intel is a bigger sea of FAST littles anyway.  Laptops rule in business now and they all have batteries now don't they?  Once you break away from Microsoft bloat ware you don't need all that much in depth processing power (what you need is fast fast fast execution and low power consumption) -- see desktop Linux and Chrome OS for proof of that concept.

A 2.2 ghz A-53 octa core at 20nm would have both of these requirements, plus having 64 bit processing ability when needed, and a natural 32 bit execution mode for all that stodgy old legacy software (that's most all of it at this point in the game).

Linaro -- Linux needs to use all them little 64 bit cores very very efficiently.  Your backplane support needs to be in the kernel and IBM's old backplane stuff needs to be readily available and all dusted off & cleaned up & modernized as the "free open source solution".   You got you a bullet-proof pedigree there on that hardware, so use it.


==========


High cost will eventually ruin both Intel and Microsoft.  Soon you will be able to buy competitive "you name it" from China at 1/4 of the bloatware cost from MS or Intel.    And there is a whammy comming for Microsoft, slowly coming out of the old Ukraine and other old Soviet states.  

It's name is REACTOS.


    ..... and do you think Balmer is dumb enough to sue a long standing open source project with close ties to WINE, Open Office, etc. etc. etc.  

I hope he does, that way everybody can do a little dance together on the deck of the MS-Titanic as the ship slowly goes down.


A whole new generation of folks are learning to live off of Android devices and Linux is accepted by those folks as a valid "full desktop experience".    Most of the rest of the world accepts this concept now and the USA is not the biggest market any more anyway.   China and India and IndoChina are the biggest upcoming markets.

American CEO upon hearing the Sony Octa story.

What !!!   Both Sony AND Samsung are going Octa?   Then WE have to go Octa or we won't be seen as competitive for Christmas !!!  

Shite, we are falling behind -- again.   We need to get out of this business, it is eating up all our resources just to be late to the stupid party every single time.

Call the board design group and get them cracking !!!   What, lead time is too long ???? -- they say we need to put a stake in the sand a year out and stick to it?   What the heck are they talking about !!??   This Octa stuff wasn't even invented six months ago !!!  Tell them to pull their heads out of their asses and smell the coffee !!

Balderdash !!   We don't have an allocation slot on the octachips from either source??  We didn't buy one when it was offered to us ???  Who dropped that ball on that one ...... fire him !!

Those little Chinese guys are eating our lunch, moving right around us like we are big rocks in a creek bed -- you go tell board design I want a total redesign in 10 days anyway or else I'll fire the lot of them.


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


Meanwhile, at Allwinner in China.   "Gentlemen, our dual and quad core A-7 product sales are good.  We are utilizing 100% of our current fabrication allocation and we are selling all that we can produce at the current price structure.  Revenues are stable for the last 3 periods.

Our decision last year to leapfrog ahead with the buy in to the new secret Xeniwa 18nm plant is coming along well, the product samples we just received from them on the 16 core A-53 Dominator chipset are performing even better than we had expected.  

Since we designed the chipset specifically to retrofit into existing board designs the much smaller chipset can be footed to drop right into the majority of our customers designs.  Minor modifications for extra features only ..... they should be able to handle the work in house if they are truly capable inside of 30 days.

No, they can use the same battery as they are using now, current draw is much lower than before so overall phone life goes up.  We should ship single samples out next week to our key customers so they can begin the retrofit work.

Xeniwa's yield on the initial samples was 90%+.  Yes, ARM did a good job on making the A53 full process capable on the new Samsung X-ray dual bypass liquid lithography process.  Processing times are a little slow right now, but that will pick up when full production handling modifications are made to the wafer loaders.   Our allocation of the new line can certainly support a new product line introduction inside 30 days.

Which partner for the introduction?   Samsung made it a condition of lease of the new lithography line that they get first rights on all new production for the first quarter, which they MAY offer to Apple if they think Apple can move quickly enough on it.  What Samsung does not buy immediately we can sell to the next customer in line at any case, it is in the contract.

If not Samsung, then perhaps Sony as a backup first customer.   Send a set of samples to Xiaoping Qinchan as a third contingency as their current main board can drop in the sample as it is right now, so we can have available a working phone to showcase if we need one.   Cousin Cho will have us a working phone back in 10 days, count on it.  

Yes, we can sell them all through Cousin Cho right here in China if Sony does not bite, it will be easier for us that way and all the profits on both ends will stay within the family.   Suggest to Sony if they show interest that Cho can make the phones for them if they do bite on their first refusal --- tell them Cho can be in production much quicker than they can.

As part owner of the production facility, our cost to produce these new chips is actually less than the products we are buying from General Foundry right now -- so any premium over the existing pricing structures is ours to keep on this one.   Since we are first out of the gate on the Dominator, the premium pricing rolls directly into our pockets.

Yes, we win on this one.   That is what we do, win.  

Allwinner is our name after all.