http://liliputing.com/2014/01/samsung-galaxy-note-3-neo-6-cpu-cores-and-a-pen...
The hexa-core chip pairs two 1.7 GHz ARM Cortex-A15 CPU cores with four 1.3 GHz ARM Cortex-A7 CPU cores. The lower-power processor cores can handle most daily activities and the Cortex-A15 components kick in when you need a performance boost.ARM has had an issue with the orientals NOT keeping pace with the ARM generational roll out plan. The orientals are actively avoiding new licensing costs by sticking with an old ARM generation until every last bit of goodie is wrung out from it.
Then the orientals HAVE to buy another expensive ARM license for something newer, and they always try to leapfrog as far ahead as they can find large production volume on the smallest lithography process they can find (or buy into).
This is slowed leapfrog stuff is dangerous to ARM because Intel is coming and the generations must roll quickly to stay ahead of slow poky Intel, who is beginning to move a little quicker of late.
ARM let Rockchip have the A-12 chipset as an interim step but that effort seems to be stalled at the word GO right now -- and Rockchip got screwed over because their competitors then leapfrogged right over the A12 generation completely by going Octa core A15 on them.
So, Rockchip is all stuck now, still pushing out their supercharged RK3188 quad core A-9 chipset. I wonder if the A-12 will ever hit production this year at all as it has been leapfrogged. I bet they tweek the A-9 some more performance-wise rather than spend all the money to complete ARM's hastily cobbled together A-12 generation.
So what
IS ARM going to do? They can't lay out a plan because the orientals simply won't follow the game plan (whatever it is) and insist on doing just what they want to do whenever they want to go do it.
Samsung plays ball though, so ARM will now let Samsung pilot a new interim move for them. ARM will now let Samsung do the "sea of cores" thing with the current A-7/A-15 generation, letting them mix and match odd core counts while enabling all cores to be used simultaneously when needed.
Supercharging again, this time from ARM itself .... reliable and standardized.
So, here you go, extreme energy efficiency with high performance, but only when you need it. Works on the commonly available 28nm lines and also works at 20nm when you have some production time on the newer lines to run the stuff on them.
Now, this is a good idea and it will get adopted soon and by everybody.
Here is the BAD NEWS for ARM though, this newly released "sea of cores" tech will allow the orientals to stick with the current A-15 generation at 28/20nm for at least another full year -- they can simply go on mixing and matching cores for their exact product needs for at least that long. It will be cheaper to add a pair of cores than buy a new license.
Intel may have time to actually figure out how to design a real phone chip at 14nm if you keep on stringing the A15 generation along for that long a period of time.===================
So, the ARM chipsets are PLENTY POWERFUL ENOUGH right now, but the lack of the broad range full functional ARM OS all that is holding things back. Certainly, next year the power and efficiency levels will be off the current charts and there will be an absolute crying need for an ARM general OS.
Linaro and the Linux Kernel guys need to step forward and put some "bios like" standards out there that allow the multitude of hardware out there to interact sensibly with the operating system so the software guys can pull together a true general purpose ARM OS or two or three.
You know there will be more than one choice out there once the hardware interface can be rationalized, that is all that is stopping it from happening now.