somebodyspecial :
If you still think Intel is leading the industry in fabs by 2yrs you are sadly mistaken. Say hello to samsung. When Samsung starts pumping out 64bit Arm chips (probably for NV etc) at 20nm look out. I'm guessing Denver (and all it's competition) will be aimed at desktop speeds (think 4ghz Arm cores, and Custom in NV's Denver case, Qcom will likely have some custom Arm 64 core also) and we will see a VERY close power race in all things at that point. I would be shocked if NV doesn't put a video card with a SOC in a 500watt box for a desktop PC. Everyone is coming for x86 market share in 2015. AMD/Intel should be very nervous. More games are already being made on mobile than PC or anything else. As they evolve (4GB in your phone next year?, holy crap), and graphics up their game so to speak, these will become pretty darn good desktops that browse, get email, and play great games. The next shoe to drop will be a full adobe suite on ARM and many apps will follow after that.
WINTEL's party is over. It might take years for them to really become even with ARM in apps but don't forget 90% of the public just browses the web, gets mail and plays games. So for most of the public we really don't need a PC these days, or at least we don't need it to be Wintel. How many people do you know that even own a copy of OFFICE? I could actually dump Office myself if it wasn't for being in IT/Enterprise.
If google wants to speed this train along, they better make their next OS require 3-4GB of memory in the phone etc Note it doesn't have to be a pig, I just mean a check for this amount in the device or the OS won't run even if it can run great on 512MB like kit kat. You want to make the norm higher when the cost is only $11 for the 1GB in iphone. The 3GB module samsung is making now can't be more than $20 probably. For $10 or so more on the phone you give developers 3GB to work with instead of 1GB no matter how much of that goes to the OS bigger memory is better. If the OS uses 512MB in a 1GB phone you only give the dev 512MB to run in. Why not give that same dev 2.5GB of 3GB to work with?
You are quite wrong about Samsung's ability to beat Intel in fabrication. Things like nand flash and RAM are very simple structures which are easy to shrink. They cannot simply apply the same process to a CPU and get it to work without a lot of tweaking to get things right. Intel is full of wizards who can make die shrinks work, and they will have a process advantage over everyone for years to come. Maybe in 10 years or so someone may overtake them, but certainly not before then. I mean, they are already playing with sub 10nm CPUs in the lab in prep for figuring out mass production in a few years. Meanwhile TMSC and Sammy are having difficulty making much simpler structures any smaller than 20nm. I am not saying they will not get there, but it will be behind Intel.
As for Wintel being over, I think you may be right, but it has a lot more to do with Microsoft's failings than it does Intel's. MS could still figure things out, and I personally hope that they do as my whole ecosystem is MS and I am not a fan of starting over any time soon, but there is a real possibility that they could go belly up in the next 10 years... which is a really scary thought.
Let's not confuse the quantity of games and apps with the quality of games and apps. Most apps on the market are a joke and can barely work as advertised. If a company puts out a workable app, it is typically just a dumbed down version of their website, and most of the time I prefer going to a mobile website than using an app. And games? please! I am not ready to concede that mobile games are anywhere near the quality of flash based games from years gone by, much less full console or desktop AAA titles. I mean who was really impressed with Angry Birds when it came out who had played 100 similar flash games 5 years previous? Not saying that I have not enjoyed myself some Angry Birds from time to time, but the idea that it is somehow better than some of those much earlier offerings is a bit short sighted. And outside of games like Angry Birds, or ports of old games like Final Fantasy, I have yet to find a single game that I really enjoy on any mobile platform... and I like a pretty wide range of games. I hope and pray that it gets better with time, but the programming and the interface bring some true challenges to the table to make games that are worth playing for the sake of playing them as opposed to games that are just there to kill time. I hope to see that some time, but have yet to see it.
The big problem with ARM is ironically an issue of efficiency. ARM was designed and built around the idea of having limited but purposeful instructions, on a lightweight processor that takes a minimum amount of power. It does this very well, which is why it has been used for years in everything from microwave ovens, to cars, to the earliest of digital cordless phones, and CD players. Heck, your average PC has a ton of ARM chips which are used to control things like your HDD, the laser in your optical drive, etc. ARM is everywhere, and it is there for a good reason; because Intel cannot cut things down small enough to be effective as a single-purpose device (try as they may with quark).
But let's not confuse 'low power' with 'efficient power use'. For those limited use platforms ARM is extremely efficient. The issue is that when ARM gets bigger and more capable then it is unable to scale that efficiency. When it comes to multi-tasking, content creation work, and other multi-use environments then ARM falls flat on it's face from a power usage perspective. Yes, the Intel chip uses more power in general, but the performance is there to get the workload done so much faster that the ARM chips use much more power for a given task rather than less. As Intel continues to wage war on power usage with their Atom and Quark product lines, while continuing to trim the fat on their server and desktop lines, I think that we will see that it is far easier to scale x86 down than it will be to scale ARM up. Intel didn't really care much about power usage until ~4 years ago, and we are just starting to see the beginnings of their power reduction efforts. But ARM has been working at getting bigger for more than 10 years and is just beginning to see a little bit of success.
x86 will not go away any time soon. For that matter, ARM is not going to go away any time soon. There is place in the world for both platforms to shine, and the back and forth competition between the two will improve each other greatly.
Facebook
Twitter
Instagram