Why those ARM licensees, such as Qualcomm, TI, Marvell, and nVidia, are so up-beat of their upcoming SoC? I doubt anyone of those will be winners in this war. They are all competing the same market segment with little differentiation. The true winner is ARM, inc. And our consumer will take advantage of this as well. The big loser is likely Intel. No wondering Intel stock has been struggling at $20 even after record earning.
I believe the marketable benefits of GPU's will mostly come from the multicore servers (there are many that are already introducing them) before you see them in handsets. Recently I saw a Dell server with quadcore CPU's and GPU's that showed some screaming performance for high end simulation tasks. I suppose we are entering a new era of computing where the CPU as we know it today is destined to become an old shell of itself, much of its chores partitioned off to other processors.
Without the available bandwidth, it makes little sense to proliferate GPU's in smart phones. To that end, we need a clearer strategy from the telco carriers, not just to the last mile, but all the way to the last node connected (end to end).
ARM based architectures do have an edge currently over x86 platforms but I agree with eewiz, the OS will actually enable all architectures and with cloud computing shifting the compute capability away from the hand-held, key adoption will rely more on user interaction and GUI and ease of use and less on raw cpu speed or compute on the device. The PC (and even the notebook) will end up being a shell and only keeping the Display as portal with the smart phone (voice, SMS) and the tablet (remote controller, entertainment launchpad, content creation and consumption)taking over for all remaining activities a PC used to do. Until then, and for nVidia, I recommend repositioning their GPUs as General Purpose Units, not just Graphics Processing Units in the existing PC/notebook space..
Tablets/smart phones replacing the computers is not a reality so soon. Considering the power required to operate these devices is still coming from batteries and they aren't really powerful enough to run for long hours. But the future devices might be some CPU to integrate STB, wireless internet router, inbuilt audio control systems.
Yes Yes ARM smartphones will bury x86 PCs How? If these smart phones will have a bigger size virtual key board and a projection screen built in side then these PCs laptops will go to museum.Or if these smartphones can read our minds and operate and project the screen also into our retina screen then the world is smart and smart.
I don't think anyone is going to dominate the future; there is plenty of room for everyone. Pocket PCs are fine for some things, but my home PC has a 27 inch monitor and my phone is a phone only. I don't want my life dominated by over- communication. I don't want to write memos, emails, etc. with my thumbs. If my bosses wants to buy me a smart phone, that's fine, but I am not going to buy one and have to pay ridiculous fees to be connected. I don't twitter or SMS or have to feel obligated to answer banal chit-chat that seems to be the rage these days. So I will have a decent sized laptop for most things and a phone with limited abilities when away, but my phone will never replace my PC. PCs are bloated with too may layers of SW; obviouly! PCs need to go on a OS diet and get more efficient, but they are still the most comfortable form factor for real computing.
The issue with Intel is not their capability but legacy. Even when successfully entering ARM's embedded market (with Atom), it is mainly because they have a Si linewidth advantage and hence can run Wintel software on a credit card.
The game will only really change when developers are freed from the Wintel legacy. Away from the desktop, there are opportunities. But at the same time, the decisive factor is increasingly power consumption. Wintel has the lead because they are ahead in silicon technology. But at the same time, legacy free software can be 10 times or more efficient. And that is the opening others can take even with an older Si technology.
Would this trend indicated that Intel was caught blindsided by his fights to extract last few % of market share from AMD. You got to love high tech, just when you think you got the giant, a second later giant is struggling against more nimble and innovative competitor. But I would not count Intel out, it is will be fascinating to watch the battle though! Kris
Blog Doing Math in FPGAs Tom Burke 23 comments For a recent project, I explored doing "real" (that is, non-integer) math on a Spartan 3 FPGA. FPGAs, by their nature, do integer math. That is, there's no floating-point ...