One more indication that ARM has won the architecture war. And it isn't even about the hardware.
ARM is a ho-hum hardware architecture, but it turns out that isn't really important. It's just good enough so that alternatives with an additional 15% or so efficiency just aren't worth the effort of building an entire architecture infrastructure. Process improvements and software inertia are just too big to fight.
What are you talking about? ARM has the little wedge on the chart in the article about a design yet to tape out on a vendor (TMSC) that has not demostrated any working logic designs on a 14nm process. You have to put your boots on first before you can win a war.
Yep- this little wedge is just one battle. The war is much bigger and it seems to me that the war will tilt this battle in the favor of ARM. Battles are always difficult to predict, but I'd put my money on the one winning the arms race...
@daleste: Re not expecting Intel to grow: The industry-wide push to SDN is driving a lot of comms apps to x86 servers just as Inte is driving a focus on comms SoCs. I expect them to gain significant ground.
If the 64bit ARM family comes out strong (good offerings x4 x8 etc.) and with higher processing speeds than have been available (2+ GHz) then I can see ARM winning big. Last I looked (a few months ago and not real hard) there was just not much real powerful silicon in production. Now we are hearing about all this 64bit multi-core work, well time will tell, but it looks like a great opportunity for ARM!
Blog Doing Math in FPGAs Tom Burke 13 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 ...