IDF 2014 – Intel Core M: architecture, performance, 2-in-1 designs
The tick-tock concept is very well known when it comes to Intel’s commitment of respecting Moore’s Law. Haswell was a tick having a new microarchitecture but being based on the same manufacturing process as Ivy Bridge (22nm), while Broadwell has the same basic microarchitecture but it’s based on the 2nd generation Tri-Gate manufacturing process on 14nm.
The 14nm process is highly optimized for low-voltage operation and also has a couple of tweaks for improved area scaling which is crucial for lower power and cost per transistor. Although the cost per mm2 went up because 14nm due to wafers and the need of additional masks, the total cost per transistor actually is below the projected improvement line because the dimensional scaling is better than projected (transistors takes less space on the die).
Intel Core M was announced at IFA 2014 in Berlin and is the first CPU based on the new Broadwell 14nm architecture. In this case, the architecture and manufacturing process were optimized for low operating voltage. The result is a impressive 4.5W TDP, with the partners having the liberty to set a lower limit (3W) or a higher limit (6W). While a fanless 4.5W TDP may seem slow to you, the results will pretty much shock you, as the CPU is a 2 core / 4 thread part running at a 1.3GHz base frequency with Turbo Boost up to 2.6GHz. Note that this is the same architecture that will be found on the Intel 5th Gen desktop and notebook CPUs (Core i3 / i5 / i7) that will launch early next year.