Image: Intel Newsroom
By tech2 News Staff / 23 Mar 2016 , 14:10
The time has come for Intel’s “Tick-Tock” strategy and “Process-Architecture-Optimization” is the new buzzword. Intel’s Tick-Tock strategy meant to introduce a new processor architecture one year, then optimise it the next, indicates a report on The Motley Fool.
This strategy has been Intel’s mantra these past few years and they’ve religiously stuck to it, till around 2013 anyway. Intel’s “Tick” cycle has almost always been about reducing the transistor size, it’s now at 14nm, and the tock about improving the performance of the new platform.
In 2014, the transition to a 14nm process was heavily delayed (Do you rememberBroadwell?) and Intel simply skipped a generation and released Skylake. Citing manufacturing and engineering challenges, the transition to 14nm was hard on Intel.
As the transistors get smaller and smaller, the chance of leakage increases, thus affecting the accuracy and reliability of the design. In other words, there’s actually a physical limit to how small a transistor can go before it becomes completely useless. Intel has already indicated that the 10nm process, the next “Tick”, is already in troubled waters. In the face of these facts, the abandonment of the Tick-Tock strategy is a sensible one.
Intel’s 10-K filing seems to indicate that they’re hitting that wall and have sensibly decided to focus on “Process Architecture Optimization” rather than on perpetuating the “Tick-Tock” cycle. Hopefully, we won’t have to upgrade our motherboards every time we want to upgrade the CPU from now on.
ARM and TSMC have already started collaborating on a 7nm process, which is much smaller than the 10nm process that Intel’s been struggling with.
No comments:
Post a Comment