Intel will reportedly bring new chips to market based on the company’s upcoming 10nm process technology in early 2017. The news came via Taha Khalifa, Intel’s general manager for the Middle East and North Africa region.
We’ve reported three weeks ago that Intel expects to roll-out 14nm Skylake parts in the second half of the year. We’ve also exclusively told you that Intel’s 10nm process technology will not show up in 2016. It’s becoming increasingly difficult every year to keep up with Moore’s law. The majority of Intel’s market segments have been stuck on 22nm for three years, despite Intel’s Tick Tock strategy. 14nm is poised to span a similarly extended life cycle to 22nm.
Intel Hopes to Launch The First 10nm Parts in 2017
We have been consistently pursuing Moore’s Law and this has been the core of our innovation for the last 40 years. The 10nm chips are expected to be launched early 2017.
Said Taha Khalifa, general manager for Intel in the Middle East and North Africa region.
Intel originally promised to introduce 14nm Broadwell processors in late 2013. But due to various technical setbacks Broadwell was delayed to 2014. And out of the 3 family’s of Broadwell processors, namely Broadwell-Y , -U and -H we’ve only seen some Broadwell-Y product introductions later that year. Intel introduced Broadwell-U parts last month and there’s wide industry speculation of Intel actually cancelling Broadwell-H parts in favor of Skylake.
It’s not just Intel that had difficulty with 14nm. The process node continues to be a challenge for the industry as a whole. And as we’ve reported yesterday, Samsung will only begin producing 14nm FinFET products in volume in the second quarter of the year. While TSMC continues to suffer setbacks with its challenging 16nm FinFET process.
Back to Intel’s 10nm node. Just like with 14nm we’re going to see smaller tablet and mobile parts introduced first on the new node. With mainstream and high performance parts coming later. It’s been revealed quite a while ago that Intel’s first 10nm family of processors will be code named Cannonlake. These will be the Tick in Intel’s Tick Tock cadence. So Cannonlake processors will be based on a die-shrink of Intel’s Skylake CPU microarchitecture.
Cannonlake should feature SoC interconnect improvements similar to AMD’s Kaveri and Carrizo. Namely shared coherent physical memory addresses and a more capable set of heterogeneous functionality. We’ve already discussed in great detail what Intel plans to bring with Skylake in relation to heterogeneous computing.
The company had released a whitepaper in 2013 in which it expressed and equally strong interest in Heterogeneous computing as AMD. Nvidia shares a similar vision as well. So it’s evidently clear that the entire industry agrees on one new direction to drive efficiency and performance to compensate for Moore’s Law recent deceleration and inevitable standstill.