News of Coffee Lake (which sounds like a whole lot of caffeine) arrives by way of PC Watch, which posted a leaked slide showing a Coffee Lake “H” processor SKU with a thermal envelope between 35 and 45 watts; two, four, and six cores; and Intel Iris Graphics 540. Also listed is a Coffee Lake “U” processor SKU with a thermal envelope between 15 and 28 watts, two cores, and Intel Iris Graphics 540.
The big note here is that Coffee Lake will finally introduce a six-core processor to the mainstream market. Six core parts have typically reserved for Intel’s “Extreme Edition” models, meant for PC gamers and high-performance enthusiasts.
Getting a little confused about what launches when? In 2014, Intel unleashed its Broadwell processors, followed by Skylake in 2015. Kaby Lake chips are slated for the end of 2016. All of these processors are based on 14nm process technology, whereas the Cannonlake processors shipping in the second half of 2017 will be based on 10nm process technology. What’s interesting is that Coffee Lake will be based on the larger 14nm process node.
That said, Cannonlake will replace Kaby Lake in the back half of 2017. Intel’s Kaby Lake is actually something of an anomaly, residing outside Intel’s tick-tock release schedule and referred as “Tock plus.” Cannonlake, however, will fall into the “tick” position, following Skylake (tock), Broadwell (tick), Haswell (tock), and so on. Could Coffee Lake be another anomaly?
Regardless, Coffee Lake will supposedly use the most advanced 14nm process node that’s optimized for processors beyond the Kaby Lake line. And because it will use GT3e graphics, the processors will feature eDRAM cache enabling faster graphics and compute processing capabilities on low-power devices and notebooks. The “H” models will likely offer more performance than the “U” chips, the former of which will presumably be aimed at high-performance notebooks.
So, what’s expected to be available in 2018 are 10nm Cannonlake “U” and “Y” processors with two cores and GT2 graphics, and the 14nm Coffee Lake chips with GT3e graphics. Intel is likely using the 14nm process node with Coffee Lake because the 10nm process reportedly still won’t have great yields by the time 2018 rolls around, so the company is relying on a matured, optimized process instead.
Still, why is Intel making two new and different processor families available on the market in 2018? The mystery will undoubtedly be unraveled as we grow closer to Cannonlake’s 2017 launch.
