Intel has teased its 8th generation Core processors, dubbed Coffee Lake, and says they'll be available this year. Here's what we know so far about Coffee Lake and Cannon Lake.
What is Intel Coffee Lake?
At a recent investors conference, Intel showed a slide detailing the roadmap for its Core processor range. The 6th and 7th-generation chips we know about - Skylake and Kaby Lake.
However, the slide - titled Advancing Moore's Law on 14nm - stated that 8th-gen Core chips would launch in the second half of 2017. So the twist was that these would also be based on a 14nm process.
No-one expected a fourth generation to use the same process, especially as Intel has already waved a 10nm Cannon Lake chip around at CES in January this year.
Some are claiming that the internal codename 'Coffee Lake' refers to the new 8th-gen 14nm chips, and since Intel hasn't said otherwise, we'll have to run with it, at least until we hear different.
In short, then, Coffee Lake is the 8th generation of Intel Core i processors, based on the same 14nm manufacturing process, and successor to Kaby Lake.
Intel Coffee Lake features and performance
So we know - roughly - when the 8th-gen Core processors are launching (likely very late in 2017 and starting with mobile chips) but how is Intel going to extract yet more performance without a process change?
Typically, or at least in the past, the company would come up with a new architecture (the chip's design) and then follow up with a 'process shrink' the following year to boost performance.
Recently, Moore's Law, which says that the number of transistors that will fit in a given area will double every 12-18 months, has slowed down. Or so it seems.
Look beyond the headline 14nm figure, and Intel has actually improved the process without changing the figure. According to some sources, Intel refers to Skylake as 14nm, Kaby Lake as 14nm+ and Coffee Lake at 14nm++.
These aren't simply names to delineate the different generations: there are performance improvements behind them. One is the move to FinFET - effectively a 3D transistor, as used by the latest graphics card GPUs - and along with other tweaks, it's probably unfair to call Coffee Lake a 14nm CPU.
In any case, the aforementioned slide quoted a 15 percent boost in performance over Kaby Lake, which itself is 15 percent quicker than Skylake.
So a Coffee Lake-powered PC should be noticeably quicker than a three-year old machine.
Intel is also likely to continue to improve the on-board GPU performance, which may well account for a decent chunk of that overall 15 percent bump. We'll have to wait until we see the benchmark results to know for sure, of course.
Quite how the chips will stack up against AMD's forthcoming Ryzen is another unknown. It could be shaping up for a very good year for CPUs.
Coffee Lake vs Cannon Lake
You're not going to be able to buy a PC of any description with a Cannon Lake chip in 2017. Intel has said that it will push 10nm processors to datacentres first, and this must be partly in order to prove the technology before rolling it out to consumers. But it could also be because server chips cost thousands of pounds, while laptop and desktop PC processors cost hundreds.
As with Kaby Lake, we're likely to see Coffee Lake first in low-power laptops and 2-in-1s, then on the desktop in early 2018. There may come a point where you have a choice of a Cannon Lake or Coffee Lake processor, but right now, it's far too early to say which will be the one to buy.