Back in January of this year, microprocessor giant Intel (NASDAQ:INTC) launched the Core i7-7700K, a quad-core chip based on its Kaby Lake architecture targeted specifically at the PC enthusiast and gaming crowd. That crowd isn't a large part of the personal-computer market, but it's one that has been experiencing phenomenal growth for quite a while now.
The chip was largely viewed as an incremental improvement from its prior-generation Core i7-6700K chip, based on the Skylake architecture. The 7700K ran at faster speeds out of the box and could be pushed to higher peak speeds through user tweaking than the quad-core Core i7-6700K could hit.
However, in what appears to be a bid to cope with intensifying competition in the enthusiast and gaming desktop market, Intel has accelerated the schedule of its Coffee Lake-based desktop processors.
Coffee Lake should come in configurations with up to six cores and, based on the various leaks that have hit the web, should even push peak frequencies up a smidgen from what the 7700K could do, thanks to the use of an improved 14nm++ technology to build the chips.
According to what appears to be an inadvertent "leak" by PC builder Eurocom, Intel plans yet another increase in core count for the desktop version of the follow-on to Coffee Lake, known as Ice Lake.
After six comes eight
By way of VideoCardz, Eurocom says Intel is planning to launch new processors with eight physical cores, each of which supports two logical cores, for a total of 16 logical cores in the second half of 2018.
These parts are almost certainly going to be based on the company's Ice Lake architecture. Ice Lake is a new processor architecture manufactured in the company's upcoming 10nm+ technology. It's Intel's second-generation 10nm technology.
A modern processor is composed of hundreds of millions, if not billions, of transistors. Since Intel's 10nm technology should enable a dramatic reduction in the amount of area per transistor, it makes sense for Intel to toss in more cores.
Additional cores are easily understood selling points, particularly as the amount of performance per core in modern PC processors isn't going up by that much with each passing generation. The trick that Intel needs to pull off is to cram those extra cores into the chip while at a minimum holding the line in terms of per-core performance -- though, ideally, this would go up as well.
Intel had better get 10nm+ to work soon
It's no secret that Intel has struggled to bring its 10nm technology into mass production. Intel had originally planned to begin mass-producing chips using it by the end of 2015, but it's looking as if production won't start on the first generation of 10nm technology until the end of this year.
For Intel to successfully launch the Ice Lake parts in the second half of 2018 as planned, it'll need to work quickly to address the issues with its first-generation 10nm technology, so that the company's manufacturing technology development teams can throw their collective weight behind getting 10nm+ ramped up into production.
I hope Intel will provide investors with an update on the progress of its 10nm technologies and products on its upcoming earnings call, which will probably take place next month.