On May 30, microprocessor giant Intel (NASDAQ:INTC) announced its new Core X-series of processors aimed at the enthusiast desktop personal computer market. These chips come in configurations with as few as four cores and scale to as many as 18 cores. This represents a significant improvement over the company's prior high-end desktop processors, which only came in six, eight, and 10-core variants.
Per an Intel fact sheet, the Core X chips with between 4 and 10 cores "are available at launch via online retail and through channel partners in all geographies," while the 12-18-core models are simply "coming soon."
Additionally, over on the ASUS Republic of Gamers forums (ASUS is a major vendor of computer motherboards and personal computers), the employee who goes by the handle "Raja@ASUS" said that the "18-core CPUs are not scheduled until later this year" and that (presumably gamers) "won't have them for a while."
In this column, I'd like to explain why it'll be a while before the 18-core Core X chip (branded the Core i9-7980XE) is available.
Getting this product right will be tough
In a previous column, I went over what I'd like to see Intel do to make the 7980XE attractive to the enthusiast/gaming market. In my view, Intel needs to be sure it isn't just selling gamers and enthusiasts chips with a lot of cores on them (that's trivial -- Intel can rebrand datacenter-oriented chips and be done with it), but that it's selling chips with lots of cores on them that can also run at high speeds in workloads that can't use all of those cores.
Intel processors have a feature that is known as Turbo Boost 2.0. The idea behind Turbo Boost -- a feature introduced years ago -- is that a chip with lots of cores on it shouldn't offer compromised performance to a chip with fewer cores on it in workloads that can only use a handful of cores.
So, in those scenarios where only a few cores are used, the few cores that get used can run at much higher speeds than they can when more (or all) of the cores on the chip are being used.
This is a good feature, but for Intel to advertise a given Turbo Boost 2.0 speed, the company needs to be sure that each core on the chip can run at that advertised speed since programs can conceivably be run on any subset of the cores on the chip.
So, for the Core i9-7980XE to not be a "compromised" product relative to the $999 Core i9-7900X (which has just 10 cores on it), it should offer a Turbo Boost 2.0 speed of 4.3GHz -- quite a high frequency.
I can't imagine that finding chips with this capability will be easy, so Intel may need some extra time to find, validate, and ship such parts.
The Xeon factor
It's also important to note that Intel's high-end desktop processors aren't based on chips that were purpose-built for the high-end desktop computer market. These are chips that were initially designed for the datacenter market that are tweaked and sold as ultra-high-end consumer products.
The market for datacenter processors is much larger than the market for high-end desktop processors, and based on everything Intel has said, it views itself as a "datacenter first" company. This means Intel's priority will be to make sure that its datacenter customers get the chips they want, so the high-end desktop customers may have to wait a bit, especially since the 18-core high-end desktop chip will need to represent, arguably, the "best of the best" of the 18-core chips that Intel produces.
The reality is that the market for processors like Intel's 18-core Core i9-7980XE is rather minuscule, even within the hardcore enthusiast gamer/prosumer market (a relatively small but fast-growing portion of the overall personal computer market).
There will certainly be buyers for this chip, and Intel is certain to make a nice chunk of profit per unit sold, but in the scheme of Intel's overall business performance, rolling out the 18-core Core i9-7980XE later than it will the lower core-count models isn't going to have much of an effect.