In addition to a brand new tier of consumer-grade processors, the Core i9 family, Intel also recently introduced the “X-series.” Here’s where things get confusing, because Core X processors don’t fit into a single line, family, or even chip architecture—it’s purely a marketing term, similar to the previous “Extreme Edition” processors that Intel offered a few years ago.
Basically, Core X processors are faster and more expensive models aimed squarely at PC performance enthusiasts, available in Core i5, Core i7, and Core i9 variants with commensurate features at each tier. Let’s take a look at the new offerings, and what changes at each level of price and performance.
A Technical Breakdown of Every X-Series Processor
Here’s the technical breakdown of the new Core X series. Note that a lot of information on the more powerful Core i9-X CPUs, coming in the latter half of 2017, hasn’t been released.
So we have nine new chips: five Core i9s (the first ever), three Core i7s, and one lonely little Core i5, all bearing the “X” suffix. The insanely powerful i9-7980XE gets an extra “E” thrown in there, because it’s still bearing the “Extreme Edition” moniker used in some of the older high-end Intel chips.
All of the X-series chips use the new LGA2066 socket and its accompanying X299 chipset, and all except the i7-7800X will support super-fast DDR4-2666 memory (or possibly faster for the more expensive models). That’s about all they have in common, though: the prices range from practically budget at around $250 to impractically ludicrous at $2000 for the 18-core i9 7980XE. The bottom two chips support “only” 16 PCI express lanes—an important factor for gamers who like to build multi-GPU rigs—up to 44 or better for the more expensive chips. That’s an important detail: to get more lanes than the previous $500-level Intel processor, you’ll need to spend at least a thousand dollars.
Core count is a big differentiation as well. The cheapest X-series chips at the i5 and i7 level will use quad-core designs with higher base clocks, with processor cores and threads increasing up to 10 for the i9-7900X (the highest Intel had produced in a consumer CPU before the X series) up to 12, 14, 16, and 18 in the even more expensive models, all of which have hyperthreading. Power draw will be high, starting at 112 watts and going up to at least 140. Extra L3 cache and support for Intel’s Turbo Boost 3.0 overclock system is limited to the more expensive models.
The X Series Isn’t a CPU Architecture…
You’d think all these new processors on a new CPU socket would be using the same architecture, bat that isn’t the case. Intel has new “X-treme” versions of existing processor designs upgraded as part of this marketing push: the i5-7640X and i7-7740X, the two cheapest in the series, are based on the Kaby Lake designs first released in early 2017. The rest of the series is actually a modified version of the Skylake architecture, which uses slightly older (~2015) versions of the 14 nanometer fabrication process. These two upgraded platforms, using the new and more complex LGA2066 CPU socket instead of the LGA1151, are labelled “Kaby Lake-X” and “Skylake-X,” respectively.
That doesn’t mean that these processors are in any way lacking, except in the obvious terms of price and value. But it does mean that Intel is in the awkward position of promoting a handful of new CPUs using an architecture that’s soon to be replaced with a second refinement of the original Skylake design. The upper tiers of the X series should be hitting the market at the same time as new Coffee Lake processors are becoming available for less “extreme” desktop applications, and perhaps only a few months before the next major shift comes in the 10-nanometer Cannonlake architecture. Patience, as always, is a virtue for both bargain hunters and bleeding-edge tech enthusiasts.
…Or a New Processor Socket…
While the X-series does come with a brand-spankin’-new processor socket, the LGA2066, that doesn’t mean that Intel is looking to make its older socket designs obsolete immediately. The LGA2066 is the new enthusiast-grade standard replacing the LGA2011 and LGA2011-3 designs, but upcoming revisions in Coffee Lake and Cannonlake will probably still be using the older and less complex LGA1151 socket (though Intel hasn’t announced anything regarding this yet).
Even if that socket itself is replaced as new CPU revisions are introduced, it will probably get something similar in size and power draw, a much more economical option than the full enthusiast-grade LGA2066. That will let Intel keep part prices down, which is especially important for its vendor partners supplying cheap and mid-level desktops to corporate and government customers by the thousand. The takeaway for the average system builder is this: immediately upgrading to a $300 (or more) LGA2066 motherboard might only grant you a few months’ worth of performance bragging rights before newer, faster Intel chips become available…even with older and cheaper socket designs.
…It’s a Marketing Term
The X-series, like the Extreme Edition chips before them, doesn’t have any set list of technical requirements. At the moment, the only thing that really binds the introductory chips together is the fact that they’re marketed (and in some senses, priced) for the enthusiast buyer and they’re unlocked to promote easy overclocking. They’re fast—oh my are they fast—and they’ll certainly be worth the money for those system builders who want to push their nitrogen-cooled rig just shy of the point of melting through their desks.
But this is the important part: because the X-series is largely an expansion of current architectures, it’s not so fast that it won’t be outdone by the best of the next major processor revision from Intel. And since those processor families are sure to get their own X-treme variants for that same enthusiast market, it would be smart to wait if you’re not in dire need of a full CPU-and-motherboard upgrade (which might also need a new power supply for the high-performance socket, faster RAM, a bigger CPU cooler…you get the idea). Gamers in particular would be wise to wait, since even the most demanding of current games probably doesn’t have the software support needed to truly take advantage of 12 or more processor cores.
Image credit: Amazon