On May 30, Intel (NASDAQ:INTC) formally unveiled its Core X series of processors aimed at the so-called high-end desktop personal computer market. The first wave of Core X processors -- that is, the parts featuring four, six, eight, and 10 processor cores -- is slated to be available for pre-order on June 19, with shipments happening a week later.
The company then plans to launch a 12-core Core X processor in August, and then finally, the members of the Core X family with 14, 16, and 18 cores are expected to be available in October.
All told, these appear to be exciting products for the high-end desktop market and easily the company's strongest showing in this market for about a decade or so (in this Fool's opinion).
However, it's always fun to speculate about the future, so in this column, I'd like to offer just that -- educated speculation about what Intel will do next for its high-end desktop customers.
No more cores -- focus on frequency
Intel's high-end desktop processors are simply its server/datacenter processors adapted for consumer use. This year's crop of high-end desktop processors is based on the company's Skylake Xeon processors slated to launch in the middle of the year for datacenter customers.
Now, it's worth noting that the company has said that it plans to release a refreshed server processor family in 2018, known as Cascade Lake. Intel hasn't formally detailed Cascade Lake, but I believe the Cascade Lake processors will essentially be the same as the Skylake Xeon processors, but with a few key differences:
- The Cascade Lake chips should be manufactured in the company's 14-nanometer++ technology, which promises performance/efficiency enhancements relative to the 14-nanometer+ technology that's used to build the current Skylake Xeon (and, by extension, Core X) processors.
- The Cascade Lake chips should support the company's 3D XPoint memory modules.
The latter feature isn't relevant to the high-end desktop market; it's aimed squarely at data center customers.
The former feature, though, should help make the Cascade Lake-derived Core X processors more attractive to consumers. Intel has indicated that its 14-nanometer++ technology delivers a roughly 10% performance boost compared to its 14-nanometer+ technology.
It's not unreasonable to expect that Intel will be able to wring out at least a 10% performance gain across the board for its next-generation high-end desktop processors due to the manufacturing technology improvements as well as potential further refinements to the physical implementations of the chips.
I don't think we should expect Intel to add more cores to the Cascade Lake chips; I only expect a core count boost once the company transitions to a next-generation manufacturing technology that brings a significant chip area reduction.
14-nanometer++ improves performance, but it isn't expected to lead to a reduction in chip area.
This is by no means bad for the company's high-end desktop lineup; the 18-core chip already has more cores than any reasonable consumer workload could make use of, so Intel is probably better off trying to increase per-core performance in this market (higher frequency, architectural improvements to the core, and so on) than trying to push beyond the 18-core mark -- at least, for now.