It's easy to look at Intel Corp.'s (INTC) unveiling of new high-end notebook processors based on its existing Kaby Lake architecture less than a year after launching the first batch as simply a pre-emptive strike against AMD Inc.'s (AMD) expected fall launch of its Ryzen Mobile processor line. And countering AMD, whose shares are down over 3% in the wake of Intel's announcement, is almost certainly one of Intel's main motivations for refreshing Kaby Lake so quickly.
But there's also probably another one: Intel has found that parts of the PC market are holding up much better than it expected when it significantly cut its PC-related R&D spend last year. Its new chips aim to keep demand in one of those segments humming along.
Intel's four new notebook processors join its U-series line, which is meant for thin-and-light devices (ultrabooks, notebook/tablet convertibles, etc.) requiring a relatively high level of performance. Two of the chips, the Core i5-8250U and 8350U, sports MSRPs of $297 and come with 6MB on-chip Level 3 (L3) caches. Two others, the Core i7-8550U and i7-8650U, sports $409 MSRPs and have 8MB L3 caches. As is the norm for U-series parts, the chips sport integrated GPUs.
Though still based on Kaby Lake, which when launching last year was referred to as Intel's 7th-gen Core CPU platform, Intel labels its new chips its first 8th-gen products due to the their architectural improvements. Interestingly, chips based on Intel's Coffee Lake CPU architecture -- the first ones are due this fall -- will also be referred to as 8th-gen. As will chips based on Intel's Cannon Lake architecture (due in 2018), the first to rely on a 10-nanometer manufacturing process. Chips based on Intel's Ice Lake architecture, which the company shared some details about last week and might not arrive until 2019, will be the first to be labeled 9th-gen.
Notably, whereas the U-series chips being replaced by Intel's new processors -- they're also based on Kaby Lake, and were launched last fall -- only contain 2 CPU cores and support 4 simultaneous threads, the new chips each contain 4 cores and support 8 simultaneous threads. One tradeoff: The base clock speeds of the cores within the new chips are only between 1.6GHz and 1.9GHz, down from the 2.2GHz to 2.5GHz range of their predecessors. However, their turbo clock speeds of 3.4GHz to 4.2GHz match or exceed those of the chips being replaced, and in spite of the extra cores, the chips match the 15-watt max power draw (TDP) of last year's parts.
Thanks to the higher core and thread counts, as well as improvements in chip design and the 14-nanometer manufacturing process Kaby Lake relies on, Intel claims its new U-series processors deliver a 40% performance improvement over the 2016 chips, as measured by the SYSmark 2014 SE benchmark. Nonetheless, Intel insists notebooks running on the chips can provide 10 hours of battery life when playing 4K video relying on the efficient HEVC (H.265) codec.
Intel is promising its new notebook processors will deliver big performance gains without sacrificing battery life.
Those performance gains are likely a big reason why AMD has sold off following Intel's announcement. Right now, some 1080p gaming issues notwithstanding, AMD's Ryzen 7 and Ryzen 5 high-end and mid-range desktop CPUs are pretty competitive on a price/performance basis with the Kaby Lake-based Core i5 and i7 desktop chips they take aim at.
And benchmarks show AMD's new Threadripper 1950X high-end desktop CPU (it costs $999, contains 16 cores and supports 32 threads) outperforming Intel's new $999 Core i9-7900X CPU (10 cores and 20 threads, based on the older Skylake architecture). That might motivate Intel to cut the 7900X's price, as well as prices for some of the soon-to-launch Core i9 chips that are set to sell for between $1,199 and $1,999.