Image source: Intel.
Back in 2015, microprocessor giant Intel (INTC -2.40%) announced that it would be breaking with its long-standing "tick-tock" development methodology. Under this model, the company would deliver two product families with a given chip-manufacturing technology over roughly two years before transitioning to a new manufacturing technology.
Starting with the company's 14-nanometer manufacturing technology, though, the company has shifted to a development model by which it does three product generations over roughly three years using a given technology.
Intel's 14-nanometer processor family will ultimately have three members: Broadwell, Skylake, and Kaby Lake.
Thanks to a leak from the generally reliable website BenchLife.info, it appears that Kaby Lake represents a substantial improvement -- at least for low-power devices -- over Skylake. Let's take a closer look.
A major step forward for Core M
Between two processors with the same underlying architecture, a processor that's rated at a higher clock speed -- usually measured in gigahertz (GHz) -- will be faster than a processor rated at a lower speed.
The fastest Core M processor based on the company's Skylake architecture runs at a 1.2GHz base frequency and features a maximum single turbo speed of 3.1GHz and a maximum dual-core turbo speed of 2.9GHz. According to the BenchLife leak, the fastest Kaby Lake-based Core M chip will feature a maximum single-core turbo speed of 3.6GHz and a maximum dual-core turbo speed of 3.4GHz.
These are fairly significant jumps in frequency, and they are likely to translate into a real boost in performance for very low-power notebooks such as the MacBook, as well as for 2-in-1 convertible devices such as the Surface Pro.
A smaller step for Core i7
BenchLife also has information on Intel's Core i7-7500U, a part rated at a 15-watt thermal design power, targeted at more traditional thin-and-light notebook/Ultrabook form factors. The CPU will apparently sport a base clock of 2.7GHz and both single- and dual-core turbo speeds will come in at 3.5GHz.
The prior generation Skylake-based Core i7-6500U runs at a base frequency of 2.5GHz with maximum single-core and dual-core turbo speeds coming in at 3.1GHz and 3.0GHz, respectively. The higher end Core i7-6600U has a base clock of 2.6GHz with maximum single-core and dual-core turbo speeds of 3.4GHz and 3.2GHz, respectively.
It would seem that with Kaby Lake, Intel is able to deliver higher speeds in a given power envelope.
Kaby Lake is looking like a reasonable product
It would seem that Intel intends to deliver generation-on-generation CPU performance improvements with Kaby Lake by way of higher clock speeds; it's not clear whether Intel has also made any improvements to the underlying CPU architecture in order to deliver an improvement in performance-per-clock.
At any rate, better CPU performance, coupled with expected improvements in the media subsystem, should make for a reasonable generation-on-generation improvement, at least for low-power systems.
It will be interesting to see if the higher-performance Kaby Lake products, particularly those aimed at high-performance laptops and mobile workstations as well as desktops, will see similar generation-on-generation bumps in performance.
