Intel claims “double-digit” GPU performance improvements in the Y-series parts and 8- to 10-percent improvements in the U-series parts. Intel promises 12- to 19-percent better CPU performance over Skylake chips sold at the same price, which is commensurate with the speeds and prices laid out below. Foundries have been known to use multiple versions of a given process, so it’s not entirely unheard of.ĬPU and GPU speed improvements come mostly from clock speed increases rather than architectural improvements, as they did in the Haswell refresh. As Intel’s manufacturing processes stick around for longer, it makes sense that the company would spend more time optimizing them rather than just moving on to the next process as it has historically done. But the changes are said to improve the performance of the transistors by about 12 percent.
The company isn’t sharing many details-Intel says that 14nm+ has an “improved fin profile” and “improved transistor channel strain,” for what that’s worth. The biggest improvement from an architectural/manufacturing standpoint is something Intel is calling “14nm+,” an optimized version of the same process node used for both Broadwell and Skylake.
The maximum supported DisplayPort version remains 1.2, however, dashing the hopes of anyone who wants to drive a 5K display at 60Hz over a single cable using DisplayPort 1.3. HDMI 2.0 and HDCP 2.2 are also supported, which (respectively) enable 4K output at 60Hz over an HDMI cable and provide the DRM required to carry an encrypted 4K signal from the thing that’s playing it to the screen that will show it. Not only does this use a fraction of the power that a CPU uses to accomplish the same task, but it frees the CPU cores up to do other things. If you don’t already know, supporting hardware acceleration for certain codecs means that the GPU (usually via a dedicated media block) handles all the processing instead of the CPU. All Kaby Lake integrated GPUs will support hardware-accelerated decoding and encoding of 10-bit HEVC/H.265 video streams and decoding of 8-bit VP9 streams. Kaby’s biggest advertised feature is improved support for 4K.
#INTEL TICK TOCK SANDY BRIDGE PC#
On the plus side, the lack of huge changes means that Kaby Lake chips can easily be dropped into existing Skylake designs, something that will help PC makers get Kaby Lake systems on the shelves in a hurry. The 100-series chipsets should still be more than adequate for most people, but they’re missing things like 10Gbps USB 3.1 gen 2, to say nothing of Thunderbolt. Kaby Lake, despite a handful of changes and improvements, is only a slightly bigger deal.įirst off, none of these mobile CPUs will include new chipsets, which means you’ll get the same connectivity options as before. In 2014, when Intel’s 14nm process delays were holding up the Broadwell architecture, the company released a lineup of “Haswell refresh” processors that used small clock speed increases to outrun 2013’s original Haswell chips. Desktop CPUs and other higher-performance chips, including those with Intel Iris GPUs, will be announced in January around CES. The first Kaby systems are slated to ship in September, and you’ll see lots of new laptops at IFA next week. Today, we’re getting dual-core low-voltage processors, the kind you’d find in thin-and-light consumer laptops. Seventh-generation Core i3/i5/i7, Core m3Īs usual, Intel is releasing the Kaby Lake processors in waves. Sixth-generation Core i3/i5/i7, Core m3/m5/m7 No major architecture has changed, and Intel is still using a tweaked version of its 14nm manufacturing process, so most changes are relatively minor and built to serve a particular market niche. Kaby (rhymes with baby, named for a lake in Canada) first appeared on Intel’s public roadmaps in mid-2015 when the company realized that Cannonlake and its attendant 10nm manufacturing process wouldn’t be ready in time.
Many PC models are refreshed once a year or so, and that means that the PC makers need new stuff to put into those computers whether or not the laws of physics want to comply.Įnter “Kaby Lake,” the first of Intel’s post-tick-tock processor architectures. Intel’s tick-tock model may be dead, but the PC industry still demands new hardware every year.
0 kommentar(er)
