In the most recent press conference, Intel outlined its upcoming line of processors designed for mobile devices under the codename Lunar Lake. Much like their previous Meteor Lake release, the adjustments made to these processors were quite substantial. As part of these alterations, Intel revamped every aspect of its processor architecture, introducing new performance-enhancing and efficiency-boosting cores while also phasing out low-power core technology. Furthermore, they have introduced notable advancements in GPU and NPU elements within their product offerings.
Lunar Lake presents a simpler design, with only two of the four tiles having significant roles in its functionality. The compute tile, fabricated on TSMC’s advanced 3nm-class process, and the platform controller tile, manufactured using an older 7 nm process. Additionally, there is a “filler” tile, which acts as a structural blank piece of silicon designed to prevent chip bending. This design is mounted on top of a passive interposer or “base” tile that provides essential interconnections between the individual chips.
Similar to Apple’s M3 Macs, Lunar Lake will also have on-package memory. They can either have 16GB or 32GB of LPDDR5X memory.
While there is the obvious disadvantage of making memory upgrades unviable, the bandwidth has been considerably expanded, providing support for as much as 8,500MTS. In addition to this enhancement, there has been a notable decline in energy consumption during idle mode, amounting to approximately a 40% decrease. Furthermore, the space taken up by memories has also undergone a substantial reduction.
Intel’s Lunar Lake makes two major changes to the CPU designs. First, what’s known as the “Skymont” efficiency core no longer has the low-power E-core that its predecessor, Meteor Lake, shipped with.
Hyperthreading has been completely disabled. All cores simply have a single thread associated with them for performance reasons.
Lunar Lake has four E-cores and four P-cores. Lunar Lake’s E-core has a number of substantial architectural enhancements — wider machine decoding and out-of-order engines, a 4MB level-2 cache shared among all four cores.
And now with hyperthreading completely disabled, the performance core within Lunar Lake, Lion Cove, is 14 percent faster than the P-core within Meteor Lake, known as Redwood Cove.
The Lunar Lake introduces the fourth generation of Intel’s AI acceleration technology. This iteration boasts impressive enhancements, exhibiting a staggering 4-fold increase over its preceding counterpart found in the Meteor Lake model.
This significant enhancement in performance is crucial as it aligns the Neural Processing Unit (NPU) of Lunar Lake with the exacting demands of Windows CoPilot+, a system that requires no less than 40 TOPs (trillions of instructions per second) to function optimally. Notably, the NPU alone satisfies this requirement, boasting an exceptional performance output of 48 TOPS.
Intel says a big wave of Lunar Lake laptops will arrive later this year, with 80 different designs across 20 hardware partners at launch.
