Intel lifted the curtain on its newest processor sequence, codenamed Lunar Lake, throughout Intel CEO Pat Gelsinger’s Computex Keynote on Wednesday. The brand new chip will likely be launched later this 12 months, powering skinny and lightweight laptops in time for the 2024 vacation season.
The brand new system-on-a-chip (SoC) introduces various improvements over the current-gen Meteor Lake SoCs just like the Intel Core and Core Extremely processors, with Intel claiming higher energy effectivity, a 4x enhance in NPU processing energy, a 50% quicker Intel Arc GPU based mostly on its next-gen Arc Battlemage structure, and on-chip reminiscence for a smaller footprint and quicker reminiscence entry.
Lunar Lake can be the primary Intel chip sequence to fulfill the minimal necessities for Microsoft’s Copilot+ PC requirements, together with larger than 40 trillion operations per second for its NPU (TOPS), and 16GB reminiscence.
The brand new SoC incorporates a extremely revamped structure that introduces a smorgasbord of small modifications to how its processors perform that collectively obtain vital efficiency enhancements whereas not utilizing any extra energy than present Intel Core Extremely chips, and may reportedly get the identical efficiency as Intel Meteor Lake chips with a 3rd of the ability, significantly enhancing the potential battery lifetime of laptops outfitted with the brand new chips.
I haven’t examined any of those claims myself, so they need to all be taken with a grain of salt till reviewers are capable of get their arms on new laptops with Lunar Lake chips in them, however in an prolonged tech tour forward of the Computex announcement, Intel engineers introduced a reasonably thorough walkthrough of the modifications that Lunar Lake will introduce and they’re pretty far-reaching, even when particular person modifications don’t transfer the needle a complete lot on their very own.
New core architectures launched
The largest innovation in Lunar Lake is the introduction of recent hybrid core structure with Lion Cove efficiency cores and Skymont effectivity cores. These cores are a serious driver of the improved energy effectivity that Intel is claiming for Lunar Lake, with various small modifications to the best way directions are processed by the CPU, the introduction of a brand new layer of cache reminiscence, and a doubling of the variety of effectivity cores over Meteor Lake.
One other main change is the lack of hyperthreading functionality. That is when a processing core is ready to perform two directions per clock cycle moderately than one, although Intel says that the optimization that has gone into Lunar Lake’s Lion Cove efficiency cores greater than makes up for the lack of hyperthreading, which comes with its personal processing overhead.
All of those modifications, taken collectively, add as much as effectivity cores that Intel says ship the identical efficiency because the current-gen Meteor Lake chips for a 3rd of the ability, and as much as 4 occasions the efficiency at peak energy due to the addition of two extra effectivity cores. On the efficiency core facet, Intel says Lunar Lake’s Lion Cove efficiency cores are capable of get 50% higher efficiency for a similar energy as Meteor Lake’s Redwood Cove cores in single-threaded efficiency, and about 5% higher efficiency with a single-threaded Lion Cove core than Meteor Lake’s multi-threaded efficiency cores.
What’s extra, Intel says that its effectivity cores are mature sufficient that they will deal with extra demanding duties than earlier generations of effectivity cores, that means that workloads that had been beforehand assigned to higher-powered efficiency cores can now be assigned to lower-powered effectivity cores, growing the chip’s total effectivity which ought to assist prolong a Lunar Lake-equipped laptop computer’s battery life to be aggressive with Arm-based laptops like these powered by Qualcomm’s new Snapdragon X and Apple’s M3 and M4 chips.
Expanded NPU and next-generation Intel Arc graphics
Different notable advances come from the opposite two main parts of the Lunar Lake SoC, specifically the Intel NPU 4 neural processor and Intel Xe2 graphics structure, based mostly on Intel Arc Battlemage.
Beginning with the GPU, the next-gen Arc graphics comes with 8 Xe execution cores, 8 ray tracing cores, a devoted matrix AI engine able to 67 TOPS, and guarantees as much as 50% higher graphics efficiency over Meteor Lake’s Arc Alchemist-based structure.
The NPU, in the meantime, quadruples the efficiency of Intel Meteor Lake’s NPU 3 neural processor, leaping from 12 TOPS to 48 TOPS due to 4 extra neural processing cores and different architectural refinements.
For individuals who are counting, this provides as much as a complete platform TOPS of 120 between the GPU, NPU, and CPU, and when Lunar Lake laptops roll out in Q3 and This fall of this 12 months, Home windows Copilot+ Runtime and different developer APIs will enable app makers to leverage all three processors as wanted to higher optimize AI duties.
Whether or not Intel Lunar Lake delivers on the promise of the AI PC this 12 months stays to be seen, however Intel is projecting lots of confidence that by 12 months’s finish, it’ll take a commanding lead within the AI PC market, nevertheless that market finally shapes up.