Sources: The Verge
Intel Lunar Lake AI laptop chip
Last year, Intel boasted that its Meteor Lake processors, dubbed Core Ultra, represented the company’s biggest architectural shift in 40 years. But Intel didn’t settle down after that: today it is revealing how Lunar Lake, its next laptop chip coming this fall, will overhaul the formula yet again.
The AI-powered PC opportunity, Intel has seemingly ditched its famous tick-tock cadence for an entirely new system-on-chip design, one that not only triples the size and more than quadruples the performance of its AI accelerator, but promises up to 14% faster CPU performance at the same clock speed, 50% better graphics performance, and up to 60% longer battery life than last year is model.
Rob Hallock, who claims Intel tweaked every part of the chip to achieve this, says it will definitely outperform Qualcomm as well.
Is the change a big one? If you buy a Lunar Lake laptop, it won’t have any flash memory or discrete chips! Lunar Lake now includes either 16 or 32GB of LPDDR5X memory on the package itself, with no way to plug in more RAM. It is a change that reduces power consumption when moving data through the system by about 40%, according to Intel. For those who need more memory, Hallock says a standalone Arrow Lake architecture is coming to laptops later this year.
No hyperthreading
Last year’s Meteor Lake packed a wild new 3D performance hybrid architecture with plenty of Performance (P), Efficiency (E), and even a pair of new Low Power (LP-E) cores in a separate tile called the Low Power Island.
That island was built like a smartphone a first for Intel with its own Wi-Fi, Bluetooth, display drivers, memory and those low-power CPU cores. The idea: It could theoretically save battery life by never heating up other tiles and larger cores unless it is performing more important tasks.
Applications like Microsoft Teams ended up heating up the entire chip. That is why Intel is removing LP-E cores after just a year, in favor of a new 4×4 system. You get up to four new Lion Cove P cores and four new Skymont E cores on a Lunar Lake chip. Those E cores now run as fast as one LP-E core at a third of the power, or scale to 2x or 4x the performance (single-threaded vs. multi-threaded).
Windows can now create containment zones that actually keep most real-world workloads on the Skymont E cores, Microsoft and Intel say.
This is key to Luna Lake’s battery life – we can run more workloads in a lower power environment on a lower power core with fewer things on and still give you a great user experience, says Intel Fellow Stephen Robinson.
More performance
Capable E cores, Intel’s Skymont has another surprise: This year’s E core is more powerful and efficient than last year’s P core at typical laptop clock speeds, with up to 20% more single-threaded performance. It simply can’t scale to that many gigahertz.
Lion Cove’s four P cores will see a 14 percent clock-to-clock performance increase, though Intel didn’t provide clock speed figures. But overall, Hallock says, performance has jumped very significantly on a generational level compared to last year.
Intel is even more confident: The company claims its Xe2 GPU offers 1.5 times the graphics performance of Meteor Lake (in 3DMark Time Spy), which in turn was twice the performance of the previous generation. It still has the same number of Xe cores and other functional units, but with a variety of performance and efficiency improvements.
Surely not all Lunar Lake chips will be created equal, Intel says it didn’t have to split its Xe laptop GPU into two different versions for higher and lower powers Xe2 can now scale across the entire light- and mid-range laptop spectrum on its own. Intel also says the GPU offers 67 TOPS of AI performance, on top of the NPU.
NPU
Meteor Lake didn’t usher in the AI-powered laptop generation in the way Intel had hoped. In fact, it arguably left early adopters in the lurch: with just 11.5 TOPS of AI acceleration, its NPU falls far short of Microsoft’s 40 TOPS requirement for Copilot Plus PCs.
Leave a Reply