Intel has just unveiled its new generation of processors for notebook PCs, aka Lunar Lake. We’re going to have to get used to the new name, as we’re now talking about Core Ultra 200V. First of all, please note that this article is intended as a quick summary to give you a few key points of understanding. We’ll be back with more details shortly.
A more important step than it seems for Intel
This processor represents a second milestone in Intel’s evolution of processor design. After the first breakthrough with Meteor Lake, Lunar Lake is a decisive second step. This generation will feature a high degree of aggregation, including memory-on-package (MoP), as with Apple Silicon SoCs. The processor comes with up to 32 GB of LPDDR5X memory integrated into the package. Understand, then, that we are clearly approaching the SoC (System on Chip) model for Intel mobile chips.
Intel Lunar Lake: several revolutions in a single chip
We’ll use the term “revolution” here to highlight the major change taking place in chip design at Intel. Within Intel Lunar Lake we’ll find 4 key components that have received a profound update.
- First, the Core CPUs
The Core Ultra 200V is designed to compete with the much-discussed Snapdragon chipset (Elite X), as well as Apple’s M3. The application therefore seems to be limited to ultra-portables. The CPU part will rely on new Lion Cove performance cores (P-cores) and efficient Skymont cores (E-cores). To keep things simple, the story that’s going to dominate the press is that Intel hasn’t equipped Lion Cove with Hyper-Threading Technology (HTT). E-Core Skymonts seem to have been the focus of Intel’s attention. This means a huge 68% increase in IPC compared to the Crestmont used on Meteor Lake. It’s this leap forward that makes Intel confident about the minimal impact of removing Hyper-Threading on its P cores.
- iGPU: the arrival of Battlemage
Regularly criticized for lagging behind AMD and APPLE, the graphics department has been on the up since the arrival of Arc. A new milestone is about to be reached, as the iGPU powering Lunar Lake is claimed to offer a 50% gain in gaming performance over the Xe-LPG iGPU powering the Meteor Lake processor. The iGPU meets all DirectX 12 Ultimate specifications and includes 8 next-generation raytracing units.
- NPU: huge performance boost
The reference point is the NPU used in Meteor Lake. The latter is known to be insufficient to drive Copilot within Windows. Lunar Lake’s NPU is four times more powerful than its predecessor. From 12 TOPS on Meteor Lake to 48 TOPS on Lunar Lake. This meets all the requirements set by Microsoft to accelerate local Copilot sessions and qualify for Copilot AI PC certification.
Lunar Lake: a brief summary and some thoughts.
It’s essential to bear in mind that Intel is essentially targeting Apple and Qualcomm with this SoC. Efficiency and autonomy are the two keys to understanding this architecture. With this in mind, it’s easy to understand why Intel has “contented” itself with minor transformations and optimizations for its P cores, and focused all its attention on the performance of its E cores. Skymont represents a major step forward for Intel.
And that’s perhaps the most important thing to remember. After all, Skymont and its derivatives could determine a lot of things for future generations. During its presentation, Intel showed a comparison between a Core Skymont (E-core) and a Core Raptor Cove (P-Core). If we give credence to this information, then Skymont is on a par with the Raptor Cove IPC. And this is achieved with a fraction of the number of transistors, and therefore a considerable saving in power consumption…
Arrow Lake will benefit from Skymont cores in greater numbers and without the cache limitations of Lunar Lake. Until I became aware of all this, I had certain questions about the relevance and performance of Arrow Lake compared with AMD chips and their multithreaded performance… But that was before.
