Intel has officially launched its new Core Ultra 200V series "Lunar Lake" processors, to take on the next-generation thin-and-light laptop market against AMD and Qualcomm.
The company has tweaked virtually everything under the hood of Lunar Lake, with new P-Cores and E-Cores, a massively upgraded integrated GPU, and even more AI performance. Intel is also debuting on-package memory with its new Core Ultra 200V series "Lunar Lake" CPUs -- RAM that is next to the chip itself.
Intel is using its new Lion Cove P-Cores for more performance than the previous-generaiton, and Skymonth E-Cores that are tuned for efficiency. For Lunar Lake, Intel is using Low-Power variants of its E-Cores called LP-E, which sit on a dedicated low-power island.
Intel is using Thread Director to provide huge efficiency gains at low-power using the low-power island, as well as a mix of magic from TSMC and its N3B process node, of course. Intel compares its new flagship Core Ultra 9 288V "Lunar Lake" CPU against Qualcomm's X1E80100 processor and its previous-gen Core Ultra 7 165H "Meteor Lake" CPU with the new 288V having a 7% performance gain, while offering an impressive 2.29x improvement in performance-per-watt.
The company is talking up a huge 2x gen-on-gen performance-per-watt gain with Lunar Lake compared to Meteor Lake in flagship form, with games like Cyberpunk 2077 seeing a huge improvement from the beefed-up CPU power, but also 50%+ more performance from its new Xe2 "Battlemage" GPU. The integrated GPU inside of the Lunar Lake SoC is 44% faster, and uses 22% less power. Impressive stuff.
Overall, Intel's new Core Ultra 200V series "Lunar Lake" CPUs use up to 50% lower package power than Meteor Lake, and that's with the new 32GB of LPDDR5X-8533 on-package memory.
Battery life is one of the most important things with a thin-and-light laptop, and Intel isn't disappointing with its claims of up to 20 hours of battery life in UL Procyon office productivity tests, and up to 10.7 hours in Microsoft Teams 3x3. That's a 39% improvement over AMD's new Ryzen AI 9 HX 370 "Strix Point" APU and a 47% improvement over Qualcomm's X1E-78-100 chip.
Another big note is that Lunar Lake is using an all-new, low-latency fabric that provides the following inter-core and DRAM latencies:
- E-Core to E-Core = ~23ns
- P-Core To P-Core = ~26ns
- E-Core To P-Core = ~55ns
- DRAM Latency = ~90ns
Intel's new Core Ultra 200V series "Lunar Lake" processors ship with next-gen Xe2 "Battlemage" integrated GPUs that offer some gigantic improvements in performance, right after AMD launches its Ryzen AI 300 series "Strix Point" APUs with upgraded RDNA 3.5-based Radeon 890M GPUs.
The new Xe2 "Battlemage" GPU has a 50% performance increase over Xe "Alchemist" inside of Meteor Lake, with more improvements in ray tracing through improved ray tracing units, better utilization of the CPU cores versus Alchemist Xe-LPG GPUs.;
Intel compares its Xe GPUs on the Core Ultra 7 155H "Meteor Lake" CPU against the new flagship Core Ultra 9 288V "Lunar Lake" CPU in a bunch of games, delivering a 31% uplift in performance, and in some games, over 60-70% more performance over Alchemist. Very nice to see.
The company compares its new Xe2 "Battlemage" GPU inside of Lunar Lake against Qualcomm's integrated Adreno GPU inside of its new Snapdragon X processors, with the Adreno GPU failing to run 23 of the games in Intel's test suite. This is because Qualcomm's Snapdragon processor is an Arm-based chip, not capable of running games made for x86 (regular processors from Intel and AMD).
The bigger comparison is between Battlemage and RDNA 3.5, with Intel claiming to have its Xe2 GPU being 16% faster than the Radeon 890M inside of Strix Point. This is the new Arc A140V GPU, compared against the Radeon 890M, and the new Xe2 "Battlemage" GPU is keeping up, or besting the RDNA 3.5 GPU in some games. Another impressive feat for Lunar Lake.
It's not just raw GPU hardware anymore, but the software and graphics division have forged ahead with Battlemage and Xe2 launches with over 120 games that support XeSS upscaling.
This means that the integrated GPU inside of Lunar Lake will have faster performance through its new Xe2 GPU architecture, but also improved performance again through XeSS through the beefed-up XMX hardware on the SoC compared to Meteor Lake.
With XeSS enabled and on Performance Mode, Lunar Lake is capable of a 62% performance improvement, and better yet, if you're hitting 80FPS to 100FPS, you could take the Quality Mode preset and enjoy a stable 60FPS with a better-looking game.
Intel wouldn't launch a new GPU architecture without mentioning ray tracing, with Xe2 packing 30% faster RT performance and a 99% percentile over 30FPS. The new Xe2 RT engine is a fully functional RT implementation, running through the DirectX12 Ultimate API.
All in all, you've got a huge upgrade in everything: CPU, GPU, NPU, media and display, I/O, power efficiency, ray tracing, and more. Intel's new Core Ultra 200V series "Lunar Lake" CPU sound quite enticing, and I can't wait to try them out now that I've been playing with AMD's new Strix Point APU for a couple of weeks.
Intel, impressive to see... can't wait to have it in the lab!