In the over two decades that I’ve been covering the PC space, I have rarely seen such confidence in a next-generation product as I saw from Intel execs at the IFA conference in Berlin last week. To say the company is assured on the performance and power characteristics of its forthcoming Core Ultra 200V Lunar Lake laptop processors would be an unquestionable understatement.
Last week, at one of the oldest consumer electronics industry shows, analysts and press were given hands-on access to Lunar Lake Core Ultra laptops from various OEM partners, along with a deep-dive view of Core Ultra 200V Series processor performance, delivered in a series of presentations and Q&A sessions. The promise of what Lunar Lake will deliver was very strong, so much so that competitors like Qualcomm immediately began picking apart some of the compares. This was to be expected of course, but let’s break down what you can expect from Lunar Lake and read between the lines a bit along the way as well.
Intel Core Ultra 200V: Rearchitected From The Ground-Up For Perf-Per-Watt
As I’ve covered previously here, Intel’s new Lunar Lake CPU architecture almost completely flips the script on previous-generation Intel mobile CPU designs. Gone is Intel Hyperthreading for multithreaded workloads, in favor of an all single-threaded 8-core design with four Performance Cores and four Efficiency Cores. However, many areas of the chip have been optimized and rearchitected to better feed Lunar Lake’s Skymont E-Cores and Lion Cove P-Cores, including a much lower latency chip fabric, bringing core-to-core and system memory latency down as much as 40% versus Intel’s Meteor Lake Core Ultra 100 series, along with wider, more dynamic scheduling in concert with Intel Thread Director technology. The result is an over 68% improved Instructions Per Clock for Core Ultra 200V’s quad E-Core complex and a greater than 14% lift in IPC for its P-Cores with up to 3X performance-per-thread processing throughput versus Intel’s previous gen Meteor Lake chips.
And with respect to Lunar Lake’s Arc Battlemage graphics engine, Intel’s claims are even more bold, exceeding the competition, including AMD’s typically dominant position in integrated graphics, by over 68% versus Qualcomm Snapdragon X Elite and over 16% ahead of AMD’s new Ryzen AI 370HX chips.
Finally, Intel closed the gap on its NPU performance with the introduction of a new 48 TOPS neural engine that will certainly qualify it for Copilot+ PC status with Microsoft. However, it’s what all these speeds and feeds mean in terms of performance and battery life that matters, and with respect to that representatives from Intel were not shy. I got a chance for a quick impromptu interview with Intel’s VP of Client Computing and GM, AI Technical Marketing, Robert Hallock and here’s what he had to say…
Lunar Lake Performance: The Proof Is In The Bar Chart, And Independent Testing
It’s almost pointless to wax poetic about Intel’s Lunar Lake performance claims versus the competition, so instead I’ll let a few bar charts do the talking and promise that our teams at HotTech and HotHardware will both be digging into specific retail ready laptops powered by Lunar Lake in the days ahead. First, let’s look at all out performance, then we’ll look at graphics and energy efficiency.
I’m not big on pure synthetic benchmarks like Geekbench, for example, but prefer benchmarks that utilize applications that people use every day. Here, there’s a mix of both, with Cinebench (based on Maxon 3D) being my favorite for the single-core throughput compare Intel makes, and UL Procyon Office and Photoshop for what I’d personally look to in the productivity and content creation workload tests. What we don’t see here is a Cinebench multi-threaded test result, which would likely show a bit softer for Intel’s new Core Ultra 9 288V 8-core chip versus Qualcomm’s 12-core Snapdragon X Elite X1E-80-100 and AMD’s 12-Core Ryzen AI 9 370 HX processor as well. The other note is that Qualcomm’s X1E-84-100 is not shown in these results, with its 100MHz higher dual-core boosts, though I don’t think it would change this performance compare dramatically. Regardless, the numbers look strong for Intel’s forthcoming Core Ultra 200V Series, again, when you look at real world application workloads. Gaming and graphics workloads, on the other hand, could offer a more stark differential.
This graph isn’t completely fair for Qualcomm’s Snapdragon X Elite showing of “DNR” (Did Not Run), since there are plenty of games that do run on Snapdragon, like Diablo IV, Overwatch, Fallout 4 and DOTA 2, just to name a few. That said, I chose this compare because Qualcomm’s gaming performance doesn’t currently offer the strongest competitive numbers while AMD’s legacy of powerful Radeon integrated graphics engines are a taller yardstick to measure by. Here, Intel is claiming a 11 – 63 % lead over AMD, for what’s shown here in these game titles. I’d look more closely at the Cyberpunk and F1 24 scores, as they’re more modern game titles, which would bring this spread down to 11 – 32% or so, which regardless is stout.
All in, Intel is claiming an average of 16% more performance for gaming versus AMD’s Ryzen AI 300 series, though the thin and light machines that Intel’s Core Ultra 200V Series will power aren’t gaming-centric laptops by any means.
What about battery life you ask? Again, I’ll let Intel’s claims do the talking (for now)…
In this compare, Intel actually takes a lower power SKU of Qualcomm’s Snapdragon X Elite X1E-78-100 that has no dual-core boost, and compares it to their high end Core Ultra 9 288V, though the system specs do note a generous 75Whr battery for all machines. Again, these metrics are prominent claims for Intel, and with 14 hours of always-on productivity performance and almost 10 hours of Microsoft Teams calls, there are now high expectations for Lunar Lake powered laptop battery life. Intel is also claiming over 20 hours of battery life with lower power Core Ultra 7 268V chip-enabled laptops as well, in similar workloads. These battery life and efficiency gains are big, gen-on-gen, and put Intel squarely back on the map in terms of power efficiency.
Intel Core Ultra 200V Laptops, Arrow Lake Desktops And The Road Ahead
Intel’s bold claims for Lunar Lake performance, in the age of the Windows Copilot+ AI PC, are probably best punctuated by this slide showcasing available AI throughput of its CPU, GPU and NPU engines, and the types of workloads they are best suited for. All told, Core Ultra 200V Series will deliver 120 total platform TOPS and comparatively, as we’ve seen in these Lunar Lake benchmark claims, Intel should have a strong entrant in the Copilot+ PC market, though I am itching for our team’s direct performance validation results as well.
What will be key moving forward, is Intel’s software enablement strategy and execution, which as of late has been particularly strong, as evidenced by the company’s advancement in graphics performance through driver optimizations. Qualcomm’s concerted effort with Microsoft in the enabling Windows on Snapdragon X Series has also been impressive though frankly, AMD needs to play catch-up on enabling apps to take advantage of its Ryzen AI NPU. Regardless, I expect Intel will continue to be a leader in software and ecosystem enablement, as Lunar Lake represent a much needed shot in the arm for the company right now.
Beyond Lunar Lake, lies Arrow Lake, Intel’s next generation desktop platform processor offering that the company notes will now be fabricated at TSMC. The early rumblings and leaks we’re hearing about Arrow Lake CPUs are also very encouraging, along with Intel’s welcomed early advancements in its own 18A chip process node that will take the company to higher ground for its own chips and Intel Foundry services as well.
Q4 should be a very interesting time in the semiconductor and computing industries. Anyone talking about an “AI bubble” needs to wake up and smell the coffee and this fresh new batch of intelligent chips as well.
