Intel launches Core Ultra 200V series: “the most efficient x86 processor ever”
Intel has launched its second-generation Core Ultra 200V series processors, codenamed Lunar Lake. Designed for thin and light laptops, Intel claims that Ultra 200V CPUs are faster than Qualcomm Snapdragon chips in every key area – while still promising up to 20 hours of battery life.
The new generation of chips can be identified by a V suffix with nine different processors at launch. These range from the Core Ultra 5 226V to the Core Ultra 9 288V. Intel also confirmed that higher performance second-generation Core Ultra chips will follow in due course.
Laptops with new Core Ultra 200V chips inside can be pre-ordered today, including the Dell XPS 13, with shipments starting on 24 September.
Most efficient x86 processor ever
Intel announced the new chips at an invite-only event in Berlin leading up to IFA 2024. “Our engineering team set huge goals in designing the product,” said Josh Newman, Intel VP Client Computing Group. “Breakthrough x86 power efficiency, a massive leap in graphics and unmatched AI compute.”
He added: “These processors not only met the goals, they exceeded them. In fact, we’re proud to say that lunar Lake has achieved history as the most efficient x86 processor ever… and the absolute best AI performance for laptops.”
Robert Hallock, General Manager of AI Technical Marketing, was equally enthusiastic, describing the Ultra 200V as a processor “that really stands out” because “there’s so much that’s new”.
What’s new about the Ultra 200V chips
Intel made the big decision to drop the number of cores and threads in the Core Ultra 200V series. Every member of the new family features four performance P-cores and four efficiency E-cores, so they are differentiated by how many watts. What’s more, the P-cores don’t include multi-threading, so the total number of threads is eight.
That compares to up 22 threads in first-generation Core Ultra chips, with Intel countering this lower figure with the promise that its P-cores are the fastest out there.
It’s important to note that this graph shows single-core performance. In tasks that rely on multiple cores, I would expect the AMD Ryzen AI HX 370 processor and both Qualcomm Snapdragon chips to be faster.
Notably, none of the tests that Intel later shared with us compared multicore performance. Instead, it was keen to focus on AI results, gaming speeds and performance per watt.
Improved efficiency = greater battery life
The latter makes particular sense, as the Core Ultra 200V chips have more emphasis on power efficiency than sheer performance. These chips are intended for thin and light laptops, after all.
Intel has put a huge amount of work into reducing the power consumption of the total package, with the most efficient version of the chip consuming 9W at peak and even the most power-hungry 33W.
Judging from figures released by Dell, this can result in phenomenal battery life when backed with the right components. At its best, Dell reckons the new Inspiron 14 – with a Core Ultra 7-256V, 16GB of RAM and a low-energy display – provides 26 hours of Netflix viewing and 30 hours of local video playback.
As this slide from its launch presentation shows, Intel is promising similar levels of battery life to Qualcomm laptops. In the UL Procyon benchmark, which is based on Microsoft Office, even longer.
Once more, we shall see when we get a chance to test these laptops for ourselves.
Gaming performance part 1: game over vs Snapdragon
Intel took great glee from kicking Qualcomm laptops where they hurt the most: gaming performance. Few Windows-based games have been ported to the Snapdragon’s Arm architecture, and that means they must go through a translation layer (built into Windows on Arm) to work. That translation layer actually works fantastically well for most software, but games demand more.
And that’s why Intel’s team had huge smiles on their faces when they showed off this graph:
I’ll save you the bother of zooming in. In the games (roughly one third) where Qualcomm-powered laptops can run, Intel’s updated GPU whips its competitor hard.
“Of the 45 games that we tested, more than half did not run on Snapdragon,” he said, a smile on his face. “As you can imagine, that has made it somewhat challenging to give you the tidy performance summary you would expect from a clean and polished slide.
“If we turn on the slide for the games that do happen to run on Qualcomm, Lunar Lake [Core Ultra 200V] is 70% faster.”
Compatibility and performance will improve over time, but no-one who’s interested in playing graphically demanding games should choose a Snapdragon laptop right now.
Gaming performance part 2: vs AMD
Qualcomm aside, the results remain compelling. “We’re averaging about 30% faster across these 45 titles [versus first-generation Core Ultra laptops],” said Hallock, “and the top quartile of these games runs 70% faster.”
Intel’s testing shows that it’s faster than AMD’s rival chips too:
At the launch, Intel showed laptops benchmarking Shadow of the Tomb Raider and Cyberpunk 2077 at 1080p Medium settings. Plus 3DMark Solar Bay. You can see the results for yourself below.
The AI argument
Things are more nuanced when it comes to AI. Qualcomm leans heavily on the NPU (neural processing unit) when making AI claims, but most real-world software that uses AI actually relies more on the GPU. And, as we can see from the above, Intel has the clear lead here.
It also has the lead for software partnerships. Intel has been working with Adobe for decades now, while Qualcomm’s relationship with the creative software giant is far more nascent. That almost certainly means that Adobe will bring out features optimised for Intel software first.
And that’s just as true for AI features as any other.
Intel also reckons it wins for performance in relative newbie Stable Diffusion, as the graphs above indicate. How much this matters to the average user is far more questionable.
Full list of new Core Ultra 200V processors
Max P-core frequency | Max E core frequency | Smart cache | Xe graphics cores | Max GPU frequency | XMX AI PTOPS | NPU AI TOPS | Memory capacity | Base power | |
Ultra 9 288V | 5.1GHz | 3.7GHz | 12MB | 8 | 2.05GHz | 67 | 48 | 32GB | 30W (min 17W) |
Ultra 7 268V | 5GHz | 3.7GHz | 12MB | 8 | 2GHz | 66 | 48 | 32GB | 17W (min 8W) |
Ultra 7 266V | 5GHz | 3.7GHz | 12MB | 8 | 2GHz | 66 | 48 | 16GB | 17W (min 8W) |
Ultra 7 258V | 4.8GHz | 3.7GHz | 12MB | 8 | 1.95GHz | 64 | 47 | 32GB | 17W (min 8W) |
Ultra 7 256V | 4.8GHz | 3.7GHz | 12MB | 8 | 1.95GHz | 64 | 47 | 16GB | 17W (min 8W) |
Ultra 5 238V | 4.7GHz | 3.5GHz | 8MB | 7 | 1.85GHz | 53 | 40 | 32GB | 17W (min 8W) |
Ultra 5 236V | 4.7GHz | 3.5GHz | 8MB | 7 | 1.85GHz | 53 | 40 | 16GB | 17W (min 8W) |
Ultra 5 228V | 4.5GHz | 3.5GHz | 8MB | 7 | 1.85GHz | 53 | 40 | 32GB | 17W (min 8W) |
Ultra 5 226V | 4.5GHz | 3.5GHz | 8MB | 7 | 1.85GHz | 53 | 40 | 16GB | 17W (min 8W) |
Intel Core Ultra 200V series: early verdict
Intel is going through troubled times, having to lay off around 15,000 people from its workforce to keep shareholders happy. It needs a win. And guess what: the Core Ultra 200V series is exactly that.
Clearly, it isn’t going to win awards for outright multicore performance. AMD’s Ryzen AI 300 series currently have that lead, and I suspect that Qualcomm’s 12-core chip will be significantly speedier too.
However, the Core Ultra 200V series is targeted at thin and light laptops. And from everything I’ve seen so far, the chips makes a compelling choice. Intel reckons that come 24 September we’ll see over 80 “unique designs” of laptops from over 20 laptop makers. That includes all the big names.
Until I test laptops for myself, using benchmarks of my own choosing, I will hold off on a full verdict. But the early signs are almost entirely positive.
Related:
NEXT UP
Kees Wolters, Chief Product Officer and Founder at Mopinion: “To fuel your creativity, you need a solid foundation in design principles”
We interview Kees Wolters, Chief Product Officer & Founder at Mopinion a user feedback software for websites, mobile apps and email campaigns
Cisco ramps up partner training with $80m investment
Cisco announced at its 2024 partner summit an investment of US$80 million to help partners upskill their workforce through Cisco U
Whisper it, but genuinely useful local AI has already arrived on laptops
Local AI is coming to laptops, but can it actually perform tasks that will save you time? We allow HP’s new OmniBook Pro to take us out to lunch to find out