Intel claims that the x86 architecture is not finished, and far from it. After months of providing bits and pieces of CPU details, Chipmakers in trouble lately has shared information about its Lunar Lake series, which is now officially called Intel Core Ultra 200V. Less than a year ago, Intel first brought us the term “AI PC” and its Meteor Lake chip. Now, instead of emphasizing the promise of NPU processing, its plan to regain mobile chip supremacy is to point out just how annoying Qualcomm’s ARM-based chips are.
Intel calls its Core Ultra 200V series “the most efficient x86” chips yet. The key measurement is a claimed 40% reduction in PHY power consumption compared to last year’s Meteor Lake. It should use 41% to 33% less battery when using Microsoft Teams or YouTube. In certain productivity tasks, power efficiency may be higher. Intel said its PCs had 20 hours of battery life on the Procyon benchmark compared to the Snapdragon The time is a little over 10 hours.
This time there are a variety of chips to choose from. The Lunar Lake series ranges from low-power Intel Core Ultra 5 to top-tier Core Ultra 9 288V. Regardless, you still get a chip architecture that includes four efficiency cores and four power cores with eight threads. In addition to power consumption, the P-core of the low-end CPU has a maximum frequency of 4.5 GHz and has an 8 MB cache. Ultra 9 288V has a maximum speed of 5.1 Ghz and 12 MB cache. Unfortunately, Hyper-Threading is not integrated this time around. Intel claims this improves energy efficiency.
Intel is also splitting its ARC integrated GPU for the first time. If you choose Intel Core Ultra 5, you get a 130V GPU. Any Intel Core Ultra 7+ version comes with a 140V GPU. The lower-end chip only has 5 Gen4 NPUs at 40 TOPS. A chip equipped with 6 Gen4 NPUs can achieve 47 TOPS or more. Either way, they should meet the criteria set by Microsoft to qualify for Copilot+ PC, in case you are one of the few who want to qualify for Copilot+ PC Finally using Windows 11 Recall.
We’ll be hearing news about the OEM’s new Lunar Lake laptops in the next few days. What Intel most wants to defeat is AMD’s Strix Point lineup anddefeating the ARM upstart to a greater extent, Qualcomm Snapdragon X Elite chip.
Since the emergence of “AI PC”, Intel has talked more clearly about Lunar Lake’s NPU
As for AI processing, Intel says the Core Ultra chip is 20 percent better than the Intel Core Ultra 7 155H at OIDN denoising using Blender, and much better at AI upscaling or slow-motion effects in some applications. The company likely chose these tests to showcase its new CPUs, especially since ARM-based Snapdragon chips can’t run Blender without simulation.
The artificial intelligence processing we are talking about today is very different Last December, Intel CEO Pat Gelsinger (who was trying to reassure concerned employees and investors with Bible verses) took to the stage to preach the glory of the NPU (Neural Processing Unit). Intel hampered by power issues with its gaming-side 13th and 14th generation products wafer, causing repeated failures. Intel promises customers these latest mobile chips won’t have the same problem.
Robert Hallock, vice president of Intel’s Client Computing Group, said on Tuesday that the NPU is more powerful. Still, it’s becoming another aspect of processing on these new mobile chips, the division of processing between CPUs, GPUs and NPUs. He compared it to the introduction of integrated GPUs, which were once “ridiculed” but are now an expected part of CPU processing.
He added that what computing experts and journalists have accomplished over the past year: TOPS, or trillions of operations per second, Not a good way to compare neural processing ability. TOPS is a value derived from a simple mathematical equation that includes parameters such as speed and accuracy.
We should also point out that with the advent of AI PCs, NPUs are nothing new. They’ve been part of mobile phone designs for years. When Intel was founded nearly 20 years ago, it never felt the need to create an entirely new brand for graphics-capable PCs. While the chipmaker didn’t say NPUs were overhyped, Hallock did say in the Q&A that there’s a “natural tendency” to focus on introducing dedicated neural processing. Application developers now rely equally on all three aspects of modern CPUs, even for AI tasks.
Lunar Lake’s GPU may not be built just for AI tasks, but it can handle 67 TOPS compared to the NPU’s 48 TOPS. In Intel’s demonstration, its new chip ran a stable diffusion artificial intelligence art generator, with the CPU relying more on GPU processing than the NPU. AI clip segmentation in Adobe Premiere relies more on the CPU than the NPU.
Lunar Lake is famous for the fact that you don’t have to emulate anything on x86
Intel claims Lunar Lake is beating benchmarks, but we’ll need to get our hands on the first laptops to verify those numbers. As for gaming performance, Intel claims its chip has better integrated graphics than Strix Point or Snapdragon. The company shows that Core Ultra 9 288V can improve FPS by 10 to 20 F1 2024 On high settings, there is no upgrade more than high level AMD Ryzen AI 300 Compared to Intel Core Ultra 7 155H, FPS is increased by about 30 times.
It’s similar in more games, with some impressive FPS on high settings, even if they don’t beat other games laptop Comes with independent graphics card.
Intel is also quick to point out that the vast majority of games built for the x86 architecture won’t work without emulation. Microsoft and Qualcomm were the first to tell consumers that the first Copilot+ PCs were not gaming consoles. Intel executives are keen to point out that ARM’s lack of comprehensive application support leaves them having to play catch-up.
This also ignores a core point of comparison between the two chip architectures. ARM PCs are expected to achieve better optimization, resulting in battery life exceeding 15 or 20 hours. Of course, this doesn’t necessarily apply to every ultra-thin PC with a Copilot+ title, e.g. Lenovo Yoga Slim 7x or Microsoft Surface Pro. Intel needs to break 10 hours to be truly impressive, so we’ll have to wait and see.
“Not only does x86 deliver faster graphics and faster core power, but it’s also battery-powered,” Hallock said in the Q&A. “I’ll probably go to my grave saying, this is not that kind of chip; chip.” This is a choice you make inside the CPU.
