Intel Lunar Lake CPU analysis - The Core Ultra 7 258V's multi-core performance is disappointing, but its everyday efficiency is good

[Redaktion]

After the release of the Snapdragon X Elite chips as well as the AMD Zen 5 CPUs, Intel has finally launched its new Lunar Lake mobile processors. These are completely new chips with a focus on efficiency—plus, Intel has done without hyperthreading for the first time. But how efficient are the new models really and has too much performance perhaps been sacrificed?

https://www.notebookcheck.net/Intel-Lunar-Lake-CPU-analysis-The-Core-Ultra-7-258V-s-multi-core-performance-is-disappointing-but-its-everyday-efficiency-is-good.893405.0.html

[Palm1r]

Did Intel marketers send you this? Power consumption during everyday use. Because it sounds like nonsense. It looks like nonsense. They might as well have just downclocked the previous generation and gotten a similar result. That's all Intel has and that's enough of receiving gifts from them, they're the last ones.

[paviko]

What a disappointment, but also a big surprise. So the latest and greatest TSMC "3nm" is not much better than Intel "7nm" (Intel 4) and actually Intel "10nm"! Lunar Lake, a new architecture is only slightly better than Meteor Lake and 13th gen Core still keeps close.
The same was observed with Apple M3 3nm vs M2 5nm - not much improvement.

A mystery is why there is no more products with Intel 4 and Intel 3 - are so close to TSMC 3

[Palm1r]

Because it was all a lie, like the graphs at the Intel presentation.

[Joe]

Did Intel marketers send you this? Power consumption during everyday use. Because it sounds like nonsense. It looks like nonsense. They might as well have just downclocked the previous generation and gotten a similar result. That's all Intel has and that's enough of receiving gifts from them, they're the last ones.

Did AMD marketers send you here? :D
Intels previous gen was bad at 15W. Power consumtion of Lunar Lake includes memory. It looks good for >90% laptop users i think. I dont know how many of laptop users run Cinebench multi for everyday use, but i bet it is few ;).
I want to see how it performs gaming when plugged in power. Does it use 37W turbo whole time?

[Theo]

Did Intel marketers send you this? Power consumption during everyday use. Because it sounds like nonsense. It looks like nonsense. They might as well have just downclocked the previous generation and gotten a similar result. That's all Intel has and that's enough of receiving gifts from them, they're the last ones.

Did AMD marketers send you here? :D
Intels previous gen was bad at 15W. Power consumtion of Lunar Lake includes memory. It looks good for >90% laptop users i think. I dont know how many of laptop users run Cinebench multi for everyday use, but i bet it is few ;).
I want to see how it performs gaming when plugged in power. Does it use 37W turbo whole time?

Pretty sure the AMD and Qualcomm power consumption figures also include memory.

[bgx]

What a disappointment, but also a big surprise. So the latest and greatest TSMC "3nm" is not much better than Intel "7nm" (Intel 4) and actually Intel "10nm"! Lunar Lake, a new architecture is only slightly better than Meteor Lake and 13th gen Core still keeps close.
The same was observed with Apple M3 3nm vs M2 5nm - not much improvement.

A mystery is why there is no more products with Intel 4 and Intel 3 - are so close to TSMC 3

Wait for Arrowlake H if you want good multicore performance, and compare many core/threads vs many core/thread.
The comparison will be fairer. At least the 185H should be destroyed in perf/watt.

Lunar Lake is a 8Core - 8Thread machine.
vs 24 Threads machines (185H, Ryzen AI 12 cores x2), it looses in multicore benchmakrs.
The fact that its only 30% below with 3 times less threads is already pretty good.

It also looses the perf / W battle, because 24 threads downclocked also beats 8 threads full clocked CPU there.

May be Intel should have pushed lunar lake with 4 more E cores (the silicon price penalty would be limited, the everyday poer consumption should be limited too), and they would fair much better in such benchmarks, may be even being on par vs doulbe thread opponent.

In such a use case, which many coming here look at (is it better at benchmark XXX), Lunar lake is not for you.

I think the conclusion is pretty clear in that respect.

Now, lets wait next month for a Arrowlake in the mutliproc benchmark battle.

[billou]

16gb of memory its not 1W or 2 , perhaps 10W

[Evan]

I own both a Yoga Slim 7x and a Zenbook S16. Despite the synthetic benchmarks, the Slim 7x feels much snappier in everyday use (mostly web based professional productivity). Plus, it doesn't have all of the downsides of every x86 laptop I've ever used including the Zenbook S16. It turns on immediately, it doesn't get hot, it knows when the lid is closed, there aren't any lag spikes. The Slim 7x is the first Windows laptop I've ever used that feels like a smartphone in its operation. I felt the same way about my M1 MacBook Air.

What is it about X86 that, no matter how power efficient the processor, the laptops can never feel like a ARM product (instant on, snappy, etc)?

[Aras]

The table shows that the new Ryzen processors are faster than the Lunar Lake in all TDP values, so the conclusion should have been that the Lunar Lake improves over the previous Intel processors in efficiency at low TDPs, but the new Ryzen processors are significantly more efficient. Eventually the TDP of Ryzen can also be configured this low and the laptop will stay cool and silent yet providing a better performance.

The idle power consumption of Lunar Lake is lower than the new Ryzen (using an external monitor on Vivobook S 14 OLED and Zenbook S 14), but this is only important in very light tasks like video playback. So I don't clearly see a big advantage of Lunar Lake over Ryzen. A small advantage of the Lunar Lake might be that its GPU is a little faster than AMD in some cases, but I don't see it as a selling point. 

[paviko]

"Power consumption during everyday use" - this section is phenomenal. Should be placed in every notebook review. Please provide that info for every laptop - what was average power usage in Wats and how long it took to complete task. It is much better than benchmarks with full load for singlethread or multithread. It gives a real picture how long laptop can last. Or maybe simpler, how battery depleted from 100% after running PC Mark 10 test.

[Don't Fear the Future]

If I have one suggestion for the reviewer is to NOT use an external Monitor for figuring out efficiency.

Why? 

My laptop, for example, uses the discrete GPU any time I plug in an external monitor.

Not sure how this works for iGPU's, but it might also cause the iGPU to fire up a little hotter than it usually would; which then may mess up the efficiency tests.

Why not turn the resolution of the laptop monitory down to its lowest, the Refresh rate to its lowest, and turn the monitor brightness all the way down.  I'd bet that uses less power than outputting to an external monitor.  

[me2]

thank you for covering every day power usage. Lunar Lake looks like twice as efficient compared to AMD's best Intel really did an amazing job with Lunar Lake.

[Baloney]

multi-core performance is only relevant to you guys who doesn't care about real-world situation - and the truth is 99% of the people won't get benefit from strong multi-core performance but single-core performance on thin-and-light laptop. So stop writing like a fool man. Be realistic

[Gallo123]

In the real world the Intel is fast enough and close enough to competitors. AMD still has no answer on Thunderbolt, so it's a non-starter for productive work.

I am long ways away from buying a new laptop but I would gladly choose Intel again despite lagging in other areas.