Intel's Lunar Lake CPUs could deliver 50% faster multi-threaded performance than Meteor...

DragonSlayer101

Posts: 372   +2
Staff
Rumor mill: Intel's upcoming Lunar Lake processors could offer a massive multithreaded performance gain over Meteor Lake, according to a new rumor. Chipzilla is expected to launch Lunar Lake later this year as the third lineup under its Core Ultra brand, with the U-series chips meant for ultra-portable laptops and convertibles.

The latest report comes from tipster extraordinaire @SquashBionic, who claims that the 17W Lunar Lake U-series chips will offer a 50 percent improvement in multi-threaded performance over the current-gen 15W Meteor Lake processors. The next-gen chips were earlier rumored to offer higher performance per watt and faster AI NPU performance, and the latest report now suggests that those claims were fairly accurate.

In another tweet, the tipster wrote that the 17W TDP was the original PL1 spec for single-fan configurations, but Intel has decided to extend the cTDP to 30W due to popular demand from OEMs. There's no word on specific SKUs, so we don't know how many of them will still ship with the original 17W TDP and how many will get the enhanced 30W spec for improved performance. However, if the 17W SKUs can offer a 50 percent higher performance than Meteor Lake, it is anybody's guess as to how much faster the 30W chips will be.

According to prior rumors, the Lunar Lake processors will feature Lion Cove Performance Cores, Skymont Efficient Cores, and a fast Neural Processing Engine (NPU) that is said to offer thrice the performance of the NPU found in the Meteor Lake chips. The next-gen processors are also expected to get a brand new iGPU based on the Battlemage Xe2-LPG architecture.

For all its improved specs, Lunar Lake will reportedly ditch hyper-threading support on its P-cores, which could theoretically affect their multi-threaded performance. If the rumors hold up, the 17W Lunar Lake CPUs will ship with up to 4 P-Cores and 4 E-Cores, totaling up to 8 cores and 8 threads, while the 15W Meteor Lake chips cited in the post offer 2 P-cores, 8 E-Cores, and 2 LP-E cores for a total of 12 cores and 14 threads.

Permalink to story.

 
This article references "Chipzilla"...in 2024 wouldnt that be Nvidia? I mean, even AMD is bigger (in market cap) than Intel in 2024, so maybe we should update our references? Intel is the underdog in the 2020s.
 
So 2x more p cores means more MT performance ? Got it. But assuming it can do it on only 17W that will be great for mini PCs and handhelds!
 
And this is how Chimpzilla is going to lose sells this year.
Greenzilla is doing fine even with that hogenergy that's 4090
 
Those who own 17W laptops. Seriously, what sort of question is this? If you have an ultralight laptop, you're shouldn't be interested in how fast it runs benchmarks?
When you see the price of Luna Lake you will demanding it runs what ever you want very fast. This is being launched as an Ultrapremium Apple competitor in the thin and light segment.
 
Those who own 17W laptops. Seriously, what sort of question is this? If you have an ultralight laptop, you're shouldn't be interested in how fast it runs benchmarks?

How many run Cinebench on handheld gaming device like Steam Deck? 0.001% users? How many run Cinebench on tablet? 0.01% users? How many run Cinebench on ultraportable? Cinebench is one of last software that interest ultraportable users.

And as proven many times, benchmarks do NOT say almost anything about usability on real world scenarios. For your question, ultraportable users should not be interested on benchmarks since usage is very far from what benchmarks tell.
 
Intel will be on fire when Arrow Lake and Lunar Lake launches later this year

https://www.linkedin.com/feed/update/urn:li:activity:7173194948257808385/

Intel sits at 78% marketshare already, more than 3 times AMD and Apple combined.

AMD should release Zen 5 ASAP to disrupt them. However both will speak at Computex. Q3 or Q4 release for both.

Lets hope AMD did not cheap out and went for 4/5nm TSMC for Zen 5, giving Intel node advantage for the first time in a few years
 
Last edited:
How many run Cinebench on handheld gaming device like Steam Deck? 0.001% users? How many run Cinebench on tablet? 0.01% users? How many run Cinebench on ultraportable? Cinebench is one of last software that interest ultraportable users.
Do you know what a benchmark is, or why a person may be interested in its results? Of course, many users -- whether on a tablet, laptop, or workstation may never run their own benchmarks -- but they're still interested in how their hardware compared to others.

And as proven many times, benchmarks do NOT say almost anything about usability on real world scenarios.
You got us. Every benchmark program developed in the last half-century is utterly useless. They've been fooling us all these years.
 
Do you know what a benchmark is, or why a person may be interested in its results? Of course, many users -- whether on a tablet, laptop, or workstation may never run their own benchmarks -- but they're still interested in how their hardware compared to others.


You got us. Every benchmark program developed in the last half-century is utterly useless. They've been fooling us all these years.

Benchmarks are useless if they don't reflect performance difference on real world usage scenario. This is about that. So yeah, you can compare benchmark results on scenarios that never exist IRL, but not much worth there...

Good that finally someone else understood this too. Benchmarks are mostly useless.
 
Intel will be on fire when Arrow Lake and Lunar Lake launches later this year

https://www.linkedin.com/feed/update/urn:li:activity:7173194948257808385/

Intel sits at 78% marketshare already, more than 3 times AMD and Apple combined.

AMD should release Zen 5 ASAP to disrupt them. However both will speak at Computex. Q3 or Q4 release for both.

Lets hope AMD did not cheap out and went for 4/5nm TSMC for Zen 5, giving Intel node advantage for the first time in a few years
I rather wait and see before concluding. Comparing this with the Meteor Lake U series chips isn’t great when it’s only got 2 miserable P-cores and a bunch of weak E-cores. So while Lunar Lake lost HT, it’s got 4 physical P-cores vs 2 on the Meteor Lake with HT. So the lost of HT may not result in performance regression in this scenario.

About market share, I believe Intel dominates the PC market decades back. However, it’s apparent they have lost quite a fair bit of market share to AMD and ARM based chips. In my opinion, it’s not AMD that’s their main competitor, but the likes of Qualcomm and ARM based chips that’s going to give Intel a major headache. No matter how good Lunar and Arrow Lake are, it’s very difficult to compete with a good ARM based SOC when it comes to performance per watt. Apple’s M series SOC is prove that ARM chips can be a very good alternative to x86 chips.
 
No matter how good Lunar and Arrow Lake are, it’s very difficult to compete with a good ARM based SOC when it comes to performance per watt. Apple’s M series SOC is prove that ARM chips can be a very good alternative to x86 chips.
How many x86-based SOCs are out there? And I'm talking about ones that have integrated memory on SOC.

Yeah, when x86 SOCs with integrated memory get out, Apple SOC's only real advantage is gone. That is something everyone seems to forget.
 
Back