Discussion Intel Meteor, Arrow, Lunar & Panther Lakes Discussion Threads

Tigerick · Aug 22, 2022

As Hot Chips 34 starting this week, Intel will unveil technical information of upcoming Meteor Lake (MTL) and Arrow Lake (ARL), new generation platform after Raptor Lake. Both MTL and ARL represent new direction which Intel will move to multiple chiplets and combine as one SoC platform.

MTL also represents new compute tile that based on Intel 4 process which is based on EUV lithography, a first from Intel. Intel expects to ship MTL mobile SoC in 2023.

ARL will come after MTL so Intel should be shipping it in 2024, that is what Intel roadmap is telling us. ARL compute tile will be manufactured by Intel 20A process, a first from Intel to use GAA transistors called RibbonFET.

Comparison of upcoming Intel's U-series CPU: Core Ultra 100U, Lunar Lake and Panther Lake

Model	Code-Name	Date	TDP	Node	Tiles	Main Tile	CPU	LP E-Core	LLC	GPU	Xe-cores
Core Ultra 100U	Meteor Lake	Q4 2023	15 - 57 W	Intel 4 + N5 + N6	4	tCPU	2P + 8E	2	12 MB	Intel Graphics	4
?	Lunar Lake	Q4 2024	17 - 30 W	N3B + N6	2	CPU + GPU & IMC	4P + 4E	0	8 MB	Arc	8
?	Panther Lake	Q1 2026 ?	?	Intel 18A + N3E	3	CPU + MC	4P + 8E	4	?	Arc	12

Comparison of die size of Each Tile of Meteor Lake, Arrow Lake, Lunar Lake and Panther Lake

	Meteor Lake	Arrow Lake (20A)	Arrow Lake (N3B)	Arrow Lake Refresh (N3B)	Lunar Lake	Panther Lake
Platform	Mobile H/U Only	Desktop Only	Desktop & Mobile H&HX	Desktop Only	Mobile U Only	Mobile H
Process Node	Intel 4	Intel 20A	TSMC N3B	TSMC N3B	TSMC N3B	Intel 18A
Date	Q4 2023	Q1 2025 ?	Desktop-Q4-2024 H&HX-Q1-2025	Q4 2025 ?	Q4 2024	Q1 2026 ?
Full Die	6P + 8P	6P + 8E ?	8P + 16E	8P + 32E	4P + 4E	4P + 8E
LLC	24 MB	24 MB ?	36 MB ?	?	8 MB	?
tCPU	66.48
tGPU	44.45
SoC	96.77
IOE	44.45
Total	252.15

Intel Core Ultra 100 - Meteor Lake

As mentioned by Tomshardware, TSMC will manufacture the I/O, SoC, and GPU tiles. That means Intel will manufacture only the CPU and Foveros tiles. (Notably, Intel calls the I/O tile an 'I/O Expander,' hence the IOE moniker.)

ikjadoon · Dec 29, 2023

controlflow said:
"Battery capacity has increased from 80 Wh on the Prestige 15 to 99 Wh on our Prestige 16. Combine this with the efficiency of the new 14th gen Intel CPU and you can get extremely long runtimes of almost 18.5 hours of real-world WLAN browsing. This is one of the longest-lasting Windows laptops we've tested with results that rival the MacBook series."
View attachment 91120

Impressive battery life here. Would be curious to see how a PHX laptop with a similar size battery performs.
Not a fan of how MSI decided to set the PL2 to 115W in this laptop. Hopefully there is at least an easy option to cap power much lower. Asus in the Zenbook seems to allow you to cap the power to 28W, which is much more appropriate for a thin and light laptop.

As you noted, this is running a pretty old bios. Would be interesting to see if things can improve with newer FW.

That is some bonkers battery life on an H-series CPU. Love the 100WHr battery. Li-on batteries are so light relative to a laptop, everyone should be going big. Now let's see some MTL-U battery life next year: that is what most consumers will buy, as Intel finally got rid of the confusingly popular P-series.

For reference, here is the original M1 MacBook Air:

The M2 Air was actually a bit worse at 910 minutes in the WiFi v1.3 test.

//

This Prestige 16 now has the 10th longest battery life ever tested in a laptop review by Notebookcheck. See the M3 MacBook Pro 16: just +9% longer with the same battery size on a very common consumer task like web browsing.

MTL appears to be a massive battery life win. Let's hope all OEMs continue to use 80+ WHr laptops in all their 13" & larger designs.

mikk · Dec 29, 2023

Khato said:
Back on MTL, another video has come along comparing the Asus new and old BIOS:

The benchmarks themselves have the usual problem of not being run at a static power level, so I wouldn't say they're of much interest. What is nice is starting around 3 minutes in are graphs of average temperature, frequency, and power versus time on a Prime95 run with both new and old BIOS. While the power graph clearly indicates that the new BIOS does allow the CPU to consume more power for a time, by the end of the graph the power consumption is equivalent between new and old BIOS... but the clock speed with new BIOS at that steady state is about 10% higher than the old BIOS. Also the new BIOS shows a markedly more consistent clock speed in general.

It uses quite a bit more power indeed, it might not even related to the pcode update golden pig was talking about. Hard to say with this power difference. On the other side the CPU frequency in Prime95 isn't that much different after PL2 drops, maybe like 15% higher (1600? Mhz vs 1400? Mhz) whereas Cinebench R24 MT run 26% faster and compiling even more faster. We don't know for the E-cores though. Depends on when Intel released the new pcode for OEMs, it's not like OEMs will immediately update and release everything new from Intel. I think it is more than just a power bump up but there is limited data in this test. Something clearly was wrong with the old bios in the compiling test, maybe the E-cores didn't run properly.

Hulk · Dec 29, 2023

mikk said:
It uses quite a bit more power indeed, it might not even related to the pcode update golden pig was talking about. Hard to say with this power difference. On the other side the CPU frequency in Prime95 isn't that much different after PL2 drops, maybe like 15% higher (1600? Mhz vs 1400? Mhz) whereas Cinebench R24 MT run 26% faster and compiling even more faster. We don't know for the E-cores though. Depends on when Intel released the new pcode for OEMs, it's not like OEMs will immediately update and release everything new from Intel. I think it is more than just a power bump up but there is limited data in this test. Something clearly was wrong with the old bios in the compiling test, maybe the E-cores didn't run properly.

Yes, looks like Intel pushed up max temp and power so performance benches better.

For what it's worth I was at my local Microcenter today and got to check out that Zenbook with MTL. Wow, gorgeous screen and really nice form factor I must say.

Khato · Dec 29, 2023

mikk said:
It uses quite a bit more power indeed, it might not even related to the pcode update golden pig was talking about. Hard to say with this power difference. On the other side the CPU frequency in Prime95 isn't that much different after PL2 drops, maybe like 15% higher (1600? Mhz vs 1400? Mhz) whereas Cinebench R24 MT run 26% faster and compiling even more faster. We don't know for the E-cores though. Depends on when Intel released the new pcode for OEMs, it's not like OEMs will immediately update and release everything new from Intel. I think it is more than just a power bump up but there is limited data in this test. Something clearly was wrong with the old bios in the compiling test, maybe the E-cores didn't run properly.

At the end of the Prime95 frequency trace it looks to be reporting about 1.3GHz on the old BIOS and 1.425GHz on the new BIOS. During that time the reported power is basically equivalent. Note that I'm assuming that the reported frequency is indicative of actual relative frequencies.

The frequency plot on the old BIOS is definitely indicative of there being something wrong. It seems like it's overcompensating for changes in loading maybe? Resulting in more drastic swings in frequency compared to the new BIOS.

controlflow · Dec 29, 2023

Hulk said:
Yes, looks like Intel pushed up max temp and power so performance benches better.

For what it's worth I was at my local Microcenter today and got to check out that Zenbook with MTL. Wow, gorgeous screen and really nice form factor I must say.

I played around with one too, really liked the 120hz OLED. Nice colors, contrast and a fast/responsive display. Too bad those screens are so power hungry. BTW, I can confirm that the retail units really do have LPDDR5 7467 (32GB for the unit I looked at). The Best Buy specs claiming 6400 is indeed a typo.

I might consider picking one of these up based on how the power/performance ends up looking after a few more bios updates. The build quality and packaging are nice and it seems to be quite capable for a sub 3lb system. $1300 is a pretty reasonable price for the specs, but maybe I've been too desensitized by Macbook Pro pricing.

Hulk · Dec 29, 2023

controlflow said:
I played around with one too, really liked the 120hz OLED. Nice colors, contrast and a fast/responsive display. Too bad those screens are so power hungry. BTW, I can confirm that the retail units really do have LPDDR5 7467 (32GB for the unit I looked at). The Best Buy specs claiming 6400 is indeed a typo.

I might consider picking one of these up based on how the power/performance ends up looking after a few more bios updates. The build quality and packaging are nice and it seems to be quite capable for a sub 3lb system. $1300 is a pretty reasonable price for the specs, but maybe I've been too desensitized by Macbook Pro pricing.

Not to derail the thread but do you know if the SSD in the Zenbook is replaceable? I'd really like a 2TB (or more) drive.

controlflow · Dec 30, 2023

Hulk said:
Not to derail the thread but do you know if the SSD in the Zenbook is replaceable? I'd really like a 2TB (or more) drive.

it is indeed replaceable according to Ultrabookreview.com. I think it is actually the only upgradable thing on the entire system at this point.

TESKATLIPOKA · Dec 30, 2023

ikjadoon said:
MTL appears to be a massive battery life win.

MTL is a massive battery life win? It clearly depends on laptop, Acer Swift Go 14 SFG14-72 ended up much worse in comparison despite having the same CPU.
On the other hand, MSI did a great job in selecting power efficient components and putting the biggest battery possible inside.

BTW, this was in what you posted.

Raptor Lake U managed 954 minutes with only a 68W battery, If I normalized It to 99.9W then It should manage 1401 minutes.
Not really apples to apples comparison for both CPU and laptop, but I wanted to show that It's not really depended on CPU, because they are very efficient light loads, but more on the rest of components(SSD, RAM, display).

TESKATLIPOKA · Dec 30, 2023

Hulk said:
Not to derail the thread but do you know if the SSD in the Zenbook is replaceable? I'd really like a 2TB (or more) drive.

Is there a laptop in which SSD is soldered to the board? Isn't It always put into an M2 slot?

cebri1 · Dec 30, 2023

Hulk said:
Yes, looks like Intel pushed up max temp and power so performance benches better.

For what it's worth I was at my local Microcenter today and got to check out that Zenbook with MTL. Wow, gorgeous screen and really nice form factor I must say.

I saw one as well, it's indeed gorgeous. I may pull the trigger on it, as I'm looking to replace my old 7700 + 1050Ti. Only downside is 16GB of RAM.

mikegg · Dec 30, 2023

ikjadoon said:
That is some bonkers battery life on an H-series CPU. Love the 100WHr battery. Li-on batteries are so light relative to a laptop, everyone should be going big. Now let's see some MTL-U battery life next year: that is what most consumers will buy, as Intel finally got rid of the confusingly popular P-series.

For reference, here is the original M1 MacBook Air:

The M2 Air was actually a bit worse at 910 minutes in the WiFi v1.3 test.

//

This Prestige 16 now has the 10th longest battery life ever tested in a laptop review by Notebookcheck. See the M3 MacBook Pro 16: just +9% longer with the same battery size on a very common consumer task like web browsing.

MTL appears to be a massive battery life win. Let's hope all OEMs continue to use 80+ WHr laptops in all their 13" & larger designs.

Do these battery benchmarks account for browsing speed? For example, number of web pages loaded/minute.

I can see CPUs running at extremely low frequency in order to increase battery life. AMD and Intel mobile CPUs are notorious for running at half the speed when unplugged from the wall while Macbooks run at full speed until low battery.

DavidC1 · Dec 30, 2023

SiliconFly said:
RWC is exactly same as previous gen, clock frequencies too are similar to previous gen & no known significant performance optimizations either. They played it too safe. So, we can't expect much performance gains with RWC at this point I guess.

Redwood Cove on client gets doubled L1i. Probably helps in some scenarios and also lower power a bit.

DavidC1 · Dec 30, 2023

TESKATLIPOKA said:
Raptor Lake U managed 954 minutes with only a 68W battery, If I normalized It to 99.9W then It should manage 1401 minutes.
Not really apples to apples comparison for both CPU and laptop, but I wanted to show that It's not really depended on CPU, because they are very efficient light loads, but more on the rest of components(SSD, RAM, display).

The U's are more power efficient but average is what matters. Alder/Raptor seems to have a bigger than usual variation. Let's see what other MTL laptops, especially based on Us do. I'm interested on U9.

Tigerick · Dec 30, 2023

DavidC1 said:
The U's are more power efficient but average is what matters. Alder/Raptor seems to have a bigger than usual variation. Let's see what other MTL laptops, especially based on Us do. I'm interested on U9.

U9 has 2P+8E with 4 Xe Intel graphics, there is nothing special on it. You should wait for a year for Lunar Lake

DavidC1 · Dec 30, 2023

Tigerick said:
U9 has 2P+8E with 4 Xe Intel graphics, there is nothing special on it. You should wait for a year for Lunar Lake

If you are looking for portability, the chances are you aren't a gamer, or at least don't slobber over graphics. At that power level, you won't run much of latest fancy pants graphics anyway. Lunar isn't going to change this.

By the way, no surprise at Meteorlake launching in December. All I see is bunch of excuses or long-winded explanations, but mobile always comes out to be 1 year release cadence. If a next gen product is coming 6 months down the road, why would manufacturers bother, especially considering it takes couple of months for BIOS/software optimizations.

mikk · Dec 30, 2023

Khato said:
At the end of the Prime95 frequency trace it looks to be reporting about 1.3GHz on the old BIOS and 1.425GHz on the new BIOS. During that time the reported power is basically equivalent. Note that I'm assuming that the reported frequency is indicative of actual relative frequencies.

The frequency plot on the old BIOS is definitely indicative of there being something wrong. It seems like it's overcompensating for changes in loading maybe? Resulting in more drastic swings in frequency compared to the new BIOS.

Sadly he used Prime95 instead of Cinebench R24 or his compile test to check out clock speed, power and temperature, it would have been much more meaningful for a real world use case.

FlameTail · Dec 30, 2023

mikk said:
Cinebench R24

AHAHAHHAHAHAHAAH

SiliconFly · Dec 30, 2023

FlameTail said:
that's nuts.

ondma said:
IDK, I am not a computer engineer, but I read an article comparing rentable units and hyperthreading, and I didnt see them come to that conclusion.
The surprising thing about ARL is that if it truly does not have hyperthreading, the leaks I saw didnt say anything about increasing E cores either.

Seems like the worst of both worlds. Loss of single thread performance due to clock regression, and loss of ultimate MT performance due to no HT.
My feeling is that Zen 5 will dominate in both.

Saylick said:
Pretty sure you cannot scale ST performance by X amount just by scaling up a theoretical core's resources by the same increase. A single thread will never fully saturate the width of a core at all times because of instruction dependencies, which is exactly why going wider doesn't give you proportional IPC uplift. It's also why SMT was created, so that you get more throughput of a given core, but not more ST performance. Rentable Units, if it's actually possible, likely means better utilization of silicon area since you don't need a separate big cores, which are not efficient from a perf/mm2 point of view since ST performance has diminishing returns with core area.

Khato said:
Regarding rentable units... Sadly the reality is quite boring, especially compared to the fanciful fiction. I bet that the term was included without context in some presentation that a non-technical 'leaker' received. So clearly some explanation for the term needed to be created in order to be able to 'leak' it.

Sorry abt the confusion. I was talking about the theoretical maximum of a hypothetical 8 core CPU with a perfect Rentable Units (RU) implementation running a perfectly sliceable workload. Just an imaginary CPU (not Intel or any other company for the matter). Don't think a real CPU like that can even exist in the real world. It was all just about the idea itself.

Actually, I think MLID's leak about RU in some future Intel architecture is (I'm assuming) maybe about just 2 cores tied together into a single cluster and sharing resources that can probably boost slightly beyond a single core's ST performance when a single thread is assigned to that cluster. But even that sounds a bit dodgy imho.

FlameTail · Dec 30, 2023

SiliconFly said:
Actually, I think MLID's leak about RU in some future Intel architecture is (I'm assuming) maybe about just 2 cores tied together into a single cluster and sharing resources that can probably boost slightly beyond a single core's ST performance when a single thread is assigned to that cluster. But even that sounds a bit dodgy imho

Smells like a resurrected Bulldozer?

mikk · Dec 30, 2023

Cinebench scores from a Thinkpad Carbon: https://www.bilibili.com/video/BV1PQ4y1J7K9/?spm_id_from=333.337.search-card.all.click

Sustained power around 30W and PL2 55W (6:45).

SiliconFly · Dec 30, 2023

FlameTail said:
Smells like a resurrected Bulldozer?

Unlike RU, Bulldozer's ST perf is limited to a single core's max perf. In theory, RU's ST perf can actually exceed a single core's ST perf depending on the thread. Whether this works in real world is a different question altogether. Sounds a bit unlikely as implementation is a nightmare.

coercitiv · Dec 30, 2023

SiliconFly said:
Sorry abt the confusion. I was talking about the theoretical maximum of a hypothetical 8 core CPU with a perfect Rentable Units (RU) implementation running a perfectly sliceable workload.

A perfectly sliceable workload would be sliced into the optimum amount of threads already. It would also reduce RU efficiency to zero as all cores would already have work to do.

SiliconFly · Dec 30, 2023

SiliconFly said:
...when a single thread is assigned to that cluster...

coercitiv said:
A perfectly sliceable workload would be sliced into the optimum amount of threads already. It would also reduce RU efficiency to zero as all cores would already have work to do.

Not multi-threaded workload. I mean a "single thread" can be "sliced" by a processor only if it has RU.

(Very different from scheduling workloads with multiple threads which is already common.)

Abwx · Dec 30, 2023

mikk said:
Cinebench scores from a Thinkpad Carbon: https://www.bilibili.com/video/BV1PQ4y1J7K9/?spm_id_from=333.337.search-card.all.click

Sustained power around 30W and PL2 55W (6:45).

From 9600 pts to 14952 pts is 1.56x better perf and should require 2.5-3x the power, and here power increase by only 1.3x, that would be perf/watt increasing hugely when power increase, that s impossible, there s a a turbo at work here.

Edit : Assuming that 9600 pts at 23W is accurate then 14952 pts is at 55W.

coercitiv · Dec 30, 2023

SiliconFly said:
Not multi-threaded workload. I mean a "single thread" can be "sliced" by a processor only if it has RU.

(Very different from scheduling workloads with multiple threads which is already common.)

A single thread that can be perfectly sliced by RU can be sliced perfectly by the programmer or the compiler as well. You asked for the ideal situation, that's what happens when things are pushed to the limit.

Discussion Intel Meteor, Arrow, Lunar & Panther Lakes Discussion Threads

Senior member

Attachments

Member

Diamond Member

Diamond Member

Golden Member

Member

Diamond Member

Member

Platinum Member

Platinum Member

Member

Golden Member

Senior member

Senior member

Senior member

Senior member

Diamond Member

Platinum Member

Golden Member

Platinum Member

Diamond Member

Attachments

Golden Member

Diamond Member

Golden Member

Lifer

Diamond Member