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	12 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 (N3B)	Lunar Lake	Panther Lake
Platform	Mobile H/U Only	Desktop & Mobile H&HX	Mobile U Only	Mobile H
Process Node	Intel 4	TSMC N3B	TSMC N3B	Intel 18A
Date	Q4 2023	Desktop-Q4-2024 H&HX-Q1-2025	Q4 2024	Q1 2026 ?
Full Die	6P + 8P	8P + 16E	4P + 4E	4P + 8E
LLC	24 MB	36 MB ?	12 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.)

9949asd · Sep 19, 2024

Look at this, ARL r23 st 2500+. Form MLID lol

This is exactly equal to what I heard r23 p-core=443 /ghz. The e core is 363.

Wolverine2349 · Sep 19, 2024

9949asd said:
16 skymont almost equals to a 5950x.

Wouldn't they be better? I mean both have 16 cores and Skymont has around Golden or Raptor or Redwood Cove IPC which is around15 percent better than Zen 3.

Or are you referring to that 5950x has SMT where Skymont does not. So Skymont 16 cores no SMT almost equals 5950x with SMT. But turn off SMT on 5950X and the 16 Skymont cores spank it in workloads that scale to infinite threads?

9949asd · Sep 19, 2024

Wolverine2349 said:
Wouldn't they be better? I mean both have 16 cores and Skymont has around Golden or Raptor or Redwood Cove IPC which is around15 percent better than Zen 3.

Or are you referring to that 5950x has SMT where Skymont does not. So Skymont 16 cores no SMT almost equals 5950x with SMT. But turn off SMT on 5950X and the 16 Skymont cores spank it in workloads that scale to infinite threads?

Of course will be the 5950x 16c32t

cebri1 · Sep 20, 2024

jdubs03 said:
Figured I’d update the thread with a couple recent GB6 scores for the 288V. Looks like it’ll be about the same as the HX 370 (Zenbook S 16) in ST.
View attachment 107912
Pretty much confirms 2800-2880s ST, and right around 11000 MT.

These are great numbers. 20% faster ST than the 155H with a slightly lower power envelope. It should match the 155H MT with half the cores.

cannedlake240 · Sep 20, 2024

cebri1 said:
should match the 155H MT with half the cores.

GB6 favors that/doesn't scale with more threads

DrMrLordX · Sep 20, 2024

OneEng2 said:
Again, they do whatever they can to make the greatest profit. If it is illegal, they calculate the risk/reward of doing it anyway. Every company would be a monopoly if they could. Every one of them.

You expect people to find this position acceptable? That's the core of assessing whether a company is moral, immoral, or simply amoral.

DavidC1 · Sep 20, 2024

Wolverine2349 said:
Assuming in this screenshot it was running at 4.6GHz for the single core score, it has an IPC of about 4-6% lower than Golden Cove.

I don't buy this theory. Geekerwan tests show RPL mobile is slower than RPL-S by a similar amount MTL is "slower" compared to RPL-S.

Meaning there are clockspeed and IO related shenanigans in play which is why we wait until desktops are released and eliminate all doubt.

Wolverine2349 said:
That is still dang impressive. Not quite Raptor Cove IPC, but Chips and Cheese details and states its IPC is very close to Redwood Cove under Final Words: https://chipsandcheese.com/2024/06/15/intel-details-skymont/

C&C is likely implying that RWC is faster than Raptor Cove and Skymont is close to RWC.

Wolverine2349 said:
And Redwood COve has like 8-9% worse IPC than Raptor Cove per this:

That's way too big of a difference. It shows about 5-7% elsewhere. The MTL system Huang is using is underperforming.

cannedlake240 said:
GB6 favors that/doesn't scale with more threads

Geekbench scales reasonably well to a certain point. I don't know whether that applies at this thread count.

Also the fallacy of this argument is relying on a user-submitted benchmark to see if we can isolate single digit differences. It is absolutely futile.

DavidC1 · Sep 20, 2024

SiliconFly said:
DNR is GNR's 2025 successor. There were no credible leaks that suggest it was cancelled in favor of a new uarch.

Or, there's nothing P core in 2025, and it'll be in 2026 or later.

Even Clearwater Forest will likely be towards latter part of the year. Wouldn't surprise me too much if CWF is September-October.

My guess:
Q4 2025: 176 core Clearwater Forest -SP
Q1 2026: 288 core Clearwater Forest -AP
Q1-Q2 2026: P core Xeon 7 -SP
Q2-Q3 2026: P core Xeon 7 -AP

Since they aren't even talking about DMR it could be even later such as -SP coming in Fall of 2026 and -AP in Q4 2026. Considering AMD is in a near 2 year cycle, I can't see why they need to release it in early 2026.

cebri1 · Sep 20, 2024

DavidC1 said:
C&C is likely implying that RWC is faster than Raptor Cove and Skymont is close to RWC.

That's way too big of a difference. It shows about 5-7% elsewhere.

RWC is much better than most people think because launch microcode was horrible. Perf/mhz increased after several updates.

511 · Sep 20, 2024

DavidC1 said:
I don't buy this theory. Geekerwan tests show RPL mobile is slower than RPL-S by a similar amount MTL is "slower" compared to RPL-S.

Meaning there are clockspeed and IO related shenanigans in play which is why we wait until desktops are released and eliminate all doubt.

C&C is likely implying that RWC is faster than Raptor Cove and Skymont is close to RWC.

That's way too big of a difference. It shows about 5-7% elsewhere. The MTL system Huang is using is underperforming.

Huan tested in December without the performance BIOS

DavidC1 said:
Geekbench scales reasonably well to a certain point. I don't know whether that applies at this thread count.

Also the fallacy of this argument is relying on a user-submitted benchmark to see if we can isolate single digit differences. It is absolutely futile.

It scales only well for arm not x86 that's from my observation

jdubs03 · Sep 20, 2024

ARL-HX leak from Jaykihn. Everything is spicing up:
https://videocardz.com/newz/intel-c...ltra-9-285hx-with-24-cores-with-5-5-ghz-boost

csbin · Sep 20, 2024

ARL-S DDR5 8533 G2

https://twitter.com/x/status/1836953693314502719

mikk · Sep 20, 2024

DavidC1 said:
I don't buy this theory. Geekerwan tests show RPL mobile is slower than RPL-S by a similar amount MTL is "slower" compared to RPL-S.

Meaning there are clockspeed and IO related shenanigans in play which is why we wait until desktops are released and eliminate all doubt.

What about mobile versus mobile? Check out Geekbench scores, at the same clock speed MTL-H is a few percent slower than RPL-H. Constantly slower.

blackangus · Sep 20, 2024

Wolverine2349 said:
WHere is proof that AMD can?

Intel already has actually done it, just not on modern IPC consumer platform nor on a ring bus.

Intel has it on modern server and Workstation Xeon platforms being Saphhire Rapids and Emerald Rapids way more than 8 big cores on a singe tile or mesh which has consistent core to core latency unlike multiple CCDs beyond 8 cores. Yes mesh is not the best for gaming, though maybe mesh overclocked it can be good. But other things that need or desire consistent core to core latency beyond 8 cores. Plus no scheduling issues

AMD does not have more than 8 big cores on a single CCD/node even on the most enterprise EPYC and Threadripper unlike Intel with SPR and Emerald Rapids and beyond.

Well I mean beyond the fact they did it with Zen5c? Or that there is intent of 32 core CCD's with Zen6? Or maybe the graphics units they design that handle far more throughput than any CPU?
I mean the Zen5 CCD is dinky they are not silicon limited, so its obviously a design decision not to do this with the current P cores to keep margins as high as possible.
Why would you think they can't do it?
No one has made a 12000 hp street car yet, by that logic no-one can do it. Yet we know they can, its just not something that more than a handful of people would buy.

I will be excited as everyone to see if that 12 core intel is generally available as that is also something I would like to see if the performance is high end RPL+ like, Im in the market for a new PC and what to get something that will do well for at least 5 years so going with the top or near top of the line for whatever I choose. ARL cant get here fast enough for comparison!

AcrosTinus · Sep 20, 2024

Wolverine2349 said:
I have a spare 12600K in a backup system.

I just ran CPUZ 2.11 benchmark the default one with default settings just e-cores off in BIOS

Single core score was 791 and clock speed was 4.85GHz.

Ran it a couple other times and very close to same score of 790 to 793 with 4.85GHz clock speed single core.

So given that the Core Ultra 265K Skymont e-cores go up to 4.6GHz per its specs: https://www.techpowerup.com/cpu-specs/core-ultra-7-265k.c3776

Assuming in this screenshot it was running at 4.6GHz for the single core score, it has an IPC of about 4-6% lower than Golden Cove.

That is still dang impressive. Not quite Raptor Cove IPC, but Chips and Cheese details and states its IPC is very close to Redwood Cove under Final Words: https://chipsandcheese.com/2024/06/15/intel-details-skymont/

And Redwood COve has like 8-9% worse IPC than Raptor Cove per this:

Intel Meteor Lake P-cores Show IPC Regression Over Raptor Lake?

Intel Core Ultra "Meteor Lake" mobile processor may be the the company's most efficient, but isn't a generation ahead of the 13th Gen Core "Raptor Lake" mobile processors in terms of performance. This isn't just because it has an overall lower CPU core count in its H-segment of SKUs, but also...

www.techpowerup.com

And since Raptor Cove is 5% better than Golden Cove and Redwood Cove 8-9% behind Raptor and thus a few percent behind Goklden Cove, appears Skymont per my Golden Cove test of 12600K and SKymont screenshot accurate, appears Skymont has within 1-3% of Redwood Cove IPC aftercall.

Once again not quite as good as the +2% better than Raptor Cove that it promises, but still darn impressive and appears Chips and Cheese is right. Though it varies by workload a lot I imagine and maybe best case is 2% better than Raptor Cove with all around performance being Redwood Cove IPC.

I wonder what the clock speeds are running all 12 cores

The Ryzen 5900X also 12 cores scores 9471. It has SMT though. SO wonder how much uplift SMT gives it giving it a higher score than the 8983 12 Skymont scores in the screenshot leaked.

What do you think?

This REDWOOD COVE IS SLOWER THAN RAPTOR COVE nonsense needs to stop.
Chips and Cheese made that claim benchmarking a mobile part on a disaggregated first gen die. The real performance of REDWOOD COVE is the real 8Wide version in XEON6. Wait for the performance of that part before extrapolating mobile P-Core performance to a desktop. I wouldn't do the same for Lunar Lake either, the P-Core there with the entire fabric is tailor made for efficiency and not performance. Hence, the Redwood cove being slower than Raptor Cove claim is false for now until more information arrives.

Josh128 · Sep 20, 2024

blackangus said:
Well I mean beyond the fact they did it with Zen5c? Or that there is intent of 32 core CCD's with Zen6? Or maybe the graphics units they design that handle far more throughput than any CPU?
I mean the Zen5 CCD is dinky they are not silicon limited, so its obviously a design decision not to do this with the current P cores to keep margins as high as possible.
Why would you think they can't do it?
No one has made a 12000 hp street car yet, by that logic no-one can do it. Yet we know they can, its just not something that more than a handful of people would buy.

I will be excited as everyone to see if that 12 core intel is generally available as that is also something I would like to see if the performance is high end RPL+ like, Im in the market for a new PC and what to get something that will do well for at least 5 years so going with the top or near top of the line for whatever I choose. ARL cant get here fast enough for comparison!

They also did 16 core per CCD /8core per CCX on Zen 4C Bergamo. Rumor still has it that 3nm Zen 5C will be 16 core CCX/CCD. That could have major performance implications in some workloads. Diminishing returns for this Im sure, but when they went from 4 to 8 cores per CCX on Zen 3, they got the largest combination increase in clocks and IPC in the history of Zen.

Wolverine2349 · Sep 20, 2024

blackangus said:
Well I mean beyond the fact they did it with Zen5c? Or that there is intent of 32 core CCD's with Zen6? Or maybe the graphics units they design that handle far more throughput than any CPU?
I mean the Zen5 CCD is dinky they are not silicon limited, so its obviously a design decision not to do this with the current P cores to keep margins as high as possible.
Why would you think they can't do it?
No one has made a 12000 hp street car yet, by that logic no-one can do it. Yet we know they can, its just not something that more than a handful of people would buy.

I will be excited as everyone to see if that 12 core intel is generally available as that is also something I would like to see if the performance is high end RPL+ like, Im in the market for a new PC and what to get something that will do well for at least 5 years so going with the top or near top of the line for whatever I choose. ARL cant get here fast enough for comparison!

Well ZenC is dense and much weaker cores with gimped cache. Maybe they could do more than 8 cores on CCX. Probably though is it possible thermal and density constraints are in play for 12 higher performing cores on a single CCX rather than just hey do not only because of margins. Or both?

The C and even weaker C cores they have more for a reason. They are smaller clocked lower and have less cache so thus 16 C cores and 32 even weaker C variants on a single CCD/CCX/ And the 32 core variant is probably dual 16 CCX weak 16 cores on a single CCD so even crossing CCX bad latency penalty.

So your looking forward to the 12 core Bartlett Lake and see if it is generally available. I am too, but part of me is skeptical it will be stable or will it even be unlocked or have decent clocks. Or heck will it even be on a ring or just a shaved off Emerald Rapids Raptor Cove mesh die of 12 working P cores to fit into LGA 1700? And looking forward to Arrow Lake. Which would you choose. I like the all in one P core but at same time like newer arch of Arrow Lake especially with the strong e-cores finally. We have 1 month and 4 days to wait for Arrow Lake. Or maybe I will stick with 7800X3D. Its a tough call. Getting a little more than 16 threads especially with no SMT/HT is intriguing for gaming and thread heavy games with consistent core to core latency. But so is having 12 homogenous cores on one die.

Ghostsonplanets · Sep 20, 2024

csbin said:
ARL-S DDR5 8533 G2

https://twitter.com/x/status/1836953693314502719

ARL IMC to support CUDIMM DDR5 9600?

Wolverine2349 · Sep 20, 2024

Intel Arrow Lake-HX Enthusiast Laptop CPU Lineup Leak: Core Ultra 9 285HX Flagship With 24 Cores, Up To 5.5 GHz Clocks

Specs of Intel's next-gen Arrow Lake-HX enthusiast CPU family have leaked & will feature the Core Ultra 9 285HX flagship with up to 24 cores.

wccftech.com

Looks like high end laptop SKU will finally have equal number of P and e-cores as desktop SJ with much lower clocks.

When if ever has Intel ever had an equal core count in a laptop SKU as a desktop SKU. I cannot remember any in recent memory. I remember some were close, but none as many cores which is 8 + 16 in this case.

Oh edit that they do with the 13900HX Raptor Lake after research.

mikk · Sep 20, 2024

Wolverine2349 said:
Intel Arrow Lake-HX Enthusiast Laptop CPU Lineup Leak: Core Ultra 9 285HX Flagship With 24 Cores, Up To 5.5 GHz Clocks

Specs of Intel's next-gen Arrow Lake-HX enthusiast CPU family have leaked & will feature the Core Ultra 9 285HX flagship with up to 24 cores.

wccftech.com

Looks like high end laptop SKU will finally have equal number of P and e-cores as desktop SJ with much lower clocks.

When if ever has Intel ever had an equal core count in a laptop SKU as a desktop SKU. I cannot remember any in recent memory. I remember some were close, but none as many cores which is 8 + 16 in this case.

Oh edit that they do with the 13900HX Raptor Lake after research.

Last time i9-14900HX same core count. HX is desktop transferred to BGA, this is not a real mobile CPU.

ikjadoon · Sep 20, 2024

mikk said:
Yes PL2 can be configured for PL1. 155H has a TDP of 28W and many OEMs set PL1 to 45W or much higher than this. Lunar Lake has a relatively low PL2 even compared to MTL-U. We won't see crazy numbers which is good, I guess around 30W will be the peak PL1 for most higher end devices in max performance mode if the cooling can handle it. Cheaper devices with suboptimal cooling might prefer 17-20W.

I read this before and forgot about it until you & others mentioned it now.

Core Ultra 7 155H

Intel Ark
Processor Base Power: 28W
Minimum Assured Power: 20W
Maximum Turbo Boost Power: 115W

Alienware:
PL1: 102W
PL2: 121W

The PL1 is over 3X the "processor base power". At least one-third of all 155H laptops tested by Notebookcheck exceeded the base TDP on Ark.

Thunder 57 · Sep 20, 2024

Wolverine2349 said:
The C and even weaker C cores they have more for a reason. They are smaller clocked lower and have less cache so thus 16 C cores and 32 even weaker C variants on a single CCD/CCX/ And the 32 core variant is probably dual 16 CCX weak 16 cores on a single CCD so even crossing CCX bad latency penalty.

What are you talking about? C core and weaker C core? 16 C cores and 32 even weaker C variants?

Wolverine2349 · Sep 20, 2024

Thunder 57 said:
What are you talking about? C core and weaker C core? 12 C cores and 32 even weaker C variants?

Th Zen 6 32 core variant is the even weaker C cores. I remember reading it somewhere on how they would get 32 onto one CCD with upcoming Zen 6. But maybe things change do not know. But if Zen 6 is coming 2025, hard to imagine they could get 32 into one CCD without making them even weaker C cores/

fastandfurious6 · Sep 20, 2024

zen 5c is actually very strong, 70-80% perf of full zen 5 (someone write the right number)

zen 6c will be similar

zen 6 with 32 cores probably will happen

cebri1 · Sep 20, 2024

I love these Zen 5 talks in the Intel Lake thread.

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

Senior member

Attachments

Member

Senior member

Member

Senior member

Senior member

Lifer

Golden Member

Golden Member

Senior member

Golden Member

Golden Member

Senior member

Diamond Member

Member

Member

Senior member

Senior member

Senior member

Senior member

Diamond Member

Senior member

Diamond Member

Senior member

Senior member

Senior member