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

[Tigerick]

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.





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.)

 

[majord]

Why am I feeling extreme Deja vu..

 

[Hulk]

Another link on the hot hardware podcast info:

Intel Identifies Issues with the Launch of Its Core Ultra 200S Desktop Arrow Lake Desktop CPUs and Promises Fixes Are Forthcoming

There's good news on the horizon for the Core Ultra 200 series Arrow Lake processors as Intel identifies performance-hindering issues.

www.thefpsreview.com

Transcript:
“I think what people have been interested to hear is what happened. I can’t go into all the details yet, but we identified a series of multifactor issues at the OS level, at the BIOS level, and I will say that the performance we saw in reviews (through no fault of reviewers) is not what we expected and not what we intended. The launch just didn’t go as planned. That has been a humbling less for all of us, inspiring a fairly large response internally to get to the bottom of what happened and to fix it.”

Over an hour interview and a couple simple questions were not asked?

1. You mentioned the testing by reviewers was not what Intel expected. I'm sorry but I'm not following? Were the tests results from the reviewers different from what Intel's testing showed?

2. Were the internal Intel test rigs somehow different than released hardware and software? Was Intel testing with special versions of Windows and BIOS? What is the reason for the "unexpected" results? Did Intel not release what they had tested in the press release info?

3. Again, it is strange that you used the term "expected" for the benchmarking results of reviewers. Do you mean they were lower than your internal testing, if so why not just use the word "lower" instead of "expected?"

4. Finally, will the results after your BIOS and Windows fixes be equal to your original press release performance data or better? If equal, then why are reviewer benchmarks lower than Intels? Again, back to question #2, was Intel using different software/BIOS?

That interview came of as a lot of spin and borderline outright deceiful. We have a good idea of what is going on but we can't tell you absolutely anything. We can even tell you if reviewer bench data was lower or higher than our internal testing only not what we expected. This is a load of BS gobblygook.

As usual Intel is too high and mighty to actually answer questions from the little people in an honest and straightforward because our little brains might explode.

I was a pretty big fan of Intel. I'm getting sick of the spin. The word "expected" really infuriates me as it is so mealy-mouthed and vauge.

"Oh we're so sorry and we're absolutely going to fix this!"

Fix what? You tested this hardware on presumably the same software the reviewers did? None of it makes sense, or perhaps I'm just stupid, which is entirely possible.

 

[511]

Skymont grew a lot because of the doubled FP section, which is likely responsible for 20-30% growth in die area alone. Without that, it would have been well under 1mm2, and be close to 1:4 ratio again.

Intel's numbers of 32% Int and 72% FP means the gain due to doubled FP is approximately 1.72/1.32 or 30%. 20-30% die area growth for 30% FP improvement.

The 30% integer gain came at the cost of roughly 30% core area increase when you take out the doubled FP, assuming ~1.6x N3B density gain over Intel 4, which is an excellent figure. It might be even more optimistic than this, depending on the actual difference between N3B and Intel 4.

For skymont we still don't know the fin Library which will complicate this and it would be using 3-2 like Apple is using to reach high clocks

 

[coercitiv]

Fix what? You tested this hardware on presumably the same software the reviewers did? None of it makes sense, or perhaps I'm just stupid, which is entirely possible.

We've had a similar case with AMD this year: engineers at were testing CPUs in a way that was also disabling some security patches in Windows, which led to better performance. Those security patches were bound to be disabled in a new build of Win, so there was indeed some performance left on the table in early reviews. Normally I would criticize the vendor for this type of screw-up, but the reality is Intel is paying through their nose for this "unexpected" performance gap. The loss in sales and in mind share is telling them all they need to know.

IMHO we should cut Robert Hallock some slack, he's doing something Intel failed to do for years now, which is to talk about internal issues and communicate intention to fix things. He's also been dealt a very bad hand to work with here. The same applied for Tom Petersen on the GPU side, it's hard to get out there and expose yourself, so they deserve more leniency than usual (as messengers).

Let's wait and see what Intel cooks up, it's not like anyone is rushing to buy Arrow Lake anyway.

 

[511]

No wonder intel is having a hard time with drivers and OS the actual engineers left

https://twitter.com/x/status/1855388673598271797

 

[Slash3]

No wonder intel is having a hard time with drivers and OS the actual engineers left

https://twitter.com/x/status/1855388673598271797

View attachment 111383

Easter egg; it's a tradition that goes back a very long time.

 

[511]

Was this posted in this thread?

https://twitter.com/x/status/1778120391657136573

 

[SiliconFly]

.

 

[511]

“Exclusive:-#Panther Lake is 35% more Efficient than #Lunar Lake.”

Looks like it’s totally made up. He posts way too much unverifiable/cooked-up stuff & lots of garbage. Wouldn’t trust him unless the news comes from a reliable source.

Yes i know he along with couple of Intel heros are total shill for intel they might run userbencmark who knows lol i am waiting for jaykihn or Bionic or Raichu for the leaks

 

[mikk]

35% is easy for PTL-H under load, it has much more cores. Doesn't tell much without more infos. Xe3 40% at what power. If it's at 50W on PTL-H this isn't impressive. At 30W however it would be quite good.

 

[maddie]

Hulk's getting angry, look out folks.

As usual Intel is too high and mighty to actually answer questions from the little people in an honest and straightforward because our little brains might explode.

I was a pretty big fan of Intel. I'm getting sick of the spin. The word "expected" really infuriates me as it is so mealy-mouthed and vauge.

"Oh we're so sorry and we're absolutely going to fix this!"

Fix what? You tested this hardware on presumably the same software the reviewers did? None of it makes sense, or perhaps I'm just stupid, which is entirely possible.

 

[511]

Hulk's getting angry, look out folks.

Hope he doesn't smash 💀
On a sidenot when is the expected time for RMA to be resolved?

 

[SiliconFly]

.

 

[FlameTail]

+35% efficiency isn't easy. In fact it close to impossible unless they pull out some major new gimmicks.

From N3B to 18A, the ppw gain (power, not density) isn't that significant I think. It's like 10% or maybe even upto 20% due to BSPDN. 35%? Thats just way off. In fact, I'm getting the feeling that it's gonna be hard for PTL-H to even match LNL efficiency (let alone +35%).

Edit: If PTL-H can comfortably surpass LNL in efficiency, it places Intel in a whole new level. A direct competitor to Apple leaving AMD in the dust. So, this +35% efficiency gain rumor should be taken with a tiny truck load of salt.

What efficiency?

If it's +35% multicore efficiency, it should be easy for PTL-H. The performance-power curve of PTL-H will be better than LNL, by virtue of it having more cores.

 

[igor_kavinski]

His anger is justified. Another 3 to 4 weeks wait is kinda painful.

I think he should RMA his current degraded CPU and specifically ask for it to be replaced with 285K. Intel owes him that.

 

[511]

I think he should RMA his current degraded CPU and specifically ask for it to be replaced with 285K. Intel owes him that.

Jut ask for a refund if it's available if not try asking for 285K

 

[SiliconFly]

.

 

[511]

What are we missing here? Are you implying PTL-H will get +35% battery life (than LNL) just cos it has more cores?

(Note: We're talking about +35% overall PTL-H efficiency (w.r.t Prakhar Verma tweet) and not specifics. And definitely not a specific point in the power/perf curve.)

+10-15% is coming from the node itself +5-8% from the microarch if PTL-U should have 15-20% gain

 

[FlameTail]

What are we missing here? Are you implying PTL-H will get +35% battery life (than LNL) just cos it has more cores?

Certainly not.

I am talking something like this;

(Multi Core Performance)

(Note: We're talking about +35% overall PTL-H efficiency (w.r.t Prakhar Verma tweet) and not specifics. And definitely not a specific point in the power/perf curve.)

What is the definition of "overall efficiency" ? Sounds rather vague.

 

[SiliconFly]

.

 

[511]

Certainly not.

I am talking something like this;
View attachment 111390
(Multi Core Performance)

What is the definition of "overall efficiency" ? Sounds rather vague.

Now it depends if it is 35% at 15W or not
I think at 50W we should expect 1400-1500 r24

 

[Hulk]

We've had a similar case with AMD this year: engineers at were testing CPUs in a way that was also disabling some security patches in Windows, which led to better performance. Those security patches were bound to be disabled in a new build of Win, so there was indeed some performance left on the table in early reviews. Normally I would criticize the vendor for this type of screw-up, but the reality is Intel is paying through their nose for this "unexpected" performance gap. The loss in sales and in mind share is telling them all they need to know.

IMHO we should cut Robert Hallock some slack, he's doing something Intel failed to do for years now, which is to talk about internal issues and communicate intention to fix things. He's also been dealt a very bad hand to work with here. The same applied for Tom Petersen on the GPU side, it's hard to get out there and expose yourself, so they deserve more leniency than usual (as messengers).

Let's wait and see what Intel cooks up, it's not like anyone is rushing to buy Arrow Lake anyway.

Where did you hear AMD was testing with some things disabled? Did AMD admit that? I don't mean that in a way that I don't believe it but good on them if they did. I get your points and they are good ones but I'm still left with an empty and betrayed after watching that video.

The thing that has me frustrated is that Intel is admitting or divulging nothing. You write unexpected performance gap. Intel won't even talk about performance gap. Only unexpected "results."

Is the released hardware that goes along with the software showing the same performance as Intel's testing as per pre release benches or not? If not why? If so then what are you fixing?

It's really simply actually. They could straighten this out in a 1 paragraph press release. But no they create a 1+ our spin video. I also love how the interviewers nod along approvingly at all of the non answers and silly excuses, "Oh yes! Of course! That makes sense!" I understand they are just happy to get the interview but what's the point if you can't ask pointed questions?

They should put Intel to the test. Ask the hard questions, ask the follow ups. If Intel refuses to answer end the "interview."

I think they played along and gave Intel a pass because its Intel.

Intel is getting smoked across the board right now by AMD. AMD has no degradation issues, AMD is more efficient, and AMD overall is more performant. It's time for Intel to level up.

This is Netburst vs. Athlon and there is no Core waiting in the wings for Intel. It's time to be self effacing, no arrogant and smug.

I don't blame the messenger, he seems like a really nice guy and is going what he can within the onerous confines Intel placed on what he can divulge. "Just tell them we'll fix it." Fix what? "You know just mix it up so they don't if if you're coming or going and say fix it and unexpected a lot."

As John Lennon said, "Just gimme some truth man."


BTW, good advice some people here regarding my 3rd degraded part. I'm going to ask for a refund. I don't want another defective part as I believe they are all the same (defective).

 

[SiliconFly]

.

 

[Hulk]

I don't think they had much of a choice. Possibly scripted and/or reviewed/approved before publishing.

That's fine. They should put a disclaimer up front explaining the following interview is "entertainment" and not journalism. You know, like how the WWE admits they are entertainment. I'm half, only half, joking here.

Good journalism is hard to find these days. We as consumers of information should nurture it by rewarding it when we see it and calling it out for what it is when we don't in a polite and respectful manner of course.

 

[Hulk]

My new first follow up.

"You say the results were unexpected. What was Intel expecting exactly?"

Then sit back and watch the dance.