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.)

 

[TESKATLIPOKA]

It will recognize faces from your photos a tad little faster.
Not happy ? My god....

Bye bye offline privacy.

 

[Thibsie]

Bye bye offline privacy.

Well this functionality is, in fact, sought after.
Either you do it with the CPU, the NPU/GPU are files are uploaded so a server can do it faster but... privacy bye-bye.

I can totally see why it is useful to have the photo software get you every photo of, let's say, daughter number 1 without having to do the job yourself.
Seems a valid use case to me.

 

[gdansk]

Well this functionality is, in fact, sought after.
Either you do it with the CPU, the NPU/GPU are files are uploaded so a server can do it faster but... privacy bye-bye.

I can totally see why it is useful to have the photo software get you every photo of, let's say, daughter number 1 without having to do the job yourself.
Seems a valid use case to me.

It is a useful feature and one that most customers would want to run on their machine instead of in the cloud whether they realize it or not. Who wants Google/Apple sending pictures of their kids to their servers if there is a good local alternative?

But it isn't clear to me if customers would prefer that processing to be fast or power efficient. I'd lean toward the latter and in that case total peak TOPs doesn't seem too relevant.

 

[eek2121]

Yeah the TOPS is the bad news.

I was expecting MTL to have 45+ TOPS NPU.

I am surprised it's not the case.

Not a single CPU released (or releasing this year) has an NPU that does 20 TOPs, let alone 40+

Bye bye offline privacy.

If Windows 12 is as bad/worse as 11, my privacy will be intact because I will finally move to linux.

 

[SiliconFly]

I think i may have figured out why Intel marketing came out with the snake oil stuff out of the blue! MTL launch is just a few days away and AMD has the tendency to crash Intel's party by saying stuff like Intel's cores are monolithic and old and their chiplets are the best. Or stuff like, Intel's nodes are bad and their nodes are superior, etc. Same goes with the cores like Zen architecture is better than Intel core, etc, etc.

And as far as MTL is concerned, everything in it is new, except for the CPU tile which has RWC which is just a rehash of RPC which in turn is a rehash of GLC. Afaik, this is Intel's achilles heel. And probably the singular issue with MTL. And they know AMD will somehow bring it up to rain on the MTL parade. So, it appears they chose a preemptive strike. A carefully orchestrated attack on AMD to ensure they doesn't get too cozy taking potshots this time.

 

[coercitiv]

A carefully orchestrated attack

This move is already adequately explained by stupidity. No need to look for another reason.

 

[gdansk]

Stupidity + opportunity = gaffe.
It doesn't have much to do with MTL.

 

[Saylick]

HXL on Twitter posted some benches from bilibili on MTL. Not sure how legit, not sure what laptops were compared, what kind of battery size, etc.

视频去哪了呢？_哔哩哔哩_bilibili

undefined, 视频播放量 undefined、弹幕量 undefined、点赞数 undefined、投硬币枚数 undefined、收藏人数 undefined、转发人数 undefined, 视频作者 undefined, 作者简介 undefined，相关视频：

t.co

Battery Life (Minutes):

 

[Saylick]

More MTL stuff from HXL (from Bilibili), this time about the iGPU and CB:

 

[dullard]

More MTL stuff from HXL (from Bilibili), this time about the iGPU and CB:

Looks like Ultra 7 155H wins some, 13500H wins some, and 7840HS wins some. But the Meteor Lake is at 28 W vs 35 W or more for the others?

Also, why Windows 10?

 

[Hitman928]

HXL on Twitter posted some benches from bilibili on MTL. Not sure how legit, not sure what laptops were compared, what kind of battery size, etc.

视频去哪了呢？_哔哩哔哩_bilibili

undefined, 视频播放量 undefined、弹幕量 undefined、点赞数 undefined、投硬币枚数 undefined、收藏人数 undefined、转发人数 undefined, 视频作者 undefined, 作者简介 undefined，相关视频：

t.co

View attachment 90049
View attachment 90050
Battery Life (Minutes):
View attachment 90051

Their 13500h results seem to be in line with notebookcheck's results but their 7840hs results are really low (outside of geekbench5 multi for some reason). If their Intel results are true, it's pretty disappointing. Battery life is improved but performance is significantly worse tier to tier (unless the 125 is supposed to be lower than the 13500H?). Without knowing what the exact PL setting and such were, it's hard to really say though.

 

[dullard]

but their 7840hs results are really low (outside of geekbench5 multi for some reason)

The 7840HS is configurable from 35 W to 54 W. I agree: without knowing the power level of the tests, it is hard to really say if the results are good or bad.

 

[Hitman928]

The 7840HS is configurable from 35 W to 54 W. Without knowing the power level of the tests, it is hard to really say if the results are good or bad.

The results shown are significantly lower than the lowest of notebookcheck's results. I mean, it's possible that their configuration is hard locked at a lower power than any of notebookcheck's tested configurations, but that seems unlikely, especially given the large difference and the inexplicable geekbench5 result which matches notebookcheck's top result. If it was a hard power limit, there's no way that could happen. Only thing I can think of is if maybe they tested the Ryzen on battery power for most of the tests except GB5 multi. That's the only thing that I can think of that would make the results make sense (but obviously still invalid). No real idea what's going on, but their Ryzen results make me question if their MTL numbers are legit either.

 

[TESKATLIPOKA]

Without actual power these results are meaningless.
At least IGP looks decent. 3077 in Graphic score is not bad at all and higher than PHX. The question is how It performs in actual games.

 

[cebri1]

Some numbers are off. 125H should have better battery life than the 155H.

 

[TESKATLIPOKA]

Some numbers are off. 125H should have better battery life than the 155H.

Battery could have a lower capacity or bad optimizations, you can see It all the time in laptops despite having comparable specs.

 

[cebri1]

Battery could have a lower capacity or bad optimizations, you can see It all the time in laptops despite having comparable specs.

That's my point. If a CPU with less cores and lower frequencies has worse battery life is that the comparison is not well made.

 

[Saylick]

More MTL stuff from HXL (from Bilibili), this time about the iGPU and CB:
View attachment 90053
View attachment 90054
View attachment 90055
View attachment 90056

HXL has this to add:

https://x.com/9550pro/status/1733146964437742034?s=20

Power
R23~40w
TS~38W

 

[Abwx]

From NBC tests it look like the 7840HS is set at about 35W, the 7840U score about 12000@30W.

Also notice the evolution from R15 to R23, no need to improve your CPU IPC, "improve" the bench instead, that work much better, same from GB5 to GB6...

Edit : The difference between 125 and 155 in battery life is not significant, in the first test it s 3%, so within error margins.

 

[dullard]

Why 3 GHz for the 155H? I thought the base was supposed to be far higher than that.

 

[dullard]

Some numbers are off. 125H should have better battery life than the 155H.

Look up "Race to idle". If both turbo to the same power level, but the 155H gets done sooner, then it will sit idle for longer. The 125H will have to stay in high power turbo mode for longer, using more energy. Thus, the 155H should have a longer battery life.

 

[Hitman928]

Look up "Race to idle". If both turbo to the same power level, but the 155H gets done sooner, then it will sit idle for longer. The 125H will have to stay in high power turbo mode for longer, using more energy. Thus, the 155H should have a longer battery life.

Most battery life tests aren’t run this way. They will do some combination of constant web loading tests, video loop tests, pdf/ebook tests, long idle, and constant load tests. Race to idle doesn’t affect any of these.

Very few do a total energy used type test where race to idle can have a large effect but usually they put them under efficiency sections.

 

[TESKATLIPOKA]

HXL has this to add:

https://x.com/9550pro/status/1733146964437742034?s=20

12000 pts at 40W is not a very good result.
6P+8E i7-1370P manages 13,159 at PL1: 55W, PL2: 30W
4P+8E i7-1360P manages 11,072 at PL1: 43W, PL2: 28W

 

[dullard]

deleted

 

[Geddagod]

If MTL has worse or ~ same perf/watt than RPL, it's gonna be ICL all over again lol