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.)

 

[511]

@Hulk it sure is tempting at that price

 

[Hulk]

Sadly you're still stuck with a board for which you won't get any refunds.

I prefer to see the glass being half full.

 

[DrMrLordX]

Some people made it process issue apparently saying Intel 7 is bad

Let's hope that isn't the case. Doesn't seem like Alder Lake had that problem.

I prefer to see the glass being half full.

Something is better than nothing.

 

[Hulk]

Igor sent me a review of the 285K that seems quite favorable to Intel. Some of the stats made me go "hmmmm?"

Accord to this review the 285K is much easier to cool than the 9950X.

Intel Core Ultra 9 285K CPU Review: Performance, Thermals & Power Analysis - Page 10 of 12 - Hardware Busters

Hardware Busters - Intel Core Ultra 9 285K CPU Review: Performance, Thermals & Power Analysis - CPU

hwbusters.com

Also showing idle system power for Intel as half that of AMD. Total load system power is much lower as well?

Not what I had previously thought?

Intel Core Ultra 9 285K CPU Review: Performance, Thermals & Power Analysis - Page 9 of 12 - Hardware Busters

Hardware Busters - Intel Core Ultra 9 285K CPU Review: Performance, Thermals & Power Analysis - CPU

hwbusters.com

 

[Abwx]

Total load system power is much lower as well?

Not what I had previously thought?

Intel Core Ultra 9 285K CPU Review: Performance, Thermals & Power Analysis - Page 9 of 12 - Hardware Busters

Hardware Busters - Intel Core Ultra 9 285K CPU Review: Performance, Thermals & Power Analysis - CPU

hwbusters.com

Lol, here what he says, that the 9950X power increased unexpectedly without even touching something since his first review, just in time apparently.

At stock the 9950X is limited to 200W.

 

[DavidC1]

You will have a hard time degrading it at 5 GHz. These chips are very easy to cool, so beside the lower voltage you get a huge benefit of low temps.

I thought it was the ring that degraded because it was too high on Raptorlake?

Let's hope that isn't the case. Doesn't seem like Alder Lake had that problem.

Alderlake had the ring clock much lower than Raptorlake. There's a 900MHz difference. I think it's 3.6GHz vs 4.5GHz. Arrowlake is 3.8GHz. Don't think it's a coincidence. Haswell with the 4790K was when they really started ramping the clocks again. Haswell was also the chip that made ring asynchronous with the clocks. Sandy Bridge's CPU clocks were identical to ring.

IO is generally harder to clock higher so that's why I keep saying back off from the ridiculous clocks.

I do believe Intel's claims that they messed up on the optimization part with Arrowlake though. Probably can get ~5% in gaming and reduce variations run-to-run without needing to clock any parts higher.

 

[DrMrLordX]

What's the point of running a 14700k @ 5 GHz max frequency versus just running a 12900k/s? Would there be any performance advantage at that point? You'd get a few more ecores, but that's about it . . .

 

[Hulk]

What's the point of running a 14700k @ 5 GHz max frequency versus just running a 12900k/s? Would there be any performance advantage at that point? You'd get a few more ecores, but that's about it . . .

I was thinking about that today. 200W on my 14900K. 8+16 CB R23 = 35000, 8+12=31000

Handbrake test on this forum time goes down by 10 seconds when you add in the additional 4 cores for the same total power.

So yes, for applications that will use the additional threads the E's, even an additional cluster make a significant difference.

 

[poke01]

I was thinking about that today. 200W on my 14900K. 8+16 CB R23 = 35000, 8+12=31000

Handbrake test on this forum time goes down by 10 seconds when you add in the additional 4 cores for the same total power.

So yes, for applications that will use the additional threads the E's, even an additional cluster make a significant difference.

Some 12900K dies can make use of AVX-512 when e-cores are disabled, you are not getting this done are you

 

[Hulk]

Some 12900K dies can make use of AVX-512 when e-cores are disabled, you are not getting this done are you

I had a 12700K a few years ago and was able to enable AVX-512 and run the Handbrake test with it enabled in the app as well. The improvement in performance wias negligible.

 

[511]

I had a 12700K a few years ago and was able to enable AVX-512 and run the Handbrake test with it enabled in the app as well. The improvement in performance with negligible.

Black friday is soon yo be upon us

 

[Kocicak]

What's the point of running a 14700k @ 5 GHz max frequency versus just running a 12900k/s? Would there be any performance advantage at that point? You'd get a few more ecores, but that's about it . . .

Raptor lake had inter-core latencies improved compared to Alder lake.

 

[SiliconFly]

.

 

[maddie]

I think it's better to get a new 14900K, set the frequency limits properly, apply the new bios/ucode updates, etc. It should then run fine for a long time.

Yep, 4th time's the charm. Great advice to a colleague.

 

[Racan]

Yep, 4th time's the charm. Great advice to a colleague.

 

[SiliconFly]

.

 

[511]

Rubbish. A new 14900K with all the fixes shouldn't pose any problems at all. What else do you suggest, some silly amd processor?

14700K is better than 14900K as an option

 

[511]

Some panther lake leaks

消息称英特尔 Panther Lake 处理器大核显版本 2025 年末率先推出 - IT之家

Panther Lake 移动处理器将采用 Cougar Cove 架构性能核、Darkmont 架构能效核、“Celestial”Xe3 架构核显、NPU 5。

m.ithome.com

 

[OneEng2]

Foolish advice IMO.

If Hulk is a gamer, the answer is clear. The 9800 X3D shortly after the release of the higher clocked models will be the best gaming processor for the money on the planet for quite some time I would expect.

There is even the opportunity for a future Zen 6 upgrade in a couple of years.

How is spending more money on an obsolete platform with inferior performance a good investment?

 

[OneEng2]

I will concede that the lest expensive cure is another Intel chip.

 

[SiliconFly]

.

 

[gdansk]

A replacement is free. No money needs to be spent. Investing money is foolish advice IMO when it can be done for free.

A refund was offered. Try to keep up!

 

[SiliconFly]

.

 

[gdansk]

He can always opt for a free replacement instead. Keep up!

If you knew a refund was an option then why complain about spending money?
The only sunk cost is the board. Meanwhile the advice here is to take another replacement or buy a 14700K and underclock it to 5GHz. Huh, okay, some people really like being abused.

 

[511]

The best bet is 14700K on Black friday safer than 14900K and he saves money for future upgrade just buy Zen6 when it comes out