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.)

 

[adroc_thurston]

as it will have APX and AVX10

This doesn't matter much.

on paper NVL should be greatest change since Alder Lake

Yeah because it has V$ variants for premium DT.

 

[fastandfurious6]

Cope Ultra 2% !!!! (dancing for the stakeholders)

https://twitter.com/x/status/1843598034250412141

 

[Josh128]

Low clocks, low cache. It's an overbloated LP ghetto.
The one in KRK should do a bit better given it's a uni-ccx with 16M L3$.

Is it though? Its not 4+4? Or is it just 8 Zen 5C cores?

 

[inquiss]

Is it though? Its not 4+4? Or is it just 8 Zen 5C cores?

Neither?

 

[poke01]

This doesn't matter much.

oh great

 

[inquiss]

This doesn't matter much.

Yeah because it has V$ variants for premium DT.

You think AMD will go that route in the future? AXV10?

 

[Wolverine2349]

Because they underperformed that's why. This is the best that both vendors can do and it wasn't enough. It's a strange question, like they've been holding back somehow.

Yes for sure.

Though Zen 5 did not regress from Zen 4 in gaming. It just was not any better or 3%.

But if Intel Arrow Lake regressed from Raptor Lake in gaming overall that's worse.

 

[Wolverine2349]

Cope Ultra 2% !!!! (dancing for the stakeholders)

View attachment 108932

https://twitter.com/x/status/1843598034250412141

2 percent improvement in gaming or overall performance improvement over Raptor Lake?

Zen 5 was joked Zen 3 percent in gaming.

 

[H433x0n]

If the above is true, that would be the main contributing factor for less than expected performance not moving to tiles.

So it'll go from 4.6GHz to 3.5GHz. That's a 31% increase in L3 cache latency. That's not only lower than Alderlake, but would be substantially slower in ST mode as Alderlake had trouble increasing ring clocks only when E cores were active.

I wonder what's contributing to that huge gap, curious if it’s node or design related.

2 percent improvement in gaming or overall performance improvement over Raptor Lake?

Zen 5 was joked Zen 3 percent in gaming.

Depends. It has strong integer performance. They’re claiming +4% over Zen 5 in SPECint 1-copy which seems believable based off of LNL.

The consensus seems to be +5% 1T and +15% nT with horrid gaming performance.

 

[adroc_thurston]

AXV10?

It's just AVX512 with smaller opt

 

[H433x0n]

Cope Ultra 2% !!!! (dancing for the stakeholders)

View attachment 108932

https://twitter.com/x/status/1843598034250412141

Everybody is celebrating but not realizing they’re also dunking on Zen 5 and cheering on the death of x86.

Both Intel and AMD having bad generation uplifts is bad for x86 and DIY PC.

 

[Jan Olšan]

This gloom will only last till Nova lake.

on paper NVL should be greatest change since Alder Lake, APX, 14A and total new architectures to boot. save money and wait till then if your on Alder lake or Zen4/5.

Nova lake should better than Zen6 too as it will have APX and AVX10. But as always things go sideways at Intel so who the ***** knows

AVX10 is strictly worse thing than AVX-512 that Zen 5 has, it is likely inferior to Zen 4's version of AVX-512.

APX okay, that is in interesting addition, but in last 15+ years, CPU cores have developed many tricks to sidestep all the deficiencies that APX seeks to correct, so there won't be some large benefit from it, IMHO.

 

[Wolverine2349]

Everybody is celebrating but not realizing they’re also dunking on Zen 5 and cheering on the death of x86.

Both Intel and AMD having bad generation uplifts is bad for x86 and DIY PC.

Well if it is on par in 1% and 0.1% lows in gaming with Raptor Lake it's fine and worth it.

Lower power stability and should age well as games get more threaded compared to stuck at only 8 cores X3D chips.

 

[poke01]

AVX10 is strictly worse thing than AVX-512 that Zen 5 has, it is likely inferior to Zen 4's version of AVX-512.

APX okay, that is in interesting addition, but in last 15+ years, CPU cores have developed many tricks to sidestep all the deficiencies that APX seeks to correct, so there won't be some large benefit from it, IMHO.

If that’s the case AMD is already ahead in some areas then. All AMD needs is N3E or N3P to get higher clocks for Zen6

 

[poke01]

Everybody is celebrating but not realizing they’re also dunking on Zen 5 and cheering on the death of x86.

Both Intel and AMD having bad generation uplifts is bad for x86 and DIY PC.

https://twitter.com/x/status/1843805974269964700

This person seems happy

 

[DavidC1]

Yeah but you also have more raw capacity at 17clk a core, plus L1.5d on top of that. LNC by all accounts should cope very well with slower L3.
Like this isn't the first time Intel L3 got slower. They coped well before that, so unlikely LNC copes any worse than SNC in TGL did.

We're talking about a general purpose processor. Every aspect matters. Also, that's just clock based regression. Cycle latency at the same MHz could have regressed on top of that.

Plus it's not like Arrowlake is bad in all games, it just has a big variance.

I still would wait independent benchmarks for actual analysis, because even if Intel isn't misleading/lying/made a mistake with the benchmarks, they are just not as thorough as dozens of sites testing with various setups and methodologies.

 

[fastandfurious6]

ARL mobile will be the actual power/thermal benchmark

desktop power draw is basically irrelevant..... that's why you get desktop...... for "unlimited" power and 5 whopping kilograms of liquid cooler lol

and I dont expect ARL-M to do any better than halo/9945hx in thermal efficiency

dragon range 7945hx @80w was performing ~80% of desktop 7950x already

 

[DavidC1]

Everybody is celebrating but not realizing they’re also dunking on Zen 5 and cheering on the death of x86.

Both Intel and AMD having bad generation uplifts is bad for x86 and DIY PC.

Both companies were at the lowest moments before their brightest. Hope that both can do the same.

This person seems happy

Presler, the 65nm shrink to Prescott also saved power. The problem was it just wasn't that impressive even with it. The new process is probably responsible for the power reduction.

 

[Josh128]

2 percent improvement in gaming or overall performance improvement over Raptor Lake?

Zen 5 was joked Zen 3 percent in gaming.

Its negative 2 percent in gaming under Raptor Lake, my dude.

 

[Josh128]

Everybody is celebrating but not realizing they’re also dunking on Zen 5 and cheering on the death of x86.

Both Intel and AMD having bad generation uplifts is bad for x86 and DIY PC.

True fans of tech arent celebrating. Im not. I am having fun dunking on the idiots over at WCCFTECH who are extremely obnoxious and base their self worth off of products that their favorite company produces, but thats a different thing altogether. As someone with money invested in Intel and also looking to upgrade from my Zen 3 system who has been less than impressed with Zen 5 thus far, I was hoping ARL would be everything Zen 5 was not. What I am seeing is that its basically a clone of Zen 5 perf wise, which is OK, but it all comes down to pricing.

 

[Joe NYC]

Low clocks, low cache. It's an overbloated LP ghetto.
The one in KRK should do a bit better given it's a uni-ccx with 16M L3$.

That's excellent news for Kraken, if true. 4x Zen5 + 4x Zen 5c are on the same ring bus sharing L3 will improve efficiency all around (power efficiency, die area efficiency, efficiency in utilizing L3). If the clock speeds are on par with Strix, this will be a winner and give Lunar Lake run for its money in the CPU department.

Lunar Lake will still lead in GPU, since Kraken will have smaller GPU.

 

[jpiniero]

but it all comes down to pricing.

Given that it uses the extremely pricey N3, don't expect much on that end.

 

[Wolverine2349]

True fans of tech arent celebrating. Im not. I am having fun dunking on the idiots over at WCCFTECH who are extremely obnoxious and base their self worth off of products that their favorite company produces, but thats a different thing altogether. As someone with money invested in Intel and also looking to upgrade from my Zen 3 system who has been less than impressed with Zen 5 thus far, I was hoping ARL would be everything Zen 5 was not. What I am seeing is that its basically a clone of Zen 5 perf wise, which is OK, but it all comes down to pricing.

As long as it performs on par in gaming to Raptor Lake. Negative 2 percent is almost margin of error if that's true.

But Zen 5 is like vanilla Zen 4 for gaming which is way behind Raptor Lake.

If it performs real world like vanilla Zen 4 or 5 ouch.

We will know 2 weeks from Thursday.

 

[Josh128]

Given that it uses the extremely pricey N3, don't expect much on that end.

I wouldnt be so sure. At least I no longer think it'll be more expensive than Zen 5 / 4nm. The reason why is Huang and his damned H100/200/ Blackwell AI GPUs. He is selling an enormous quantity them at an enormous price and TSMC picked up on that a year or so ago and increased 4nm pricing. As of now, Im not too sure of any really big players besides Intel and to a lesser degree, AMD, that are buying up that 3nm production line. NV's next GPU may be on 2nm or even 18A (or both). I have come to believe that 4nm wafer price hike by TSMC is the reason that Zen 5 launched with seemingly mismatched pricing and has seen relatively muted price drops despite by most accounts, having terrible sales.

 

[Joe NYC]

Given that it uses the extremely pricey N3, don't expect much on that end.

It will be tough for Intel to price Arrow Lake competitively with expensive BOM.

While at the same time compete with 9800x3d that will outperform it by miles at ~ $449 price. How can Intel price the 285V?

The best option may be just supplying small numbers of these chips and redirect N3B capacity to Lunar Lake.

Also, it is unclear if there will be a supply of CPUs with smaller 6+8 dies, given the switch of these from Intel 20A to TSMC N3B is so recent.