Discussion Intel Nova Lake in H2-2026: Discussion Threads

[adroc_thurston]

They cancelled MTL-S

That thing never reached product design stage.

 

[511]

At this point i totally believe you have personal vendetta against Pat 🤣

 

[moinmoin]

... to at least push the halo part to greater than 10% IPC (through frequency gain ...

🤔
Is nobody going to tell him?

 

[jpiniero]

Seriously, textbooks will be written on how Pat fumbled the whole thing. He had complete control. Kept making grave mistakes. Over and over. His problem is that he didn't take much interest in engineering matters. Left it up to the fools. The same fools who came up with Rocket Lake and then Raptor Lake. I wouldn't be surprised if disabling AVX-512 was Pat's decision and it was made in a desperate attempt to upsell the Xeon W CPUs which are still priced extremely high. Intel had no interest in making money. Just wanted to keep up appearances of being their former self. A technology leader. Without any technical leadership.

The decision to give up on AVX-512 on big.LITTLE long predates Pat.

Fun Fact: Desktop Skylake went through 4 1/2 years of refreshes. Alder/Rocket/Bartlett would be 4 years if Bartlett gets released at the end of the year.

 

[igor_kavinski]

🤔
Is nobody going to tell him?

My mistake. I probably meant to say something other than IPC but can't think of it.

 

[poke01]

10% increase in ST is pathetic especially on N2. No wonder this company is dead

 

[poke01]

Like seriously AMD will lap them easily in nT too with less cores.

Who cares about wider vectors when you can’t even get a proper core design…

 

[adroc_thurston]

No wonder this company is dead

They're not dead.

Like seriously AMD will lap them easily in nT too with less cores.

Oh no, Intel should be ahead.

 

[MS_AT]

They're not dead.

Oh no, Intel should be ahead.

Ahead in which metrics?

 

[adroc_thurston]

Ahead in which metrics?

Cinememe 2024 nT.

 

[Kepler_L2]

Cinememe 2024 nT.

There's no way Zen6 is <60% MT

 

[adroc_thurston]

There's no way Zen6 is <60% MT

Remember that Cinememe2024 has hitlertastic hatred for AMD cores (relative to R23).
I haven't profiled it, but given Apple/QCOM results, the thing loves fat private caches, and AMD hates fat private caches.

 

[poke01]

Cinememe 2024 nT.

Wow, such leadership

 

[adroc_thurston]

Wow, such leadership

That's the only thing it could've ever won.

 

[Io Magnesso]

Intel seems to have created garbage with TSMC N2
The idea that using TSMC's process can improve is a silly idea.
The problem is to devise what kind of cooking can be done to provide an excellent CPU while using the TSMC process.

 

[Io Magnesso]

That being said, it's a big deal just to improve the IPC by 10(？)%. It is quite difficult to try to improve the IPC by improving the architecture compared to the past.
There is no magic

 

[poke01]

Intel seems to have created garbage with TSMC N2
The idea that using TSMC's process can improve is a silly idea.
The problem is to devise what kind of cooking can be done to provide an excellent CPU while using the TSMC process.

lets hope its an early estimate. They can improve the the P core, it won't launch for another 18 months.

 

[Kepler_L2]

lets hope its an early estimate. They can improve the the P core, it won't launch for another 18 months.

The ARL-S leak turned out pretty accurate https://www.igorslab.de/en/intels-i...n-for-raptor-lake-s-refresh-and-arrow-lake-s/

 

[branch_suggestion]

So lets make some estimates for NVL-S perf.
Gaming ~5% faster than 9800X3D, well behind Z6X3D.
1T, ~3900 GB6.4 vs ~4k+ Z6
nT, ~4k CB24 vs ~3750 Z6 24c (nearly Intel best case, AMD wins in anything FP heavy)
Not good famalam, considering this is an iso-node showdown.

 

[adroc_thurston]

Gaming ~5% faster than 9800X3D, well behind Z6X3D.

Gaming you have to account for bLLC options and idk the L3 perf on NVL.

 

[Kepler_L2]

So lets make some estimates for NVL-S perf.
Gaming ~5% faster than 9800X3D, well behind Z6X3D.
1T, ~3900 GB6.4 vs ~4k+ Z6
nT, ~4k CB24 vs ~3750 Z6 24c (nearly Intel best case, AMD wins in anything FP heavy)
Not good famalam, considering this is an iso-node showdown.

Regular NVL-S isn't getting anywhere near 9800X3D gaming perf, and bLLC is very TBD considering how garbage their current L3 is.

 

[branch_suggestion]

Gaming you have to account for bLLC options and idk the L3 perf on NVL.

I already did, that is my bLLC estimate.
Also assumes the L3 is pretty decent but not AMD level.
Without hybrid bonding it is hard to see them beat a similarly powerful core in extended L3 apps like gaming.

Regular NVL-S isn't getting anywhere near 9800X3D gaming perf, and bLLC is very TBD considering how garbage their current L3 is.

That is bLLC estimate, vanilla is getting crushed still of course.

 

[adroc_thurston]

Without hybrid bonding it is hard to see them beat a similarly powerful core in extended L3 apps like gaming.

EMR was like 5M@stop and purportedly decent.
Fingers crossed their L3 stops sucking donkey balls, especially now that they're having shared L2 with less net capacity per core.

 

[poke01]

N2 is wasted on Intel. IMO, desktop is okay but mobile looks grim.

 

[Darkmont]

I am curious as to what clocks they’re getting (or not) out of a quality node jump