Discussion Intel current and future Lakes & Rapids thread

[Geddagod]

Parallel development isn't a big issue for Intel anymore. Remember Jim Keller's involvement...

Don't think Keller had anything to do with that. IIRC Keller claimed most of what he had to do with is the organizational changes of Intel.

starting from LNC all Intel core designs are purely node-agnostic. Makes it a lot easier to port from one node to another.

SNC actually. Intel claimed that even before RKL launched IIRC.
If you are talking about Intel + TSMC, then crestmont might be the first. Though I'm also not sure there aren't any irrelevant core ip (or tbh even gpu ip) that also wasn't straight up ported between the two fabs.

I think ARL is gonna take efficiency to the next level. Unquestioned leadership? Maybe... maybe not... too early to speculate

If ARL doesn't win the perf/watt crown, it's going to be extremely embarrassing for Intel, esp at lower power levels. I'm still holding out hope LNC isn't completely a bust, and that even if they threw away a bunch of perf, they might have kept the good perf/watt scaling aspect of the arch, but who knows.

 

[TESKATLIPOKA]

Is MTL a good product? I think so too. But definitely a lot better than previous gen.

MTL will be a better product than Raptor, I agree. I think It will be better than Phoenix too in some things and comparable in others.
The question is how much time will take until Strix Point is out, because in ST, MT, IGP and efficiency in my opinion Strix will win.

At least I would wait for CES 2024. I think AMD will show something, so then we will see If It's worth the wait for Strix or not.

 

[Goop_reformed]

Too early to say. Looks like ARL is also all about efficiency like MTL (which isn't a bad thing I'd say).

But can't totally rule out ARL performance gains this early. LNC might move the needle a bit.

Well the projection slides said otherwise, and so far there has nothing that supports a higher performance gain.

 

[SiliconFly]

If ARL doesn't win the perf/watt crown, it's going to be extremely embarrassing for Intel...

True. If Intel screws-up ARL, it's actually over for them. They'll lose all respect and may never bounce back. Well, truth is, they don't command much respect even now. And parts like RPL-R only makes things even worse.

 

[SiliconFly]

Well the projection slides said otherwise, and so far there has nothing that supports a higher performance gain.

Well, the projection slides were discussed before as it was generally agreed upon that it may not reflect final performance as it lacks context.

 

[SiliconFly]

Don't think Keller had anything to do with that. IIRC Keller claimed most of what he had to do with is the organizational changes of Intel.

During his time at Intel:

(1) He fixed some mistakes of the past. <----- (Process related)
(2) He started the new core design (LNC & beyond) <----- (MLID's babble about royal core)
(3) He pushed Intel to switch to TSMC (with mixed results) <----- (Rumors say major fallout due to this)
(4) He made the core designs node agnostic. <----- (Starts with ARL instead of MTL)

Among other things.

Guy is truly a rock star!

 

[Goop_reformed]

Well, the projection slides were discussed before as it was generally agreed upon that it may not reflect final performance as it lacks context.

There have been zero slides that substantiate your claim so far. The only slides we have are from Igor and an user in this forums. Everything else's, including our alleged resident leaker's info, are only farts in the wind.

 

[mikk]

Well the projection slides said otherwise, and so far there has nothing that supports a higher performance gain.

This is for desktop. Arrow Lake mobile is in a much better place because the clock speeds for Raptor Lake mobile are lower and even lower on Meteor Lake (ST max speed). MT speeds depends more on efficiency for power limited mobile CPUs, ARL should have a big advantage there.

 

[SiliconFly]

There have been zero slides that substantiate your claim so far. The only slides we have are from Igor and an user in this forums. Everything else's, including our alleged resident leaker's info, are only farts in the wind.

Unfortunately for you, people in this forum are usually well informed and don't deal with farts.

There have been twitter/X leaks that are more credible than the slides you mentioned. With all due respect, time for you to do some homework cos we've already done it & are not gonna do it again for you. Kindly do it yourself.

 

[Goop_reformed]

Unfortunately for you, people in this forum are usually well informed and don't deal with farts.

There have been twitter/X leaks that are more credible than the slides you mentioned. With all due respect, time for you to do some homework cos we've already done it & are not gonna do it again for you. Kindly do it yourself.

Well unfortunately, the general consensus both here and twitter is rather bleak toward arrow lake. And ease off on the animosity, I'm merely saying what's circling around. And if you want to talk about leaks from lenovo I'm all ears.

 

[H433x0n]

Well unfortunately, the general consensus both here and twitter is rather bleak toward arrow lake. And ease off on the animosity, I'm merely saying what's circling around. And if you want to talk about leaks from lenovo I'm all ears.

I’d be happy to talk leaks out of Lenovo but far as I know those leaks don’t contain any solid info. As far as twitter I completely disregard majority of the info since most of the prominent leakers are AMD super fans.

So far the only real info we have is IgorsLab and that slide from a member on this forum.

 

[Goop_reformed]

I’d be happy to talk leaks out of Lenovo but far as I know those leaks don’t contain any solid info. As far as twitter I completely disregard majority of the info since most of the prominent leakers are AMD super fans.

So far the only real info we have is IgorsLab and that slide from a member on this forum.

Well Siliconfly keeps insisting on higher performance gain even though all existing evidence plus leaks indicate otherwise. If you read my posts they are actually grounded in reality.

 

[jpiniero]

MTL will be a better product than Raptor, I agree.

So far the leaked scores for Meteor have been underwhelming. It may not be that much better than Raptor. Let alone Phoenix or Strix Point.

 

[SiliconFly]

Well Siliconfly keeps insisting on higher performance gain even though all existing evidence plus leaks indicate otherwise. If you read my posts they are actually grounded in reality.

None of the ARL leaks so far are credible enough to be taken seriously. They either lack context or mere speculations or are not trustworthy enough. Same goes with the "leakers". Always trace the leaks to the source.

The performance of LNC is not known yet. Even though we can speculate a bit on power efficiency, there's no way to judge the performance of LNC final as they don't even have an ES as of now.

 

[biostud]

Without reading any rumors whatsoever my guess will be that the first intel4 chips will have problems reaching high clocks, and it will take a couple of generations moving from laptop to desktop just like Cannon lake->Ice lake->Tiger lake->Alder lake.

 

[DrMrLordX]

None of the ARL leaks so far are credible enough to be taken seriously.

That leaked Intel slide seemed pretty legit, even if it was using a PL2 value (250W) that we now know won't be in the final product (more likely to be PL2 = 177W).

 

[igor_kavinski]

(more likely to be PL2 = 177W).

That looks more to me like a TSMC silicon limitation. If Intel could push it to 250W, why wouldn't it?

 

[yuri69]

True. If Intel screws-up ARL, it's actually over for them. They'll lose all respect and may never bounce back. Well, truth is, they don't command much respect even now. And parts like RPL-R only makes things even worse.

"Respect"? Respect of the niche DIY builders community? The rest of the whole x86 market is commanded by OEMs/hyperscalers with their massive orders. Intel will be fine since it can always flood the market with *something*.

 

[Goop_reformed]

None of the ARL leaks so far are credible enough to be taken seriously. They either lack context or mere speculations or are not trustworthy enough. Same goes with the "leakers". Always trace the leaks to the source.

The performance of LNC is not known yet. Even though we can speculate a bit on power efficiency, there's no way to judge the performance of LNC final as they don't even have an ES as of now.

Those slides are from intel, what are you even talking about? Intel is a leaky ship, this has been known for years. Disgruntled HB1 visa holders talked and oh boy do they talk a lot.

 

[coercitiv]

That leaked Intel slide seemed pretty legit, even if it was using a PL2 value (250W) that we now know won't be in the final product (more likely to be PL2 = 177W).

That leaked slide did look legit, yet AFAIK it lacked context: we did not know if the ARL estimate was final target. It was still a problem though, as ARL needs to be a generational upgrade like ADL was.

That looks more to me like a TSMC silicon limitation. If Intel could push it to 250W, why wouldn't it?

Not only a node limitation, but also possibly a tile design limitation. Monolithic designs handle heat dissipation better, on top of that the tiled design may be subject to influence from thermal effects (expansion, contraction) which may impose further temperature limits. All this can (not will) lead to reduced efficiency in heat dissipation, leading to a lower max power target.

On a personal note, just be glad we're no longer getting 250W+ PL2., although the harm is mostly done as AMD has already followed suit with Zen 4.

 

[DrMrLordX]

That looks more to me like a TSMC silicon limitation. If Intel could push it to 250W, why wouldn't it?

We don't know if that slide pertained to N3B or 20a compute tiles.

That leaked slide did look legit, yet AFAIK it lacked context: we did not know if the ARL estimate was final target.

Yeah but why put an early buggy ES estimate into a slide deck? Something there doesn't add up.

On a personal note, just be glad we're no longer getting 250W+ PL2., although the harm is mostly done as AMD has already followed suit with Zen 4.

How would that slide have looked if the PL2 target were 177W instead of 250W? A 7950X loses very little MT performance moving from 170W TDP/230W PPT down to ~125W PPT. But a 13900k/ks/14900k loses more . If Arrow Lake doesn't scale well with increased PL2 values then it may still be relatively competent @177W PL2.

 

[coercitiv]

Yeah but why put an early buggy ES estimate into a slide deck? Something there doesn't add up.

Possibly because some exec out there wanted reassurance they are on track. You put the buggy early ES estimate to reassure some nervous manager they won't miss they next holiday bonus. It's all speculation though, like many other partial leaks about ES performance it helps us set the lower bound for performance, helps us very little with the upper bound.

How would that slide have looked if the PL2 target were 177W instead of 250W?

ARL should have looked better @ 177W relative to RPL. In fact it should get better and better as it moves down the power curve.

 

[Goop_reformed]

So here are the slides again: https://www.igorslab.de/en/intels-i...n-for-raptor-lake-s-refresh-and-arrow-lake-s/

It clearly says performance projection, now why on earth would they project an ES performance? Makes zero sense. The only thing missing here is clock speed, and that's it!.

AFAIK, putting a number as a goal is not how chips, nor any other engineering projects are done. The designers would clear out the low hanging fruits first then do what they can with the rest with compromises. Again, putting a number early is almost never done. Getting more performance at the last minute like with zen 1 over Excavator took a huge toll and decision like that is almost always abandoned.

Trying to make excuses to circumvent facts, which are clearly shown, is mental gymnastic and it doesn't help with the discussion at all.

Having said, even after asking a bunch of people here and on my alma mater, I still think arrow lake final performance number will be better than a meager single digit ipc uplift. It's just now possible for a company like intel to fail this hard.

 

[DrMrLordX]

Possibly because some exec out there wanted reassurance they are on track. You put the buggy early ES estimate to reassure some nervous manager they won't miss they next holiday bonus. It's all speculation though, like many other partial leaks about ES performance it helps us set the lower bound for performance, helps us very little with the upper bound.

If I were in that position, I'd use my simulations (which is probably what those numbers were based off of, or very early samples returned from the fab) to guesstimate what the final product would look like. Which is what I'm assuming they did there, but who knows, there might have been another slide explaining exactly how they arrived at those numbers.

ARL should have looked better @ 177W relative to RPL. In fact it should get better and better as it moves down the power curve.

Well that's the rub, in a lot of media venues Arrow Lake-S @ 177W will be compared to Raptor Lake-S @ 250W. And that's not fair in a lot of ways, but such is the state of tech journalism.

 

[igor_kavinski]

It's just now possible for a company like intel to fail this hard.

It is possible if Intel was confident that there was no way AMD could surpass their "world class engineering". Internally, they probably figured that ADL would demolish AMD and before it could get up again, the meagre 10% IPC uplift of ARL would finish them off and send them back into the sad Bulldozer era of embarrassing struggles. ADL/ARL designs were finalized years before Intel could even imagine something like Zen 3 eating their lunch. Now the same thing is supposed to repeat with Zen 5 and later on, with Zen 7.