• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

Alder Lake-S leaks

But the discussion is about Alder Lake, that's the point. Or should discussion be confined purely to a little tidbit of information and then halted completely until something else comes up? There is nothing to discuss beyond speculation and rumour until a real product lands on our doorstep. Such speculation is, of course, performance figures and how it will stack up to the competition. Ppart of that involves Intel's marketing numbers, which automatically fall under scrutiny because A: don't trust marketing, B: especially don't trust Intel's marketing, C: Intel's proven track history with delivering on those numbers, or not as the case often is.

It's just unfortunate that such discussion will inevitably be negative because Intel have not delivered anything close to their marketing numbers for quite some time. That's just the way it is. So I'm sorry if you think we're going on bashing Intel, but that's just the way it is right now. Intel haven't delivered anything like they've claimed for ages and a lot of their marketing has been heavily skewed, biased and outright paid shill nonsense (especially true since Ryan Shrout got hired).

This is not emotive or bias or fanboyism, this is cold, hard, fact. This is also a valid part of the discussion.

So your comments on the slide were about Alder Lake? If they were then apologies, I misunderstood.
 
Well, yeah. That low-end 8+8 core Xeon is Alder Lake-S, and Sapphire Rapids is based on Golden Cove, the same as Alder Lake. That's fostered further discussion that the mid-range HEDT is going to be Ice Lake, a 2 generation old mobile architecture rumoured to compete with only Zen 2, yet everything still hinges on goddamn Cascade Lake (REALLY?) until Sapphire Rapids arrives.

Overall it remains a depressing year for Intel.
 
Something tells me that AlderLake-S for Crimbo will be the system to have going into 22.

You can only create so many duds before you come up with a straight flush.

moving away from 14++++++++++++++++++++++++++++ will prove dividend.
 
Something tells me that AlderLake-S for Crimbo will be the system to have going into 22.

You can only create so many duds before you come up with a straight flush.

moving away from 14++++++++++++++++++++++++++++ will prove dividend.
But Intel have only made 1 dud in Rocket Lake. The endless of Skylake rehashes are tired yes, but hardly duds. Moving to 10nm is going to take a huge amount of pressure off power and thermal requirements for sure, but you're still creating a turd if the arch isn't up to scratch.

And here's the thing: the big little concept is unproven in x86. It took Intel years to get Lakefield scheduling half-decently and now there's twice as many little cores and 8 times as many big cores to deal with. In very real terms, Alder Lake could just end up being yet another 8 core CPU. The 20% IPC bump - if it happens - would handily retake the gaming crown, but that's all it'll do. Intel still have a core deficit, still have a productivity deficit and still have no answer to AMD in any of the markets that these businesses actually care about.

And then Zen 4 lands.

Christmas 2021? I gather it was only ever going to be a paper launch anyway with volume due H1 next year. And this is entirely reliant on DDR5 also being available in volume at a consumer level. Are Intel going to allow mixed memory motherboards, or dictate vendors must choose one or the other? Given the change in CEO recently, are Intel going to continue their convoluted and largely fictitious marketing messages and rush Alder Lake out the door with DDR4 saying "but if you want DDR5 you have to wait for a bit", or just EOL the DDR4 boards after a quarter once DDR5 is more available? Maybe the entire thing will be held off until Microsoft actually sort a half-decent scheduler so Alder Lake's entire design philosophy won't crash and burn.
 
But Intel have only made 1 dud in Rocket Lake. The endless of Skylake rehashes are tired yes, but hardly duds. Moving to 10nm is going to take a huge amount of pressure off power and thermal requirements for sure, but you're still creating a turd if the arch isn't up to scratch.

And here's the thing: the big little concept is unproven in x86. It took Intel years to get Lakefield scheduling half-decently and now there's twice as many little cores and 8 times as many big cores to deal with. In very real terms, Alder Lake could just end up being yet another 8 core CPU. The 20% IPC bump - if it happens - would handily retake the gaming crown, but that's all it'll do. Intel still have a core deficit, still have a productivity deficit and still have no answer to AMD in any of the markets that these businesses actually care about.

And then Zen 4 lands.

Christmas 2021? I gather it was only ever going to be a paper launch anyway with volume due H1 next year. And this is entirely reliant on DDR5 also being available in volume at a consumer level. Are Intel going to allow mixed memory motherboards, or dictate vendors must choose one or the other? Given the change in CEO recently, are Intel going to continue their convoluted and largely fictitious marketing messages and rush Alder Lake out the door with DDR4 saying "but if you want DDR5 you have to wait for a bit", or just EOL the DDR4 boards after a quarter once DDR5 is more available? Maybe the entire thing will be held off until Microsoft actually sort a half-decent scheduler so Alder Lake's entire design philosophy won't crash and burn.

Depends entirely what you consider a dud, and your perspective.
Gaming crown is very important, even more so than productivity crown, when AMD took the crown with 5-series the prices skyrocketed.
Big little concept might be unproven in x86 but Intel didn't design it on a fag packet, but invested billions in R&D so your point there is kind of mute.
I think that first Alder Lakes will probably be DDR4 based, then it will migrate to DDR5 when Zen 4 actually lands, god knows when that will be..............
as they are still struggling with manufacturing process for Zen 3 and RNDA2, making AMD inherently tied in with TMCS. Intel will not have such restrictions.

It could come out in numbers and hit AMD hard in the fall. Only thing that might save them is Zen3+ refresh.
 
You can only create so many duds before you come up with a straight flush.

https://en.wikipedia.org/wiki/Gambler's_fallacy

Just wait and see then pick whatever best meets your needs, no need to assume anything.

But Intel have only made 1 dud in Rocket Lake. The endless of Skylake rehashes are tired yes, but hardly duds. Moving to 10nm is going to take a huge amount of pressure off power and thermal requirements for sure, but you're still creating a turd if the arch isn't up to scratch.

Intel has been building up to Rocket lake for years, I don't think we should view it as bad luck or that it ensures future releases will be good. Put another way, CPUs don't seem to be the result of luck or happenstance but rather careful planning, if the chip is underwhelming then it's a mistake made years ago.

Hopefully Intel and AMD will have done a good job for their upcoming chips but we don't need to assume anything, just wait and see what they actually produce.
 
Depends entirely what you consider a dud, and your perspective.
Gaming crown is very important, even more so than productivity crown, when AMD took the crown with 5-series the prices skyrocketed.
Big little concept might be unproven in x86 but Intel didn't design it on a fag packet, but invested billions in R&D so your point there is kind of mute.
I think that first Alder Lakes will probably be DDR4 based, then it will migrate to DDR5 when Zen 4 actually lands, god knows when that will be..............
as they are still struggling with manufacturing process for Zen 3 and RNDA2, making AMD inherently tied in with TMCS. Intel will not have such restrictions.

It could come out in numbers and hit AMD hard in the fall. Only thing that might save them is Zen3+ refresh.

In what way is Rocket Lake not a dud? The only area it wins in is AVX512, which nothing uses... Everything else it's slower and hotter than even the previous gen Intel.

It'll be interesting to see how well the 'gaming crown' thing works going forwards, fundamentally it's actually irrelevant as to see the differences you need to be running 1080p, or even lower, with a 3090... Most people don't have a 3090 and most people don't game at <1080p. For marketing sure, for sales, who knows.

Just because a company spent millions/billions in R&D doesn't make something good, Bulldozer and Rocket Lake are good arguments for that.

AMD aren't exactly struggling with the process, just capacity, and that's largely because of consoles. With Apple dropping to 3nm opening up 5nm for next gen processors/GPUs I don't see it being a problem for Zen3+/Zen4.

If Alder Lake comes out in 'the fall', actually hits the 20% IPC in more than just one specific case *and* can still hit the clock speeds then they'll take the lead... That's a few ifs. I doubt Alder Lake will see any real availability in 2021, I also doubt the 20% IPC claim (it will likely hit that in one or two very specific benchmarks so it can still be claimed), and given Ice Lake results I'm not sure they'll hit the clock speeds on 10nm they can on 14nm (with all the +'s)

Who knows what Zen3+ will be and when it will come out, if it's 'just' Zen3 tweaked on 6nm or even 5nm that could well be enough to be in the same ballpark as Alder Lake, and quite possibly out earlier as well.
 
In what way is Rocket Lake not a dud? The only area it wins in is AVX512, which nothing uses... Everything else it's slower and hotter than even the previous gen Intel.

It'll be interesting to see how well the 'gaming crown' thing works going forwards, fundamentally it's actually irrelevant as to see the differences you need to be running 1080p, or even lower, with a 3090... Most people don't have a 3090 and most people don't game at <1080p. For marketing sure, for sales, who knows.

Just because a company spent millions/billions in R&D doesn't make something good, Bulldozer and Rocket Lake are good arguments for that.

AMD aren't exactly struggling with the process, just capacity, and that's largely because of consoles. With Apple dropping to 3nm opening up 5nm for next gen processors/GPUs I don't see it being a problem for Zen3+/Zen4.

If Alder Lake comes out in 'the fall', actually hits the 20% IPC in more than just one specific case *and* can still hit the clock speeds then they'll take the lead... That's a few ifs. I doubt Alder Lake will see any real availability in 2021, I also doubt the 20% IPC claim (it will likely hit that in one or two very specific benchmarks so it can still be claimed), and given Ice Lake results I'm not sure they'll hit the clock speeds on 10nm they can on 14nm (with all the +'s)

Who knows what Zen3+ will be and when it will come out, if it's 'just' Zen3 tweaked on 6nm or even 5nm that could well be enough to be in the same ballpark as Alder Lake, and quite possibly out earlier as well.


In Intel's recent slides they talked about the HEDT Alder Lake models having an all core clock of 4ghz, so that will give us some indication of where the clocks will land
 
In Intel's recent slides they talked about the HEDT Alder Lake models having an all core clock of 4ghz, so that will give us some indication of where the clocks will land

Ouch!

If Rocket Lake (11900K) is potentially capable of boosting to 5.1GHz all-core surely that means Alder Lake needs a 25% IPC across the board just to compete?

Although if that's equivalent of 'TB2', so 4.7GHz it seems for 11900K, 20% would put it fractionally ahead (17.5% would match).
 
In Intel's recent slides they talked about the HEDT Alder Lake models having an all core clock of 4ghz, so that will give us some indication of where the clocks will land
Intel's HEDT has never clocked as high as the desktop parts though, excluding the insane desperate stunts like the W-3175X.

I doubt we'll see 10nm SuperFET push past 5GHz though...
 
Big little concept might be unproven in x86 but Intel didn't design it on a fag packet, but invested billions in R&D so your point there is kind of mute.
Surely Intel are one of biggest examples of why throwing billions at things doesn't necessarily work.
Now most of the really bad examples are outside their core competence (which is basically: milking Intel getting very luck with IBM's decision to use the 8086 for their PC), but the all these 'failures' cost millions or billions:
  1. McAfee
  2. 5G modems
  3. Atom for tablets panic mode
  4. Larabee
  5. Itanium
So throwing money at problems might not work.
 
Last edited:
Whoever gets DDR5 into circulation first will sell through the roof. Certainly wish Intel well, if they deliver a world-beating product I'll gladly shell out for it.

Arm proved that mixing core specs can work, hopefully Microsoft can optimise Windows to understand it.
 
Intel's HEDT has never clocked as high as the desktop parts though, excluding the insane desperate stunts like the W-3175X.

I doubt we'll see 10nm SuperFET push past 5GHz though...

That's not really true. Nehalem/X58 clocked through the roof. The 920, 975x and 990x were all very good clockers and blew the desktop sku's out of the water.
 
Last edited:
Whoever gets DDR5 into circulation first will sell through the roof. Certainly wish Intel well, if they deliver a world-beating product I'll gladly shell out for it.

Arm proved that mixing core specs can work, hopefully Microsoft can optimise Windows to understand it.
So DDR5 by itself is a reason to buy in? And presumably pay a lot more the actual DDR5 memory compared to DDR4 memory prices in the process.
So you think that in early 2022 a poor implementation of DDR5 in a slow product would outsell good implementing of DDR4 in a faster product even if DDR5 was twice the price of DDR4?
Well, 5 is a higher number than 4, but why?
 
So DDR5 by itself is a reason to buy in?

I think people will at least value the support for it for sure. It may even be possible on the earlier chipsets to continue running DDR4 while you decide. That will give anyone who offers that capability a significant advantage for the longevity of the platform.

And presumably pay a lot more the actual DDR5 memory compared to DDR4 memory prices in the process.

You can presume all you want... There is also the argument that as transition is made stocks of high end DDR4 will deplete leaving it as either more expensive or lower end than we know it today.

So you think that in early 2022 a poor implementation of DDR5 in a slow product would outsell good implementing of DDR4 in a faster product even if DDR5 was twice the price of DDR4?
Well, 5 is a higher number than 4, but why?

DDR5 will have a number of advantages over DDR4 besides the speed, such as built-in voltage regulation. It will also be innately registered memory with some ECC capabilities, even on Desktop. It has many advantages beyond the speed, or in your case, the obsession with it being a "one up".
 
Back
Top Bottom