News Core i9-12900KS with 5.2 GHz All-Core Boost Allegedly Incoming

[Admin]

The new AIDA64 up supports Intel's unannounced Core i9-12900KS Alder Lake processor.

Core i9-12900KS with 5.2 GHz All-Core Boost Allegedly Incoming : Read more

 

[NightHawkRMX]

I can smell the burning from here

 

[tamalero]

I can smell the burning from here

Agree, if the current offerings churn 200+ Watts in stock.
I imagine these will be above the 250Watts, rivaling a high end Threadripper and harder to cool.

 

[alceryes]

@Admin You're missing a couple "S"s when talking about the 12900KS chips.

 

[HyperMatrix]

Guys...come the freak on...what's with all the errors and lack of clarity in your articles...

"The Core i9-12900K will most likely retain the same 8 + 8 configuration as the Core i9-12900K "

"As the "S" suffix implies, the Core i9-12900K is a special-edition SKU, meaning it'll come with higher clock speeds than the vanilla Core i9-12900K. "

"The previous Core i9-9900K boosted to 5 GHz on all cores, 300 MHz higher than the typical all-core boost clock speed for the regular Core i9-9900K. "

"According to VideoCardz's sources, the Core i9-12900KF reportedly features a 5.2 GHz boost clock across all eight Golden Cove cores."

"Of course, it's arguable that you can overclock the Core i9-12900KS to match the Core i9-12900K regarding the all-core boost clock. "

Highlighted Red text is a mistake.

 

[spongiemaster]

Agree, if the current offerings churn 200+ Watts in stock.
I imagine these will be above the 250Watts, rivaling a high end Threadripper and harder to cool.

Depends on how well the chips are binned. I believe it was the 9900KS that was significantly more efficient than the standard 9900K.

 

[Co BIY]

I think this release indicates good news for the process and yields.

Also that later production 12600K /12700K processors might overclock very well even if they don't bin them out as a 12600KS.

 

[jacob249358]

literally just a space heater that happens to destroy cinebench

 

[Alvar "Miles" Udell]

100w higher power consumption to get you 100 more points in benchmarks!

 

[VforV]

Not only Alder Lake does not sell amazing, but with this move intel proves it feels the fire burning under it's *** with Zen3D approaching... #intelScaredofZen3D.

 

[watzupken]

Depends on how well the chips are binned. I believe it was the 9900KS that was significantly more efficient than the standard 9900K.

While I agree this is true, but it is not easy pushing another 300 Mhz all core turbo even with a good bin. If the usual 12900K have power spike to 270W or higher, pushing that from 4.9 to 5.2 is not going to come without a cost in power. Having said that, this chip will likely only interest those who are extreme overclockers or for hardcore enthusiast due to the pricing and also the need for some high end motherboard and cooling solution.

 

[watzupken]

Not only Alder Lake does not sell amazing, but with this move intel proves it feels the fire burning under it's *** with Zen3D approaching... #intelScaredofZen3D.

Alder Lake is not selling well because of multiple reasons. Adding this to the mix will not improve sales. As a matter of fact, they may see better sales with the lower end SKUs. It is just for Intel to try and defend their performance crown from whatever AMD is going to announce/ release in early 2022.

 

[-Fran-]

The "S" stands for "Scared of Zen3D", right? LOL

And I'm all for these better binned chips, as a proud previous owner of an i7 2700K, but this CPU is already pretty good on its own, why is there a need for a "special" edition? Why does Intel feel the need to milk its fanbois even more? Ah, wait, nevermind. There's always a need to milk fanbois more.

Regards.

 

[TerryLaze]

The "S" stands for "Scared of Zen3D", right? LOL

And I'm all for these better binned chips, as a proud previous owner of an i7 2700K, but this CPU is already pretty good on its own, why is there a need for a "special" edition? Why does Intel feel the need to milk its fanbois even more? Ah, wait, nevermind. There's always a need to milk fanbois more.

Regards.

If you have extremely good yield that is better than what you expected then why wouldn't you release an even better version of a cpu?
Why does AMD have an X version of almost any cpu that does add nothing at all except for a higher price?
The answer to both is, because they can. Pecunia non olet. If they'll be sold, they will be made.

(the S stands for suck on this amd/tsmc)

 

[-Fran-]

If you have extremely good yield that is better than what you expected then why wouldn't you release an even better version of a cpu?
Why does AMD have an X version of almost any cpu that does add nothing at all except for a higher price?
The answer to both is, because they can. Pecunia non olet. If they'll be sold, they will be made.

(the S stands for suck on this amd/tsmc)

Curve your expectations, I'd say. No one knows how yields are for Intel. The last KS they put out was the 9900KS, which was like 14nm++++++++ or was it "+++++"? Well, at a maturity level where they didn't really need to anyway.

Much like with the 2600K and 2700K, 3600 and 3600X and 3600XT, this new KS is going to be as dumb as all of them. I mean, the 3800XT that I own, mind you, it's one of the dumbest CPUs of the last few years and this KS is going to battle for that position head on.

Make no mistake this is a dumb desperate move from Intel to steal some of the spotlight from AMD's announcements in Jan. We all know, more or less, what to expect. We know it'll be good and Intel is worried.

Regards.

 

[TerryLaze]

Curve your expectations, I'd say. No one knows how yields are for Intel. The last KS they put out was the 9900KS, which was like 14nm++++++++ or was it "+++++"? Well, at a maturity level where they didn't really need to anyway.

What does this have to do with any bodies expectations? They have to have enough units for them to make a new sku which means that they have a very high yield.

Much like with the 2600K and 2700K, 3600 and 3600X and 3600XT, this new KS is going to be as dumb as all of them. I mean, the 3800XT that I own, mind you, it's one of the dumbest CPUs of the last few years and this KS is going to battle for that position head on.

Dumbness depends on what somebody needs from their cpu, for a render box yes, extremely dumb, for other purposes the best thing ever.

Make no mistake this is a dumb desperate move from Intel to steal some of the spotlight from AMD's announcements in Jan. We all know, more or less, what to expect. We know it'll be good and Intel is worried.

Intel already stole all of the spotlight with the normal lineup, you get "up to" 5950x performance with $200 less, amd will have to drop prices on a new tech that costs them more to make, even if this avg 15% uptake is going to be across the board.

In cine 23 single, 15% more for the 5950x will not even reach the 12900k running at 50w, 1638 will turn into 1883, and we have seen from the last years that lower multithreaded performance isn't keeping anybody from buying intel cpus.

https://www.techpowerup.com/review/...er-lake-tested-at-various-power-limits/2.html

 

[NightHawkRMX]

This just in, a single core doesn't use much power, so a power limit won't effect a single core benchmark.

 

[-Fran-]

What does this have to do with any bodies expectations? They have to have enough units for them to make a new sku which means that they have a very high yield.

Having high or low yields have nothing to do with making a special binned CPU. That's the point. This is Intel's Marketing thinking it's a good idea to make noise with his "KS" SKU out of nowhere.

Dumbness depends on what somebody needs from their cpu, for a render box yes, extremely dumb, for other purposes the best thing ever.

True, but it's still dumb.

Intel already stole all of the spotlight with the normal lineup, you get "up to" 5950x performance with $200 less, amd will have to drop prices on a new tech that costs them more to make, even if this avg 15% uptake is going to be across the board.

In cine 23 single, 15% more for the 5950x will not even reach the 12900k running at 50w, 1638 will turn into 1883, and we have seen from the last years that lower multithreaded performance isn't keeping anybody from buying intel cpus.

https://www.techpowerup.com/review/...er-lake-tested-at-various-power-limits/2.html

I don't know what your point is here, but I'll just say it doesn't matter? I'm sure most 12900K/KF SKUs can reach the same speeds with about the same power usage as the new KS will. If the current 12900K/KF performs well, then the KS is not going to shatter any new records or anything really relevant for the lineup. That's why it's just noise from Intel to deviate the attention from AMD's potetial announcements.

Regards.

 

[TerryLaze]

This just in, a single core doesn't use much power, so a power limit won't effect a single core benchmark.

Still more than 15% faster...

Having high or low yields have nothing to do with making a special binned CPU. That's the point. This is Intel's Marketing thinking it's a good idea to make noise with his "KS" SKU out of nowhere.

What you are talking about is true for a paper launch, it doesn't matter there if they can release enough units because it's only supposed to be shown in benchmarks and not sold.
I very much doubt this is going to be the case for these.

I don't know what your point is here, but I'll just say it doesn't matter? I'm sure most 12900K/KF SKUs can reach the same speeds with about the same power usage as the new KS will.

No they won't, not out of the box, not without a lot of tweaking, a lot of them not at all, no matter what.
Same performance yes, but not at the same power.
And sure, a lot of people won't care but then again it's going to be what, $30-50 bucks more? A lot of people won't mind paying that little bit extra for not having to mess with settings, which is the same reason people buy the x skews from amd.

 

[-Fran-]

Still more than 15% faster...

I still don't understand why it's important to point that out... We know the 12900K/KF and future KS will be faster than Zen3 in ST works. As I said, it's not important here.

What you are talking about is true for a paper launch, it doesn't matter there if they can release enough units because it's only supposed to be shown in benchmarks and not sold.
I very much doubt this is going to be the case for these.

I'll agree there, as to be fair here, whenever Intel releases a "shpeshial" SKU, it always has availability. Not necessarily wide, but good enough to not be smoke and mirrors.

No they won't, not out of the box, not without a lot of tweaking, a lot of them not at all, no matter what.
Same performance yes, but not at the same power.
And sure, a lot of people won't care but then again it's going to be what, $30-50 bucks more? A lot of people won't mind paying that little bit extra for not having to mess with settings, which is the same reason people buy the x skews from amd.

Hm... Looking back, you're a bit wrong. If you go and look into most of the sites that keep track of the average speed each SKU can reach on average (I can't remember the site name!), they are all really close, making it so that it's more to people looking to break records and such.

And I would put that price difference onto something else... Like a better 360 AIO, a windfarm or more hamsters to feed the beast. Heh.

Regards.

 

[TerryLaze]

I still don't understand why it's important to point that out... We know the 12900K/KF and future KS will be faster than Zen3 in ST works. As I said, it's not important here.

What else would intel be so afraid of from the 3dzen? As I already said intel was selling extremely well the last few years without having any multithreaded crown, so it's not that intel is afraid of and single is going to stay in favor of intel.

Hm... Looking back, you're a bit wrong. If you go and look into most of the sites that keep track of the average speed each SKU can reach on average (I can't remember the site name!), they are all really close, making it so that it's more to people looking to break records and such.

it's silicon lottery that sadly closed down due to cpus reaching max performance with mobo default settings.
And if you look at the 9900ks, as the only example we have, you got 100Mhz more at lower Vcore across the board, lower vcore means less heat which means that the higher cost of the cpu can be considered a wash due to the lower cost of the cooling.
Not that anybody would cheap out on cooling but anyway.
https://siliconlottery.com/pages/statistics

9900KF	4.80GHz	4.60GHz	1.275V	100%
9900KF	4.90GHz	4.70GHz	1.287V	Top 92%
9900KF	5.00GHz	4.80GHz	1.300V	Top 37%
9900KF	5.10GHz	4.90GHz	1.312V	Top 8%
9900KS	4.90GHz	4.70GHz	1.225V	100%
9900KS	5.00GHz	4.80GHz	1.250V	Top 95%
9900KS	5.10GHz	4.90GHz	1.287V	Top 28%
9900KS	5.20GHz	5.0GHz	1.325V	Top 4%

 

[tamalero]

What does this have to do with any bodies expectations? They have to have enough units for them to make a new sku which means that they have a very high yield.

Dumbness depends on what somebody needs from their cpu, for a render box yes, extremely dumb, for other purposes the best thing ever.

Intel already stole all of the spotlight with the normal lineup, you get "up to" 5950x performance with $200 less, amd will have to drop prices on a new tech that costs them more to make, even if this avg 15% uptake is going to be across the board.

In cine 23 single, 15% more for the 5950x will not even reach the 12900k running at 50w, 1638 will turn into 1883, and we have seen from the last years that lower multithreaded performance isn't keeping anybody from buying intel cpus.

https://www.techpowerup.com/review/...er-lake-tested-at-various-power-limits/2.html

Nice cherrypicked test.

 

[TerryLaze]

Nice cherrypicked test.

It's the only single threaded test that is being done and somewhat accepted as being of a useful app, if you don't like that one then take whatever result you do like and divide by 32 threads for the 5950x and 24 threads for the i9.

 

[Conahl]

It's the only single threaded test that is being done and somewhat accepted as being of a useful app

thats right, but only because intel is in the lead with this test. when intel isnt in the lead, its a useless app

 

[TerryLaze]

thats right, but only because intel is in the lead with this test. when intel isnt in the lead, its a useless app

Same goes for both camps, for all the time intel was in the lead it was always a shill app by shill devs for the amd crowd.

Bottom line, benchmarks don't have enough single threaded tests, basically only cinebench which is also useless because nobody in their right mind would run it on only one core, but it is what it is, it's the only thing we get.