Intel Starts Shipping Core i9-12900KS to Early Customers

[btarunr]

Intel's ambitious new flagship desktop processor, the Core i9-12900KS, started shipping to early customers before retail embargo. By this we don't mean enthusiasts with privileged ties to the company, but retailers. PC enthusiast DAGINATSUKO was able to purchase one online for roughly USD $790, before the store they purchased from took down the listing.

Pictures of the retail i9-12900KS reveal a darker-themed box than that of the i9-12900K, with "Special Edition" written on the front-face. The chip features an S-Spec code "SRLDD." We also learn a few interesting tidbits about the i9-12900KS from this source. Apparently, its processor base power (PBP) is set at 150 W, and its maximum turbo power (MTP) at 260 W. The standard i9-12900K comes with 125 W PBP and 241 W MTP. The increased power limits support a more aggressive boosting algorithm, with the maximum Turbo Boost clocks on the P-cores set at 5.50 GHz, compared to 5.20 GHz on the i9-12900K. The P-core base frequency is increased by 200 MHz, too, now at 3.40 GHz.



View at TechPowerUp Main Site | Source

 

[DeathtoGnomes]

Intel Starts Shipping Core i9-12900KS to Early Guinea Pigs.​

FTFY.

 

[Mussels]

AMD: Let's try adding 3D cache and experiment with new designs

Intel: ADD. MORE. WATTAGE.

 

[Dr_b_]

Its not worth $800. $499 maybe. but not 800. It uses too much power, will cost a lot to keep running over time, and everyone has high utility bills right about now.

 

[InVasMani]

What if Intel were to do something like 3C big + 12C LITTLE chip with using the 4 chip dies to segment 50/50 between big/LITTLE for EDRAM/HBM or use 2 chips for that purpose and other remaining 2 chips to further beef up the integrated graphics further. It seems like that would've been both a good mobile chip and a interesting value chip for desktop. You could even alternate chip dies even/odd between big little on each side then do the opposite on the other inverted and mirrored and rotated much strikingly similar to what Apple did with it's recent switched together chip.

The plus side to that is you'd have memory chips top and bottom so you access them more easily and quickly where the further CPU chips otherwise would have more of a latency penalty incurred. There is still latency involved, but the round trip latency would peak quite as severally potentially. The other plus is heat distribution would be kind of ideal using the efficiency L chips as a buffer between P cores.

Its not worth $800. $499 maybe. but not 800. It uses too much power, will cost a lot to keep running over time, and everyone has high utility bills right about now.

That's just mining helping against fiat currency allegedly.

 

[R-T-B]

AMD: Let's try adding 3D cache and experiment with new designs

Intel: ADD. MORE. WATTAGE.

Honestly, BIG.little on the desktop is a pretty big experiment in it's own right. Not one I particularly like, but yeah.

That's just mining helping against fiat currency allegedly.

Unsure what mining has to do with this at all? No one mines seriously with CPUs, really.

 

[Mussels]

Honestly, BIG.little on the desktop is a pretty big experiment in it's own right. Not one I particularly like, but yeah.


Unsure what mining has to do with this at all? No one mines seriously with CPUs, really.

I'll agree on the big.little, actually

I still think they need to go all the way and mix arm64 with x86, letting the OS run off super low wattage parts (or run true 15W TDP x86 parts and lock the OS to them)

 

[R-T-B]

I still think they need to go all the way and mix arm64 with x86, letting the OS run off super low wattage parts (or run true 15W TDP x86 parts and lock the OS to them)

Thanks, I hate it.

 

[Mussels]

Thanks, I hate it.

same, but the OS having the efficiency of a chromebook or ipad with gaming/demanding apps using the bigger cores as needed? That's the long term future we know is coming.

 

[Melvis]

Good for winter if you dont have a heater, should set things on fire pretty quick

 

[JustBenching]

Good for winter if you dont have a heater, should set things on fire pretty quick

Heard the same about the original 12900k, but law and behold, its extremely efficient. It even beats the m1 pro in cinebench r23 in efficiency

 

[mama]

Heat. Trouble.

 

[Tomorrow]

Heard the same about the original 12900k, but law and behold, its extremely efficient. It even beats the m1 pro in cinebench r23 in efficiency

Still runs hot. Cant imagine what kind custom loop this KS version would require.

 

[JustBenching]

Still runs hot. Cant imagine what kind custom loop this KS version would require.

No it doesnt. Im maxing out at 76 running cinebench r23 on a small single tower cooler

 

[Tomorrow]

No it doesnt. Im maxing out at 76 running cinebench r23 on a small single tower cooler

Yes it does. Instant near 100c when running r23. Temps from around 3h15min mark:

And this was using 360mm AIO. Tho admittedly they were in a hot room and the fans were running at low rpm/not tuned in bios.

 

[JustBenching]

Yes it does. Instant near 100c when running r23. Temps from around 3h15min mark:

And this was using 360mm AIO. Tho admittedly they were in a hot room and the fans were running at low rpm/not tuned in bios.

I dont need a review to tell me how my cpu performs. I have it you know...

 

[Tomorrow]

I dont need a review to tell me how my cpu performs. I have it you know...

So you base your knowledge on your one sample and apply it to others. Seems "logical".

 

[InVasMani]

Still runs hot. Cant imagine what kind custom loop this KS version would require.

Hot yes, but don't confuse heat concentration with inefficiency at the same time. The E cores get more work done for multi thread workloads for the die space occupied, but with higher heat concentration because it's doing a lot of added work relative to space occupied. The E cores need more multipliers to become more granular if they wish to improve performance and efficiency balancing of them especially consider speed step would readily see more upsides from those E cores in doing so.

That's part of why the 12900K is more efficient than the 12700K and the 12600K is less efficient than the 12700K. It's attributed in part due to the additional multipliers and speed shift along with the additional caches. They do add more heat, but that's expected.

Where the P cores excel is single thread performance, but they definitely don't fare as well on multi-thread performance for the die space they take up. The E cores were assumed to help with heat in part to muster and extract more single thread performance from the P cores aiding them by offloading some performance tasks to the E cores. I'd say in practice they don't do that as convincingly as they probably could and should right yet, but if they subdivide them into more multipliers on the E cores that could very quickly and dynamically provide a nice granularity change for the better.

Intel needs to better work around the design limitations with it's follow up. They could absolutely stand to learn from what hardware and software makers did with the SID chip and with NES sound chip working around those limitations in clever ways. That's the type of ingenuity that's needed. Find solutions and work around angles that makes them still overall compelling as a whole.

I could get lost just tinkering with the 129000K chip seeing how it works and the number of ways to extract performance and efficient out of it. It's probably much less simple and straight forward than people might presume because there are implications depending on usage and favoring pushing the E cores or P cores to a higher clock frequency and even mixing a bit of scaling between each of them linearly within the flaws of both designs and temperature limits of the chip as a whole. It's complex beast once you realize there are multipliers and frequencies for both, but also BCLK plays another role.

Speaking of BCLK it would be nice if it could get to the point where invidiaul cores/cluster have their own BCLK domain that can be adjust to push BCLK higher or lower and adjust the memory divider of them higher or lower dynamically. The implications on heat and power are fairly relevant especially if you could turn clusters on/off and switch to a lowest power cluster then speed shift back to a higher power one when and where needed.

 

[JustBenching]

So you base your knowledge on your one sample and apply it to others. Seems "logical".

And that is what you are doing as well?

 

[Deleted member 24505]

I'll enjoy my 12700k "heater" over winter, though dual rad loop negates its ferocious heat output a tad.

 

[NDown]

So you base your knowledge on your one sample and apply it to others. Seems "logical".

Aint this what you did aswell though? So, everyone that has the chip will experience what X reviewer told you ? XD

 

[R-T-B]

So you base your knowledge on your one sample and apply it to others. Seems "logical".

Same as you applying one review (thus far) to others yes.

I have no dog in this race just thought I'd point that out.

 

[TheinsanegamerN]

Its not worth $800. $499 maybe. but not 800. It uses too much power, will cost a lot to keep running over time, and everyone has high utility bills right about now.

If you can't afford the $2-3 difference over an entire year of running this in a gaming PC VS a 12700k, you cant afford a 12900ks in the first place.

Same as you applying one review (thus far) to others yes.

I have no dog in this race just thought I'd point that out.

True, however, there are multiple reviews from multiple sites that show the 12900k as being a hot headed little chip. It's not one annecdotal account saying "dude trust me".

 

[ThrashZone]

Hi,
Yep better hurry and release it for 800.us before amd releases it's 3d version for 450.us that is suppose to match 12900k performance

No it doesnt. Im maxing out at 76 running cinebench r23 on a small single tower cooler

Hi,
Hard to take you seriously seeing you haven't even bothered to fill in your system spec's which would show under your avatar

Whether or not you've subbed any benchmarks "I sure haven't seen any so far" sadly the R20-R23 boards aren't updated very often or at all
But feel free to sub on my realbench leader board with clocks/ temp min-max showing using hwinfo64 open love to see some low temps

Leader Board Show your RealBench score

Hi, Surprised there isn't a realbench thread not going to add a board so post if you want too. 9940x 4k memory 4.9 1.3v Refer to system spec's for setup this was done with the new Optimus Signature CPU water block User name and total system score in text please ThrashZone---265.365 User name...

www.techpowerup.com

 

[Tomorrow]

Hot yes, but don't confuse heat concentration with inefficiency at the same time.

Im not. In fact nowhere in this thread did i say it's inefficient. No doubt it's the fastest in most workloads. Tho based on TPU's review it's also clear that at max 241W it's power efficiency specifically is in the middle of the pack and only with lower power limits it starts to shine:

While the default 241/241 configuration is less efficient than all Zen 3 CPUs, the Ryzen 7 5800X is beat as soon as you go below the 200 W limit. At 190/190, the 5600X can no longer keep up. The most energy-efficient configuration turns out to be 75 W, which would make the Core i9-12900K the second most efficient CPU in our test group, only beaten by the Ryzen 9 5950X. Of course, such low limits will drastically reduce performance—you're trading longer runtime for lower overall power usage. The sweet spot is near 125 W, I'd say, but it also depends on the application.

https://www.techpowerup.com/review/intel-core-i9-12900k-alder-lake-tested-at-various-power-limits.html

And that is what you are doing as well?

Aint this what you did aswell though? So, everyone that has the chip will experience what X reviewer told you ? XD

Same as you applying one review (thus far) to others yes.

I have no dog in this race just thought I'd point that out.

Intel Core i9-12900K Review - Fighting for the Performance Crown

The Intel Core i9-12900K is Intel's flagship processor for the Alder Lake architecture. In our testing, we saw fantastic gaming performance from this new processor. Not only low-threaded tests have improved, the 12900K can even beat AMD at highly threaded workloads.

www.techpowerup.com

• Very high heat output / power usage
• Energy efficiency worse than AMD Zen 3

All this energy has to go somewhere, and that's the job of the CPU cooler. We've been using a Noctua NH-U12S for all our CPU reviews, and it almost feels like it's too weak for the Core i9-12900K at stock (!!). When fully loaded, temperatures quickly climb into the 90°C range and above. Given a heat output of 240 W, this isn't surprising at all. AMD's Ryzens, on the other hand, run at much lower temperatures using the same cooler. We also did some testing with an Arctic Liquid Freezer II 360 mm, which could handle the heat output with ease and achieved much better temperatures.

If i could be bothered i could easily find other reviews that say it runs hot but i decided to link TPU's own as this is where we are. So no im not basing this on one example. If your 12900K runs cooler than in the reviews then great but also remember that R23 is a short workload so by nature it wont get as hot a longer renders or games that don't stress all cores that run cooler.
Also im not seeing people across forums etc saying how cool their 12900K runs like it's a common thing. Rather opposite in fact.