Friday, July 3rd 2026
Intel Readies Two 22-Core "Nova Lake-S" SKUs with 108 MB bLLC Cache for Gaming
According to well-known Intel leaker Jaykihn, Intel is preparing two "Nova Lake-S" CPU models with a 22-core configuration, enhanced by the big Last Level Cache (bLLC). These processors are part of the Core Ultra 5 400S family, and Intel is developing two mid-range configurations, each with 22 cores. Both models will feature 6 "Coyote Cove" P-Cores, 12 "Arctic Wolf" E-Cores, and 4 LPE-Cores. This design includes a single compute die aimed at mid-range gaming rigs. Alongside the 22 cores, Intel plans to add an extra 108 MB of bLLC cache to boost gaming performance, similar to AMD's approach with its Ryzen X3D gaming CPUs, which currently lead in gaming performance. Intel aims to match raw CPU compute power with newer CPU cores and use bLLC to improve latency-sensitive tasks like gaming. The only difference between the two new models is that one is an unlocked 125 W TDP version, part of the "K" overclockable family, while the other is a 65 W base TDP model.
Recently, it has been observed that Intel's upcoming Z9x0 platform for "Nova Lake-S" will support up to 52 cores in a single LGA-1954 socket. Some early rumors suggest a PL2 power draw of 474 W, which is a significant increase over previous designs. This is understandable, considering that the design brings workstation-like core counts into the consumer sector for the first time. Regarding the socket, Intel's LGA-1954 has been pictured with a 2-lever independent loading mechanism (ILM), similar to what Intel's HEDT sector uses. This is typically reserved for larger CPU packages, and with the pin count now nearing 2,000 socket pins, it is a necessary choice for Intel. Importantly, the first "Nova Lake-S" CPU SKUs are scheduled for release in just a few months, likely leading to more rumors as we approach the final launch.
Source:
Jaykihn on X
Recently, it has been observed that Intel's upcoming Z9x0 platform for "Nova Lake-S" will support up to 52 cores in a single LGA-1954 socket. Some early rumors suggest a PL2 power draw of 474 W, which is a significant increase over previous designs. This is understandable, considering that the design brings workstation-like core counts into the consumer sector for the first time. Regarding the socket, Intel's LGA-1954 has been pictured with a 2-lever independent loading mechanism (ILM), similar to what Intel's HEDT sector uses. This is typically reserved for larger CPU packages, and with the pin count now nearing 2,000 socket pins, it is a necessary choice for Intel. Importantly, the first "Nova Lake-S" CPU SKUs are scheduled for release in just a few months, likely leading to more rumors as we approach the final launch.
279 Comments on Intel Readies Two 22-Core "Nova Lake-S" SKUs with 108 MB bLLC Cache for Gaming
You must explain whether this situation of the Nvidia driver running on E-cores actually happened, how to observe it, and what problems it would cause. The driver is located at the underlying level with extremely low performance overhead, so it is even less likely to cause any performance issues. The physical layout of the CPU also has no impact on OS scheduling, unless the program explicitly specifies calling CPU 0, 1, 2, 3, but this process generally does not require program intervention and is determined by the OS.
Some of these arguments about idle power draw hint at some confusion going on.
tpucdn.com/review/amd-ryzen-9-9950x3d2-dual-edition/images/power-idle.png
That chart right there is full system power draw at idle (read the explanation).
Measured at the wall socket (230 V) so it's not the actual power draw of the system but what the PSU is pulling from the wall socket.
The values in the chart must be multiplied by the PSU's efficiency at that load.
The TT GF3 1200 PSU would be in the green zone (85-90%) for the higher values (like 7950X) but it could end up in the orange zone (80-85%) for the lower values (like 14700K).
www.cybenetics.com/photos/models/230/2073_2.jpg
Let's say it's 87% for a 7950X and 83% for the 14700K, then 118 W from the wall means 102.7 W for the AMD system and 86 W from the wall means 71.4 W for the Intel system.
Also keep in mind that the test setup isn't favoring AMD for idle power draw because the mobo is X870E, it has two chipsets compared to one for Z790 and Z890. The power consumption for Z790 and Z890 chipsets is given as 6 W while one PROM21 chipset has 7 W.
So every little bit adds up.
I am honestly expecting idle numbers to be significantly different (lower) using a B850/X870 mobo.
Also memory speed affects idle power draw.
www.techpowerup.com/review/ddr5-memory-performance-scaling-with-amd-zen-5/3.html
Same for Intel (6000 vs 8800):
tpucdn.com/review/intel-core-ultra-7-270k-plus/images/power-idle.png
These charts have CPU only idle numbers:
cdn.mos.cms.futurecdn.net/8gYv2RahVFAe4y3DkNhic3.png.webp
cdn.mos.cms.futurecdn.net/jEAn6tyAEGppU49bTMkgEU.png.webp
From here and here.
/////
Regarding Nova Lake's idle power draw the addition of the 4 LPE cores on the SoC tile hints at optimization. The compute tile gets deactivated thus the potential for low idle power is very high. Intel is definitely playing the bipolar game, where idle is optimized for low power and active state is optimized for maximum performance aka Raptor Lake time! 🥵
With AMD, idle power depends on your RAM speed, IF speed and SoC voltage. Every single AM5 CPU I've had idled at 15-ish W with 4800 MHz RAM and 1 VSoC, and at 25-ish W with 6000 MHz and 1.2 V.
Note that the 7800X3D eats a lot less power than the 9800X3D.
Sure, it's not a massive difference for the whole system, but it can mean something for your temperatures. Now, luckily, Intel is much easier to cool, so in this example, it's pretty much equal to AMD:
Anyway, I don't have any assumptions of the new Intel CPU. I'll see how it does when the review comes out.So another maybe £0.80 a month? That's definitely gonna break the bank (and I don't have AC, but that's besides the point).Either that, or a single monitor. You choose.That's a valid argument if load power is equal. But it's not. Look at the diagrams above.
If a CPU idles at 25 W, and consumes another 20 W at load, it's gonna run more economically than one that idles at 10 W, but consumes 60 W more at load.
AMD isn't dumb and that's the lesser of their concerns right now, but AMD fanboys will still protect them because now it's about "X3D" which doesn't make a difference on laptops the moment you have a QHD display.
Btw, if I ever try to game on battery on my current Ryzen-based laptop, I'm burning through the battery in less than a single hour with a full charge despite CPU "sipping" 20W, dGPU sitting at 10W, playing at 60Hz and lowest brightness. Even the effects of a smaller battery size doesn't explain that phenomenon. I should maybe try that on my i5-7300HQ laptop as well, it's turning 10 next year and it's still functional.
For what I care, I'll buy the best compromise of either Zen6 or Nova Lake bLLC if within my budget, but my priority is sound/global latency (which my 2 Ryzen-based laptops sucks/sucked at, even with the best SODIMM DDR4 kit specs I could get) and core count (because I play games which can use them all, regardless of their type).
I got curious and just tested how my system with one of the most powerful desktop CPUs available - 24 core 270K at stock settings and 5080 at stock settings (no undervolting, frequency limiting) behaves - with one 4K 60 Hz monitor hooked up.
IDLE: Total system power draw from the wall: 60W, and that includes three 4TB SSDs, 6800 MT/s memory (again stock). CPU 12W, GPU 16W.
Youtube video playback at 1080p: system 69W, CPU 15W, GPU 20W.
I would like to see, what a STOCK AMD system with their most powerful desktop CPU draws idle and playing a youtube video.
So for me I established some baseline what a well specced system needs to draw idle and with very light load and anything above it is a waste of energy.
I hope that Intel tunes the new CPUs well for light loads.
I never suggested undervolting, and I never will.See my above example. Arguing about a few W at idle while ignoring power at high load is what we call "penny wise, pound stupid" here in England.
A few watts on idle is nothing, if I'm not doing anything the PC gets turned off, power draw at load is a more valid concern IMO.
When I upgraded from AM4, the 13900KS easily doubled performance over the 5950X, Nova Lake will be that type of upgrade
AMD now selling weak and very expensive 8 core CPUs for long years is in my eyes the same as Intel selling 4 core CPUs in the past. Not very impressive and commendable.
But I'm pretty sure most software just see cores/threads and throw everything they have at them, in which case it's still accelerated but not double the performance.I still prefer having 12+ cores regardless of them being mixed or not, rather than 8 cores with more cache. I don't really benefit from cache despite gaming intensely on some games and I pay less than half the price of X3D for more cores with similar IPC. I'm also already heavily CPU limited in the very few games I play with a 6-core CPU to the point I may just play at 1440p or 4K without even using DLSS or any upscaler on a 3060 Laptop GPU and barely lose any fps due to the CPU limitation being so huge.
It's not about the cost as such, you are acting like we should be throwing cans of soda down the drain cause it's only 30 pounds a year...A 7950x on its own hits ~70w just scrolling youtube comments. Kid you not, hit the scroll wheel on youtube ---> 70 watts. My brother was crying about it and replaced it with a 285k :D
In Canada you really only need AC 3 months out of the year the other 9 months its cold enough.
And where I live
So no, you're not pouring that soda down the drain. You're exchanging it for a chocolate bar.He encountered a few milliseconds of boost while the web page loaded something. Wow, that's unacceptable. I'm sure it increased his temperatures and his electricity bill tremendously. ;)
Not to mention the 7950X is not the CPU you're looking for if you want maximum efficiency.
Or I can have water instead, which is healthier and nearly free, well at least until the AI overlords pollute it all with datacenters. The soda analogy makes no sense, but you're trading soda for throwing food away instead, much more expensive to constantly exchange parts being worried about a few watts on idle.The power consumption from TPU's own testing measured at the wall is minuscule, he spent hundreds more on a platform that won't last as long to save a few cents on the power bill.
www.techpowerup.com/review/amd-ryzen-7-9850x3d/22.html
Treadrippers have an absurd high idle power consumption as well. My 5950X has a 45W idle power; that's completely normal for the dual CCD / 64MB cache.