• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

Intel Expands 14th Gen Core Desktop Processor Series with 65W Mainstream Models

At 65W? This is impressive! Near top tier performance but without all the massive power draw? Very cool!


It's actually 5.8 GHz TVB target according to Ark (which places it at the same level as the i9-13900K), but the noteworthy thing is that the maximum turbo power is actually 219 W. A single P-core can draw 65 W at 6 GHz, trust me, from experience, and this on a 13900KS... probably even worse on a 14900K.

They haven't really improved anything this "generation" against the 13th after all. The tradeoff of all 14th gen scam CPUs is that they boost more aggressively, but that also means they saturate heat faster/hit their power target faster, and thus have a steeper fall off.
 
Sure they have. Like every generational uptick, there have been refinements and improvements. They're just not dramatic.

Sadly, this time around Intel can't even claim to that... the CPUs haven't undergone any optimization whatsoever, even their hardware stepping is the same. There's zero changes in the silicon and no erratum got fixed, there's... just nothing. The entire "gen" is just shipping the same existing CPUs with slightly more aggressive clocks, or the "new" configuration at the i7 level that has 8P+12E (so 3 out of 4 E-core clusters) that was technically possible earlier, just not done to space the i7 and i9 a little further apart.
 
Just checked the review, wow. On single core its almost on par with a much more power hungry 14900k lol. Thats probably more of an indicator than all core benches.

Imagine how good this would be if it had the best binned parts.
 
Just checked the review, wow. On single core its almost on par with a much more power hungry 14900k lol. Thats probably more of an indicator than all core benches.

Imagine how good this would be if it had the best binned parts.
I can’t prove this because I haven’t seen anyone test this, but isn’t it most likely that the 14900k and 14900non-k draw a similar amount of power? There’s no way a single core draws over 65W, so you don’t need to increase the efficiency of the part to maintain single-threaded clocks.
 
I can’t prove this because I haven’t seen anyone test this, but isn’t it most likely that the 14900k and 14900non-k draw a similar amount of power? There’s no way a single core draws over 65W, so you don’t need to increase the efficiency of the part to maintain single-threaded clocks.

As I mentioned, the non-K has a turbo power target of 219 W, this is just a little less than the K's 253 W target. Intel can't magically make a 65 W 24-core processor capable of reliably hitting 5.8 GHz.

You have to remember that despite such high clocks, Intel's still using a 10 nm process, so yes, the power consumption figures get bizarre real fast if you're used to looking at what Ryzen tends to sip. 300 W is not difficult to hit overall with any Raptor Lake i9, they're designed to be heavily power limited even in their KS versions at stock. Add 6 GHz + 10 nm + AVX and you have the deed done. If not as high as 65 W on an individual core, at least 50 you'll hit, guaranteed.
 
I can’t prove this because I haven’t seen anyone test this, but isn’t it most likely that the 14900k and 14900non-k draw a similar amount of power? There’s no way a single core draws over 65W, so you don’t need to increase the efficiency of the part to maintain single-threaded clocks.

That is true, and thats my point, they didnt nerf the performance enough for there to be a major change in single core performance so basically one can buy this chip with a much lower overall power ceiling out of the box whilst maintaining the most important performance metric.

But is is also relevant to multi core loads as well, the data we have is a all core bench, and a single core bench, however there is a in between, a multi core load that doesnt load all cores, and single core performance will be relevant to that as well. Many software I have used e.g. will still utilise multiple cores either via thread sharing or lightly threaded games and they are also within the TDP of this chip, that tells me based on the single core bench that for those games this chip will still perform very well with a much lower than 20% performance differential.
 
That is true, and thats my point, they didnt nerf the performance enough for there to be a major change in single core performance so basically one can buy this chip with a much lower overall power ceiling out of the box whilst maintaining the most important performance metric.

But is is also relevant to multi core loads as well, the data we have is a all core bench, and a single core bench, however there is a in between, a multi core load that doesnt load all cores, and single core performance will be relevant to that as well. Many software I have used e.g. will still utilise multiple cores either via thread sharing or lightly threaded games and they are also within the TDP of this chip, that tells me based on the single core bench that for those games this chip will still perform very well with a much lower than 20% performance differential.

It's only 30 W lower, though... achievable through the reduction of that last mile + a 56sec Tau, you're looking at an eventual 14900T (the 13900T had a 106 W maximum power target, maximum turbo 5.3 GHz with base frequencies being 1.1 GHz P and 800 MHz E core), with the hope that said 14900T has 300 MHz+ higher targets in the same envelope to even get close to that.
 
Back
Top