Intel Core i9-14900K Raptor Lake Tested at Power Limits Down to 35 W

[thunderingroar]

As I was saying for the past few years, the T versions of Intel CPUS are untouchable in terms of efficiency. You can clearly see now the 14900k at 35w easily topping the charts.

Yeah and it performs on the level on R5 3600, so whats your point lol? Buy the most expensive top of the line CPU only to cut its legs off and make it run at 50% speed?

 

[Lew Zealand]

As I was saying for the past few years, the T versions of Intel CPUS are untouchable in terms of efficiency. You can clearly see now the 14900k at 35w easily topping the charts.

Lol the 14900T sounds like a great way to spend $500, to get barely half of the 14900K. Oh whoops, the 14900Ts will be notably less well binned so say byebye to a good amount of that efficiency. It's almost like people don't buy T SKUs except for OEMs who cram them into USFF cases so they can claim "i9" on the front but deliver i7 or i5-level performance.

 

[Dr. Dro]

When my brain begins to hurt like this, I go to my 17 year old home made, 80 Proof, Kalhua. Double the fire power of the store brought brand and made with real dark brown sugar instead of sugar.

Double shot of my stuff on ice and milk and the pain goes away. Yup, it goes away with a smile.

Oh and thank you W1zzard for the article. A nice read overall.

Ain't bottle a wonderful thing.

Yeah and it performs on the level on R5 3600, so whats your point lol? Buy the most expensive top of the line CPU only to cut its legs off and make it run at 50% speed?

Honestly at that strict a power limit, you can probably optimize per use case. I'd shave 7 of the P-cores and three clusters of E-cores, and attempt to maintain frequency.

Probably would still kick for a lot of applications

 

[thunderingroar]

I'd shave 7 of the P-cores and three clusters of E-cores

At that point i wouldnt even buy i9 to begin with

 

[Dr. Dro]

At that point i wouldnt even buy i9 to begin with

It's all about versatility (and hopefully, the higher quality silicon would help). But in general I agree.

Something to keep in mind however is that the T SKUs have a 35 W base power but their turbo target is actually 106 W

 

[Minus Infinity]

Can you set a custom power limit? I was thinking maybe something like 150-180W with an undervolt. Would still make a decent improvement to efficiency and temperatures and performance hit would be minimal.

However, why bother when you can get the much more efficient AMD X3D variants.

 

[Gica]

Good test to remember how impressively efficient the 7800x3D is. It would be interesting to see how well the X3D performs under an even lower TDP.

And how expensive it is. 7800X3D has the price of i7-14700K and the performance of i5-13500 in applications. It is strong in gaming, but there is a problem: not everyone uses the RTX 4090. A really interesting X3D scalability test is the one that shows us how much this processor helps weaker video cards. A test that all the reviewers avoided and I have an idea why they do it. An indirect answer can be found right in this review.

 

[Hardware1906]

And how expensive it is. 7800X3D has the price of i7-14700K and the performance of i5-13500 in applications. It is strong in gaming, but there is a problem: not everyone uses the RTX 4090. A really interesting X3D scalability test is the one that shows us how much this processor helps weaker video cards. A test that all the reviewers avoided and I have an idea why they do it. An indirect answer can be found right in this review.

View attachment 319213

on Newegg, R7 7800X3D is 45$ cheaper than the i7, using DDR5 configuration, the R7 is 10$ cheaper than the i7 14700K

 

[Outback Bronze]

It's all about versatility

This is the main reason why I buy top end CPU's, much more flexibility to do what you want with it as well as better silicon/binning.

I can turn a 13900K in to a 13100 but cannot turn a 13100 into a 13900k. I always have fun manipulating them for much better efficiency.

 

[Denver]

And how expensive it is. 7800X3D has the price of i7-14700K and the performance of i5-13500 in applications. It is strong in gaming, but there is a problem: not everyone uses the RTX 4090. A really interesting X3D scalability test is the one that shows us how much this processor helps weaker video cards. A test that all the reviewers avoided and I have an idea why they do it. An indirect answer can be found right in this review.

View attachment 319213

Very cheap for the performance and efficiency it delivers, the diluted average means you won't see games where the X3D beats the i9 by 30-40% like TW. Plus, People who really need powerful CPUs because "time is money" should go for TR workstation, not desktop CPUs.

What are you talking about? MSRP? Because I see the price of the X3D dropping to almost the price of the i5 13600k often.

 

[Deleted member 185088]

Thanks for the review, it gives me lot of interesting things to try especially when playing at 4k, limiting FPS and lowering power could result in huge power saving.

 

[Enterprise24]

How much mv achieved for 95w undervolt testing.

 

[Ferrum Master]

As I was saying for the past few years, the T versions of Intel CPUS are untouchable in terms of efficiency. You can clearly see now the 14900k at 35w easily topping the charts.

I can agree to it. I got a 11900T from some broken OEM crap PC and put it in my diy NAS PC(I do remote git compiles, so I needed those cores), that had a Pentium before... they both eat ~15W during idle from the wall, It is so universal I can put it a GPU and make into a LAN party PC also, and I did it few times. I was really impressed.

 

[JustBenching]

Yeah and it performs on the level on R5 3600, so whats your point lol? Buy the most expensive top of the line CPU only to cut its legs off and make it run at 50% speed?

It's the fastest CPU at 35w, hence the most efficient. So - what do you mean "whats the point"?

Plus as Ive said, with a little UV it hits 16k cbr23 score

 

[Cruzy]

I’d love to see an extensive undervolting test (with a detailed explanation of all settings) as it was mentioned that this might be a crucial part of the daily driver config.
I understand that binning is a big thing (gamers Nexus did a crazy video on that) so milage will vary, but an extensive article about that would be great. As a great addition, I’d compare the 13900K/S to see how much Intel improved the binning process and if this actually influences the undervolting capabilities of such high end CPUs.

 

[Pumper]

3% gaming performance increase going from 125 to 253W! Desperation in all its glory...

Why is it desperation when Intel does it, but 7950X performing at a 3-4% MT difference and identical in gaming at half the TDP is somehow a ok?

 

[LuxZg]

While not exactly true, these tests also show that a nice 14600 would be possible at true 65W (as it should be) while 14400 would be possible even at 35W (edit: ok maybe 45W). Sure, 14900 has way more silicone to achieve required performance, but still... would be so awesome if they actually worked at those power levels

 

[JustBenching]

Why is it desperation when Intel does it, but 7950X performing at a 3-4% MT difference and identical in gaming at half the TDP is somehow a ok?

The 7950x draws around 66-70% more power than the 7950x 3d for 3% more MT performance. But that's not desperation cause it's AMD

 

[HD64G]

Why is it desperation when Intel does it, but 7950X performing at a 3-4% MT difference and identical in gaming at half the TDP is somehow a ok?

Intel has chosen for many CPU gens now to compete just by increasing power limits vastly while gaining a few % in MT performance while even reducing the core number from gen 10 to 11. AMD did so only after those many years and only with AM5 in order to keep Intel in check. I disagree with both of those companies in choosing to have CPUs consume more than 200W at stock. My posts in Ryzen 7000 reviews relative to that topic speak for themselves. Now, where is your critisism over the even worst behaviour of Intel for many years prior to AMD?

 

[JustBenching]

Intel has chosen for many CPU gens now to compete just by increasing power limits vastly while gaining a few % in MT performance while even reducing the core number from gen 10 to 11. AMD did so only after those many years and only with AM5 in order to keep Intel in check. I disagree with both of those companies in choosing to have CPUs consume more than 200W at stock. My posts in Ryzen 7000 reviews relative to that topic speak for themselves. Now, where is your critisism over the even worst behaviour of Intel for many years prior to AMD?

Criticizing the out of the bot performance / power draw of the K line of SKUS is like criticizing the out of the box brightness or volume of your TV. Doesn't matter, takes 3 seconds to change it.

Intel has 3 different SKUs each with their own power limit. Choose one of them if you don't want to change the out of the box limits of the K line.

 

[Math75]

no like-for-like comparison w/ 13900k(s) at the same power limits? mehh

Hello,

I did an undervolting (stable) of my 13900K at -0,150V. With power limit max at 90W (PL1 and PL2 at 90W) : I have 30k in Cinebench R23.

During all the bench, my cpu stays at 90W load (same in handbrake).

Maybe I will test with 95 and 125 Watts. Just to see

(my first message here )

 

[Cruzy]

Hello,

I did an undervolting (stable) of my 13900K at -0,150V. With power limit max at 90W (PL1 and PL2 at 90W) : I have 30k in Cinebench R23.

During all the bench, my cpu stays at 90W load (same in handbrake).

Maybe I will test with 95 and 125 Watts. Just to see

(my first message here )

I’d be interested in the results! I have a 12900K and would like to fiddle around with it too as it likes to jump to 100C even with a 360 AIO.

 

[LuxZg]

Good test to remember how impressively efficient the 7800x3D is. It would be interesting to see how well the X3D performs under an even lower TDP.

It would be awesome to see something like top Intel vs top AMD at approximately same average power draw (the real one, at the socket). And then same with the more affordable options like 14600K vs 7800X3D. Aiming for total system cost at same total system power draw would be great. Oh well, wishful thinking but it would be a lot if fun seeing mid level options duel it at actual 65W

 

[Teutorix]

The 7950x draws around 66-70% more power than the 7950x 3d for 3% more MT performance. But that's not desperation cause it's AMD

The 7950x3D is an improved version of the 7950x not the other way around. AMD added 3D cache and lowered power consumption to get similar performance that's progress and good.

 

[yeeeeman]

all this work for retesting a 13900ks basically.