Intel Core i9-13900KS Posts 5% Single-Thread Performance Uplift Over i9-13900K

[btarunr]

Intel's upcoming flagship Core i9-13900KS "Raptor Lake" processor posts a 5% single-threaded performance lead over the company's current top-chip, the i9-13900K. The i9-13900KS is shown in leaked benchmarks scoring 2366 points in the Cinebench R23 single-thread, compared to 2243 points of the i9-13900K, 1948 points of the upcoming i9-13900 (non-K), and 2057 points of the AMD flagship Ryzen 9 7950X. This works out to a 5.43% gain for the i9-13900KS in comparison to the i9-13900K, and an impressive 15% gain over the 7950X. The multi-threaded score is 3.3% higher between the i9-13900KS and the i9-13900K.

The Core i9-13900KS wins on account of higher frequency, with its P-cores boosting up to 6.00 GHz, compared to 5.80 GHz of its predecessor, and its 150 W processor base power providing better boost state residency. The locked Core i9-13900 is held back by lower clock speeds of up to 5.60 GHz P-core boost, and power limits of just 65 W PBP and 221 W MTP. What's interesting in these numbers, though, is that even the i9-13900 is shown beating the 7950X in the multi-threaded test, scoring about 1.6% higher. Its single-threaded performance is marginally lower than the 7950X. Intel is expected to announce the i9-13900KS alongside locked 65 W 13th Gen Core processors, and 13th Gen Core mobile processors, in its 2023 International CES Keynote address, with product availability expected later in the month.



View at TechPowerUp Main Site | Source

 

[Nephilim666]

How many additional watts for that 1.6% over the 7950X?

 

[Dirt Chip]

How many additional watts for that 1.6% over the 7950X?

13900(non k) Actually use a bit less, as it's a 221w vs the 230w of the 7950x.

And regarding the ks, just about any 13900k can do 2350 in single core.

 

[Crackong]

Should be easily a 350W CPU

 

[stimpy88]

WOW! I must order 5.

 

[evernessince]

13900(non k) Actually use a bit less, as it's a 221w vs the 230w of the 7950x.

And regarding the ks, just about any 13900k can do 2350 in single core.

Remains to be seen in reviews. What we do know is that the 13900K needs higher power to scale as you see decent performance benefits all the way to 305w. The 7950X on the other hand can get identical to stock performance at 165w. When both are limited to 165w the 13900K is a significant 32% slower. From HWUB:

By extension I'd expect the efficiency of AMD's non-x CPUs to vastly exceed that of Intel's non-K CPUs. That or you can just power limit your 7950X to get vastly better efficiency. AMD did not even bother binning the 7950X like it did the 5950X because they knew Intel's new CPUs would be power hungry. They will instead use those chips in server at higher margins.

 

[Dirt Chip]

Remains to be seen in reviews. What we do know is that the 13900K needs higher power to scale as you see decent performance benefits all the way to 305w. The 7950X on the other hand can get identical to stock performance at 165w. When both are limited to 165w the 13900K is a significant 32% slower. From HWUB:

View attachment 275216


By extension I'd expect the efficiency of AMD's non-x CPUs to vastly exceed that of Intel's non-K CPUs. That or you can just power limit your 7950X to get vastly better efficiency. AMD did not even bother binning the 7950X like it did the 5950X because they knew Intel's new CPUs would be power hungry. They will instead use those chips in server at higher margins.

Yep, RL is a dirty power hog. But at raw preformance it deliver as expected out of top end CPU.
If you need, you can also tune it to get more efficiency as do 7950x can

 

[JustBenching]

Remains to be seen in reviews. What we do know is that the 13900K needs higher power to scale as you see decent performance benefits all the way to 305w. The 7950X on the other hand can get identical to stock performance at 165w. When both are limited to 165w the 13900K is a significant 32% slower. From HWUB:

View attachment 275216


By extension I'd expect the efficiency of AMD's non-x CPUs to vastly exceed that of Intel's non-K CPUs. That or you can just power limit your 7950X to get vastly better efficiency. AMD did not even bother binning the 7950X like it did the 5950X because they knew Intel's new CPUs would be power hungry. They will instead use those chips in server at higher margins.

That graph is wrong, or to put it better, this graph is so obviously wrong I don't understand how people can't instantly figure it out just by looking at it. Steve (the reviewer) said so in his twitter, there was something wonky with his motherboard. Most likely he is using fixed voltage for his power limited numbers.

The actual difference between 7950x and 13900k in efficiency at that specific task (cbr23) is between 5 and 12%, depending on the wattage. Yes the 7950x is more efficient, but marginally so

Here is an actual stock 13900k with 125w power limit, no UV or anything, everything is left on auto

 

[Kapone33]

Excuse did you see my 5%? I thought I left it in the previous word.

 

[Neo_Morpheus]

Are the oven mittens included?

Joke aside, i would wait fir real reviews, since intel has zero problems in lying about their numbers.

 

[Deleted member 185088]

Are the oven mittens included?

Joke aside, i would wait fir real reviews, since intel has zero problems in lying about their numbers.

Now it seems everyone lies, even the so called reviewers.
I would say wait for results from reputable reviewers.

 

[Bomby569]

reviewers should re-check their results, i only trust GN's Steve

 

[Fasola]

Remains to be seen in reviews. What we do know is that the 13900K needs higher power to scale as you see decent performance benefits all the way to 305w. The 7950X on the other hand can get identical to stock performance at 165w. When both are limited to 165w the 13900K is a significant 32% slower. From HWUB:

View attachment 275216


By extension I'd expect the efficiency of AMD's non-x CPUs to vastly exceed that of Intel's non-K CPUs. That or you can just power limit your 7950X to get vastly better efficiency. AMD did not even bother binning the 7950X like it did the 5950X because they knew Intel's new CPUs would be power hungry. They will instead use those chips in server at higher margins.

https://twitter.com/x/status/1583204564492115972

 

[lightning70]

I want to see the power consumption

 

[Darksword]

I want to see the power consumption

Don't worry, you'll be able to feel the heat output from wherever you're currently physically located.

 

[Dirt Chip]

Are the oven mittens included?

After ZEN4 temps, those are everywhere.

 

[lightning70]

That graph is wrong, or to put it better, this graph is so obviously wrong I don't understand how people can't instantly figure it out just by looking at it. Steve (the reviewer) said so in his twitter, there was something wonky with his motherboard. Most likely he is using fixed voltage for his power limited numbers.

The actual difference between 7950x and 13900k in efficiency at that specific task (cbr23) is between 5 and 12%, depending on the wattage. Yes the 7950x is more efficient, but marginally so

Here is an actual stock 13900k with 125w power limit, no UV or anything, everything is left on auto

View attachment 275224

How do we see this screen?

 

[Why_Me]

der8auer: Testing and Tuning the new 13900K for Efficiency

 

[Godrilla]

Just in case if you want i9 13900ks performance today

 

[chrcoluk]

der8auer: Testing and Tuning the new 13900K for Efficiency

Shame with almost every reviewer, they dont show enough comparison data, they compare to say 7950X but no one is going from a 7950X to 13900K but dont show older gen cpus going back 4-5 years.

Really curious how old and dated my 9900k is now but hard to find it comparison to 13900k.

 

[Veseleil]

Shame with almost every reviewer, they dont show enough comparison data, they compare to say 7950X but no one is going from a 7950X to 13900K but dont show older gen cpus going back 4-5 years.

Really curious how old and dated my 9900k is now but hard to find it comparison to 13900k.

It's almost as they omit those on purpose.

 

[ThrashZone]

Hi,
Yeah it's funny 16 threads is not worth mention but still has a lot of life left.
13 series is obviously an improvement but this is just sells hype and remonds me of nvidia saying just get rt now surely they aren't bias

 

[Dirt Chip]

Shame with almost every reviewer, they dont show enough comparison data, they compare to say 7950X but no one is going from a 7950X to 13900K but dont show older gen cpus going back 4-5 years.

Really curious how old and dated my 9900k is now but hard to find it comparison to 13900k.

Here you go- 6th gen up to 13th.
It`s content creation orianted, but give a good glance at the general advencment.

 

[evernessince]

That graph is wrong, or to put it better, this graph is so obviously wrong I don't understand how people can't instantly figure it out just by looking at it. Steve (the reviewer) said so in his twitter, there was something wonky with his motherboard. Most likely he is using fixed voltage for his power limited numbers.

The actual difference between 7950x and 13900k in efficiency at that specific task (cbr23) is between 5 and 12%, depending on the wattage. Yes the 7950x is more efficient, but marginally so

Here is an actual stock 13900k with 125w power limit, no UV or anything, everything is left on auto

View attachment 275224

Interesting, is there another review I can look at for 13900K power scaling? Google doesn't seem to turn up anything.

 

[Neo_Morpheus]

Now it seems everyone lies, even the so called reviewers.
I would say wait for results from reputable reviewers.

You are correct, because some of these reviewers are simply not trustworthy anymore.