Intel Core i9-13900K

[birdie]

13900K and 7950X at 125W and 65W:

Intel Core i9-13900K vs. AMD Ryzen 9 7950X at 125W and 65W | Club386

Prefer to keep temperature and power consumption down to lower levels this winter? Here's what happens when the best CPUs are scaled back.

www.club386.com

Looks like both CPUs are needlessly overclocked out of the box.

 

[Dirt Chip]

The race for the highest wattage consumption is on.
For a sec it seems AMD reached an equal-point, but along came 13900k and just smoke everyone (and everything) around.

 

[Denver]

RL manages to be even hotter than Zen4. I know winter is coming, but this is a very appealing sales tactic. Just kidding

 

[phill]

Amazing review there @W1zzard, above and beyond as always mate, amazing work and effort.

I've gotta say, reading that quickly, its not really earth shattering and massively in front of AMD for me as I was expecting it to be. Its a little faster in games lower res's but if you have these CPUs and a GPU to match, you won't be gaming at 720P I doubt...

Wow... Power and temps, jesus......

 

[W1zzard]

Yes, the simplest way is tweaking the DC Loadline until VID (a calculated value) and vcore (actually measured CPU voltage) match as close as possible in a multithreaded workload, which can take some time. If not for more or less correct CPU chip-only power measurements, at least this would ensure that power limits are actually working as intended.

That gives you a match for voltage, and only if you measure vcore with a DMM. What about current? So you can do U*I = P

 

[Pumper]

Something is not right with the Cinebench results. All other reviews show 38k~40k. Wonder how affected the other tests are by whatever is causing it in your setup.

 

[Toss]

Efficiency scaling 13900K vs 7950X

View attachment 266328

125W is sweet spot

 

[JustBenching]

Something is not right with the Cinebench results. All other reviews show 38k~40k. Wonder how affected the other tests are by whatever is causing it in your setup.

Yeah, was the same with the 12900k, their power limited numbers were waaay off.

 

[HTC]

I wonder: will those that slammed AMD and their new generation due to high temps now purchase this CPU because of it's performance, with EVEN HIGHER temps?

Dunno about the other CPUs of this family: will they too have such temps?

 

[oxrufiioxo]

I wonder: will those that slammed AMD and their new generation due to high temps now purchase this CPU because of it's performance, with EVEN HIGHER temps?

Dunno about the other CPUs of this family: will they too have such temps?

Pretty sure the 13600k will be just fine and maybe the 13700k. This cpu is pushed beyond reason to compete with the 7950X.

 

[W1zzard]

Something is not right with the Cinebench results. All other reviews show 38k~40k. Wonder how affected the other tests are by whatever is causing it in your setup.

I noticed that too .. I get 40,000 if I disable power limit, i.e. enable ASUS MCE. That's not "stock" though

 

[JustBenching]

I noticed that too .. I get 40,000 if I disable power limit, i.e. enable ASUS MCE. That's not "stock" though

Εveryone else scores 38 to 39k at 253w stock though.

 

[W1zzard]

Εveryone else scores 38 to 39k at 253w stock though.

Is that with Adaptive Boost on or off ? On Windows 11 with VBS?

 

[Solid State Brain]

That gives you a match for voltage, and only if you measure vcore with a DMM. What about current? So you can do U*I = P

There's no straightforward way that I am aware of for making sure that the internal current value used by the CPU for calculating power ("Package Power") is truly accurate. Since among other things it is used for internal CPU limits (e.g. IccMax / peak current limit), it could be assumed to be factory-calibrated. When current readings are present from motherboard sensors (e.g. "VR IOUT"), they are usually from the digital VRM controller on the motherboard, not the CPU itself.

Usually just making sure that the DC Loadline has been correctly configured is enough for most practical purposes. In principle, the DC loadline is supposed to be configured to the impedance/slope of the VRM loadline so that the CPU knows how much input voltage will drop with input current. It will vary with the configured LLC setting and occasionally CPU type/class.

 

[W1zzard]

is enough for most practical purposes

For actual usage, definitely .. even for a review where a few percent difference in measurement can be make or break? Not sure

 

[OfficerTux]

I'll wait for the 5800X3D + RTX4090 re-bench. If that also shows the 5800X3D is within striking distance I will just upgrade my AM4 system for the third time, get a mid to high-range graphics card from either team green or red as soon as those are out and be fine for the next years. Does not look like new Motherboard + DDR5 RAM is worth it at the moment.

 

[HTC]

Question for @W1zzard

For Zen 4 CPUs, AMD claims to the highest temp these CPUs can achieve before the PC shuts down is 115º, which is why 95º is a safe temp.

This CPU went as high as 117º: how high does Intel claim it to be able to reach before shutting down due to temps?

 

[W1zzard]

This CPU went as high as 117º: how high does Intel claim it to be able to reach before shutting down due to temps?

it's throttled at 115° + 2°C from normalizing from delta T to 25°C room temp

 

[JustBenching]

Is that with Adaptive Boost on or off ? On Windows 11 with VBS?

Guru3d uses windows 11. Don't know about adaptive boost but it shouldn't make a difference, if they are scoring 38-39k @ 253w.

 

[HTC]

it's throttled at 115° + 2°C from normalizing from delta T to 25°C room temp

In that case, the Intel CPU has "an unfair advantage" since it can boost to it's absolute max temp while the same can't be done with the 7950X AFAIK due to the max temp being 95º in BIOS: as i understand it, one can lower them max temp in BIOS but NOT raise it, though they might change this in future BIOS updates.

 

[TheinsanegamerN]

In that case, the Intel CPU has "an unfair advantage" since it can boost to it's absolute max temp while the same can't be done with the 7950X AFAIK due to the max temp being 95º in BIOS: as i understand it, one can lower them max temp in BIOS but NOT raise it, though they might change this in future BIOS updates.

Doubt it's sustainable, techspot had it throttle at 90c

 

[Xuper]

so 95'c is normal on intel ? just like Zen4 ?

 

[HTC]

Doubt it's sustainable, techspot had it throttle at 90c

Strange: TPU made a CPU cooler experiment on the 7950X using an AIO, the cooler of this review @ different fan speeds and a spiral wraith cooler also @ various fan speeds and, when they reached 95º (mostly in rendering) they all throttled WHILE KEEPING 95º.

 

[TheinsanegamerN]

Strange: TPU made a CPU cooler experiment on the 7950X using an AIO, the cooler of this review @ different fan speeds and a spiral wraith cooler also @ various fan speeds and, when they reached 95º (mostly in rendering) they all throttled WHILE KEEPING 95º.

The 13900k, not the 7950x

 

[HTC]

The 13900k, not the 7950x

Ahhhh: i misunderstood.