Question Anandtech.com article on 13900k and 7950x power scaling

[adamge]

A Lighter Touch: Exploring CPU Power Scaling On Core i9-13900K and Ryzen 9 7950X

www.anandtech.com

It's an interesting article, but IMO severely crippled by the reliance on "configured" power limit rather than measuring power usage. It's even more frustrating when one of the datapoints is measured power usage, so the author has access to this data.

In this scenario where configured power limits are basically hand waving guidelines, measured power usage (specifically, average power usage by the CPU during the benchmark run) is the only useful, meaningful, and concrete way to analyze the performance results relative to the power usage.

 

[igor_kavinski]

The browsing power usage was the most eye opening, as the 7950x looked like it was drawing on average about 120w, while RPL was about 80w.

Good to know that. However, consider two people who do similar stuff with their CPUs, one with 13900K and the other with 7950X. Intel user will save power in browsing and lose a lot of power in multicore. AMD user will lose power in browsing but will save a lot more in multicore heavy workloads. It kinda balances out in the end for the average user. And if the user is constantly running Blender and other heavy duty workloads, they win with the 7950X with a lower power bill. It all depends on the usage scenarios. Every user has to do their research to see which company gives them the best value. It's no longer the case like Zen 2 vs. Comet Lake where AMD seemed a bit weak or Zen 3 vs. Comet Lake where Intel was clearly the loser.

 

[Just Benching]

will save a lot more in multicore heavy workloads.

According to the PC world review just posted, not really. The difference in heavvy MT at 85w is 7%. So the AMD system will save 5 watts per hour in heavy MT workloads, and then use 40w more while browsing. That's nutty.

 

[igor_kavinski]

According to the PC world review just posted, not really. The difference in heavvy MT at 85w is 7%.

Why would the average user use these CPUs at 85W? The average user just wants to get work done. They don't tinker with settings. So assume it will be 230W for AMD and 254W for Intel (likely higher if MCE is on by default in the BIOS).

 

[Just Benching]

Why would the average user use these CPUs at 85W? The average user just wants to get work done. They don't tinker with settings. So assume it will be 230W for AMD and 254W for Intel (likely higher if MCE is on by default in the BIOS).

Because we are talking about a user that cares about efficiency. If they don't then the whole conversation is moot and the 900k wins cause its generally faster, and although it consumes a bucketload of power to do so, if the user doesn't care about efficiency then it doesn't matter does it?

 

[Mopetar]

According to the PC world review just posted, not really. The difference in heavvy MT at 85w is 7%. So the AMD system will save 5 watts per hour in heavy MT workloads, and then use 40w more while browsing. That's nutty.

Sure if you compare the power when both are maxed out Intel looks better. However, if we tested using one of the non-X chips where AMD didn't push the TDP to such pointless extremes the Intel chip is going to get clobbered in efficiency even if it's restricted to more reasonable power levels as well. The charts from the AT article that you've posted illustrate this quite clearly.

The main takeaway is that AMD hits a wall once they get to around 100W. Intel scales beyond that point, but they start to fall off heavily the further they push. I don't know to what extent this is a design choice on AMD's part or a process advantage that Intel has over TSMC, but it's pretty evident that Zen cores are significantly more efficient which is what's allowing AMD to make such significant inroads in the server market. A core count advantage is rather pointless if it's a cooling nightmare.

 

[Hitman928]

I don't think this video has been posted yet, but PC World did a 13900K power consumption review. They tested total power consumption and didn't do performance per watt testing, which is a shame because that would have been more impactful to this debate.

Basically though, they corroborate what @Herald and I have been saying, that in light to moderate CPU usage, RPL has an advantage over Zen 4 while in heavy thread usage, Zen 4 has the advantage. The browsing power usage was the most eye opening, as the 7950x looked like it was drawing on average about 120w, while RPL was about 80w.

There they are measuring total system power which makes it very difficult to compare CPU efficiency as the choice of motherboard can easily swing power usage up or down tens of watts.

 

[igor_kavinski]

Because we are talking about a user that cares about efficiency.

Oh, you must be referring to Kocicak. We have exactly one user like that on this forum.

 

[Hitman928]

I don't think way more is an exaggeration. The techpowerup showed a 52% efficiency delta, and that's while the 13900k was faster. At similar performance testing that difference could easily climb to 100%, meaning that basically the amd part would use double the power for similar performance.

Now sure, you wanna call that insignificant, go ahead, but then don't tell me that the 10-15% efficiency difference in heavy MT workloads is significant because thats double standards

You are not incorrect, but percentages don’t mean much without looking at absolute values and scenario. If the 7950x could sleep with 0.01 W versus 0.1W for the 13900k, no one would care that the 7950x is 900% more efficient in this scenario.

 

[Just Benching]

Sure if you compare the power when both are maxed out Intel looks better. However, if we tested using one of the non-X chips where AMD didn't push the TDP to such pointless extremes the Intel chip is going to get clobbered in efficiency even if it's restricted to more reasonable power levels as well. The charts from the AT article that you've posted illustrate this quite clearly.

What? First of all there is no non X 7950. Secondly, even if there was, it wouldn't perform any better than a 7950x with 65w eco mode. And we can clearly see from the pcworld review that an eco65w 7950x consumes 85w, so setting the 13900k at 85w gives us similar power draw. In which case, the 7950x is just 7% faster. So the massive efficiency difference we are talking about is 7%, or to put it better, 5 watts. That's in heavy multithreading

Meanwhile doing basic tasks on the amd part like freaking browsing draws 40 watts MORE than the intel part. 40 watts is the average power consumption I have on my 13900k during a day (i just checked, 37w according to hwinfo). So what my whole CPU draws on average during a whole day, the 7950x consumes as EXTRA just browsing the web. That is just absolutely insane and it really needs fixing.

 

[igor_kavinski]

So what my whole CPU draws on average during a whole day, the 7950x consumes as EXTRA just browsing the web. That is just absolutely insane and it really needs fixing.

And the person doing only web browsing on an expensive 7950X needs no fixing?

 

[Just Benching]

And the person doing only web browsing on an expensive 7950X needs no fixing?

What person? Who said anything about a person doing only web browsing?

 

[igor_kavinski]

Well, it's done. The silicon is out there and it can't be fixed. At least not until Zen 5 I suppose. In the mean time, I guess you will be valiantly warning users about the power wasted by Zen 4 during internet browsing. Keep up the good work, Sir.

 

[Just Benching]

Well, it's done. The silicon is out there and it can't be fixed. At least not until Zen 5 I suppose. In the mean time, I guess you will be valiantly warning users about the power wasted by Zen 4 during internet browsing. Keep up the good work, Sir.

Only as long as youll be warning them about the 5 watts in heavy mt

 

[IEC]

Something smells... familiar.

This is exactly why AMD correctly decreased the stock TDP for the X3D parts to a more reasonable cap. It's also why Intel held back the best binned chips for their 13900KS - to try to look better since they have nothing else to offer (for the time being). There will be a larger efficiency gap once X3D parts launch - in AMD's favor. Perhaps outside of the cherry-picked ones above where posters continue to confuse and conflate TDP, PPT, and power draw.

It sure will be nice for the X3D parts to take the performance crown while being more power efficient *AND* not requiring $350 premium DDR5 memory to get that last 1% of gains which all two of the RTX 4090 enjoyers who game at CPU-limited resolutions swear by. Then whichever flavor of CPU you prefer, they're all bound to get cheaper, right? Win/win.

The elephant in the room is of course the Ryzen 7 7700. It makes both the 7950X and the 13900K/S look like a complete and utter waste of money for gaming. And with a single CCD and a reasonable TDP it's got the best showing out of the box for efficiency.


A Ryzen 7700, 7600, 5800X3D or Intel 13600K look like compelling values for gaming and light productivity work. And unlike the 7950X or 13900K/S you're not paying $arm or $leg and you're not losing efficiency powering on more glued-together cores than you need. Whether you prefer the IF glue or the E-core glue.

If you buy a 7950X or a 13900K/S and you're spending most of your time idle or browsing... I question whether you truly needed that level of chip to begin with.

 

[DAPUNISHER]

Something smells... familiar.

This is exactly why AMD correctly decreased the stock TDP for the X3D parts to a more reasonable cap. It's also why Intel held back the best binned chips for their 13900KS - to try to look better since they have nothing else to offer (for the time being). There will be a larger efficiency gap once X3D parts launch - in AMD's favor. Perhaps outside of the cherry-picked ones above where posters continue to confuse and conflate TDP, PPT, and power draw.

It sure will be nice for the X3D parts to take the performance crown while being more power efficient *AND* not requiring $350 premium DDR5 memory to get that last 1% of gains which all two of the RTX 4090 enjoyers who game at CPU-limited resolutions swear by. Then whichever flavor of CPU you prefer, they're all bound to get cheaper, right? Win/win.

The elephant in the room is of course the Ryzen 7 7700. It makes both the 7950X and the 13900K/S look like a complete and utter waste of money for gaming. And with a single CCD and a reasonable TDP it's got the best showing out of the box for efficiency.
View attachment 75202

A Ryzen 7700, 7600, 5800X3D or Intel 13600K look like compelling values for gaming and light productivity work. And unlike the 7950X or 13900K/S you're not paying $arm or $leg and you're not losing efficiency powering on more glued-together cores than you need. Whether you prefer the IF glue or the E-core glue.

If you buy a 7950X or a 13900K/S and you're spending most of your time idle or browsing... I question whether you truly needed that level of chip to begin with.

Well stated.

BTW, that smell is a an old sock that is being used for sock puppet theater.

 

[Markfw]

Something smells... familiar.

This is exactly why AMD correctly decreased the stock TDP for the X3D parts to a more reasonable cap. It's also why Intel held back the best binned chips for their 13900KS - to try to look better since they have nothing else to offer (for the time being). There will be a larger efficiency gap once X3D parts launch - in AMD's favor. Perhaps outside of the cherry-picked ones above where posters continue to confuse and conflate TDP, PPT, and power draw.

It sure will be nice for the X3D parts to take the performance crown while being more power efficient *AND* not requiring $350 premium DDR5 memory to get that last 1% of gains which all two of the RTX 4090 enjoyers who game at CPU-limited resolutions swear by. Then whichever flavor of CPU you prefer, they're all bound to get cheaper, right? Win/win.

The elephant in the room is of course the Ryzen 7 7700. It makes both the 7950X and the 13900K/S look like a complete and utter waste of money for gaming. And with a single CCD and a reasonable TDP it's got the best showing out of the box for efficiency.
View attachment 75202

A Ryzen 7700, 7600, 5800X3D or Intel 13600K look like compelling values for gaming and light productivity work. And unlike the 7950X or 13900K/S you're not paying $arm or $leg and you're not losing efficiency powering on more glued-together cores than you need. Whether you prefer the IF glue or the E-core glue.

If you buy a 7950X or a 13900K/S and you're spending most of your time idle or browsing... I question whether you truly needed that level of chip to begin with.

Exactly. And for someone like me that uses then %100 load 24/7 INCLUDING avx-512 a fair amount of the time, I am saving about 100 watts, all day every day for each of my 3 7950x's. And even reducing the wattage using PBO/CO, I have 95% of the capability while saving even more power. Yes, you can down watt the Raptor lake, but at the same power level, the 7950x wins virtually everything, especially due to having avx-512.

If you game, yes, the 7700x is the choice. And for some the 7600x or the 7900x, depending on your use case.

The only way I see the Raptor lake as a viable option, is if you game and don't care that it uses more power. And when the 7xxxX3d's come out next month, that option becomes a bad option.

 

[Just Benching]

Something smells... familiar.

This is exactly why AMD correctly decreased the stock TDP for the X3D parts to a more reasonable cap. It's also why Intel held back the best binned chips for their 13900KS - to try to look better since they have nothing else to offer (for the time being). There will be a larger efficiency gap once X3D parts launch - in AMD's favor. Perhaps outside of the cherry-picked ones above where posters continue to confuse and conflate TDP, PPT, and power draw.

It sure will be nice for the X3D parts to take the performance crown while being more power efficient *AND* not requiring $350 premium DDR5 memory to get that last 1% of gains which all two of the RTX 4090 enjoyers who game at CPU-limited resolutions swear by. Then whichever flavor of CPU you prefer, they're all bound to get cheaper, right? Win/win.

The elephant in the room is of course the Ryzen 7 7700. It makes both the 7950X and the 13900K/S look like a complete and utter waste of money for gaming. And with a single CCD and a reasonable TDP it's got the best showing out of the box for efficiency.
View attachment 75202

A Ryzen 7700, 7600, 5800X3D or Intel 13600K look like compelling values for gaming and light productivity work. And unlike the 7950X or 13900K/S you're not paying $arm or $leg and you're not losing efficiency powering on more glued-together cores than you need. Whether you prefer the IF glue or the E-core glue.

If you buy a 7950X or a 13900K/S and you're spending most of your time idle or browsing... I question whether you truly needed that level of chip to begin with.

Where exactly did anyone confuse ppt with power draw? The pc world review made it quite clear, an eco 65w 7950x consumes 85w and its 7% more efficient than the 13900k. So you are saving 5 watts an hour in heavy mt workloads. Or 15w in total if you are like Mark and have 3 of those. Not bad, but nothing to write home about.

 

[Markfw]

Where exactly did anyone confuse ppt with power draw? The pc world review made it quite clear, an eco 65w 7950x consumes 85w and its 7% more efficient than the 13900k. So you are saving 5 watts an hour in heavy mt workloads. Or 15w in total if you are like Mark and have 3 of those. Not bad, but nothing to write home about.

That is totally wrong. The difference is over 100 watts, 115 to be exact. From this page:

A Lighter Touch: Exploring CPU Power Scaling On Core i9-13900K and Ryzen 9 7950X

www.anandtech.com

And later in the article, at 125 watts each, the 138900k only gets 78% of its full power benchmark, while the 7950x gets 96%. Thats efficiency domination.

 

[Abwx]

I don't know about idle power but intel beating amd in mixed workloads isn't anything new, the same was true between ALD and zen 3.

Anandtech was testing heavy MT, and even in those scenarios the difference in efficiency was tiny (8%), the zen 4 part was consuming considerably more power than the intel counterpart when both were limited at the same wattage.

For example

At 65w the difference in efficiency might look huge, but then you realize this

The amd part was drawing way more power. If you actually do the math 22911 / 71 = 322pts vs 346 pts for the amd part. So the AMD part is 7.4% more efficient at 65w in heavy MT

Those datas say that the 7950X@90W score 31179 in CB R23 while the 13900K@143W score 31615 pts.

At same perfs the 7950X is 55% more efficent if we take account of the 2% score difference in this comparison, that s the real numbers, not your gibberish maths.

 

[Carfax83]

In EVERY scenario I saw, the 13900k was consuming a LOT more power(like as much as 100 watt or more than the 7950x), except this one test. Way to cherry pick.

Wow, every scenario? I don't know why you bother even commenting on it since it's clear you didn't watch it at all. I mean, it's quite easy to disprove what you are saying. The gaming sections are very close in terms of power draw for instance.

Also, browsing the internet is by far the most common form of computing that the vast majority of people do, so I don't consider that cherry picking. It emphasizes exactly what others and I have been saying about RPL having an advantage in light to moderate loads.

 

[Just Benching]

That is totally wrong. The difference is over 100 watts, 115 to be exact. From this page:

A Lighter Touch: Exploring CPU Power Scaling On Core i9-13900K and Ryzen 9 7950X

www.anandtech.com

View attachment 75254

And later in the article, at 125 watts each, the 138900k only gets 78% of its full power benchmark, while the 7950x gets 96%. Thats efficiency domination.

Im not sure you can read graphs. Thats not an efficiency graph, thats just a power limit graph.

 

[Markfw]

Wow, every scenario? I don't know why you bother even commenting on it since it's clear you didn't watch it at all. I mean, it's quite easy to disprove what you are saying. The gaming sections are very close in terms of power draw for instance.

Also, browsing the internet is by far the most common form of computing that the vast majority of people do, so I don't consider that cherry picking. It emphasizes exactly what others and I have been saying about RPL having an advantage in light to moderate loads.

So, instead of using a real review, with test setups, and charts of the results, you expect a deaf person to watch a video, and thats what you base your opinion on.

And, @Herald , no its a power usage graph, try and read.

 

[Just Benching]

Those datas say that the 7950X@90W score 31179 in CB R23 while the 13900K@143W score 31615 pts.

At same perfs the 7950X is 55% more efficent if we take account of the 2% score difference in this comparison, that s the real numbers, not your gibberish maths.

Im comparing the equivalent power limits. At eco 65w the 7950x is 7% more efficient than the 13900k at 65w, is it not? Thats the exact same number pc world measured in their review.

Your comparison is flawed cause you are pushing 60% extra wattage to the 13900k. Of course in that scenario its less efficient and it shows your bias. Cause let me flip the script, i bet you a paycheck the 13900k at 90w is going to curbstomp the 7950x at 145w in efficiency as well 😄

 

[Just Benching]

So, instead of using a real review, with test setups, and charts of the results, you expect a deaf person to watch a video, and thats what you base your opinion on.

And, @Herald , no its a power usage graph, try and read.

Power usage is not efficiency. The difference in efficiency is a few percentage points. Or 5 in watts at eco65 settings. So congrats, you are saving 5 watts. But not while replying to me, since you are using 40 extra watts to reply to a forum post 😒

 

[Carfax83]

Good to know that. However, consider two people who do similar stuff with their CPUs, one with 13900K and the other with 7950X. Intel user will save power in browsing and lose a lot of power in multicore. AMD user will lose power in browsing but will save a lot more in multicore heavy workloads. It kinda balances out in the end for the average user. And if the user is constantly running Blender and other heavy duty workloads, they win with the 7950X with a lower power bill. It all depends on the usage scenarios. Every user has to do their research to see which company gives them the best value. It's no longer the case like Zen 2 vs. Comet Lake where AMD seemed a bit weak or Zen 3 vs. Comet Lake where Intel was clearly the loser.

Yeah I don't disagree with this at all, although it's a bit more nuanced. Both CPUs have their strengths and weaknesses. If someone was doing lots of rendering or folding, I would tell them to get a 7950x without hesitation.

But there is so much room for tweaking and improving with Raptor Lake when it comes to power consumption and thermals. For example, here is a CB23 run at 5.3ghz 235w.

Spoiler: CB23 @ 5.3ghz 235w air cooled