Intel Meteor Lake to Feature 50% Increase in Efficiency, 2X Faster iGPU

[Minus Infinity]

While the 13900K is a bit less efficient in gaming, it's more efficient than the 13700K in CBench, which actually pushes the CPU.
Inefficient? No. Using way too much power,? Yes.

View attachment 282752

I've even seen similar comments about the 4090. Just because it's using a lot of power it doesn't mean it's inefficient.

13900K and Raptor Lake's weakness comes at lower power. It scales very poorly compared to Zen 4 in the 35-105W range. 7950X at 35W is more than 2x as fast as 13900K and still have a generous lead at 105W. For mobile applications Intel knows a new architecture is needed. Meteor Lake is the beginning, but next gen doesn't really hit its stride until Arrow Lake.

 

[JustBenching]

13900K and Raptor Lake's weakness comes at lower power. It scales very poorly compared to Zen 4 in the 35-105W range. 7950X at 35W is more than 2x as fast as 13900K and still have a generous lead at 105W.

That is absolutely not true. At 35w infact the 13900k is faster than the 7950x. You probably read anandtechs review, but on anandtech their 35w limit was a 35w TDP on the 7950x. TDP on ryzen CPUS isn't power consumption, with a TDP of 35w the 7950x consumed 45w, with a tdp of 65 it consumed 90w etc.

Yes the 7950x is in fact more efficient, but the difference is way smaller than people think it is. At same wattage it's around 10 to 15% depending on what wattage you run the test.

 

[Max(IT)]

It’s time to see some efficiency instead of the usual insane race to performance

 

[Nhonho]

In this Ryzen 7000 generation, AMD should have made the CPU in a single die, with 2 MB of L2 cache per x86 core and with only 32 MB of L3. When the memory controller is on the same die of the x86 cores, the x86 cores have a very fast memory access (the fastest possible) and thus AMD CPUs would not need to have huge (and expensive) amounts of L3 cache memory. 32 MB of L3 on a CPU that has direct access to RAM is equivalent to ~64 MB of L3 on a CPU that has an intermediate chip between the x86 cores and the memory RAM. The main die (the CPU die that has the memory controller integrated) should be like this and, to increase the amount of x86 cores, it would be enough to add a chiplet of x86 cores, and the operating systems should be optimized to use mainly the cores of the main die.

AMD could even put the SATA controllers, USB controllers and other components on the chipset die and other components that don't need to be made in an expensive lithography.

In my view, AMD made a series of bad decisions on this Ryzen 7000 series (such as the lack of compatibility with DDR4 memories and socket AM4) and Radeons RDNA3 offboard graphics cards with MCM scheme. A GPU die necessarily needs to be a single die to have the lowest possible latencies.

 

[Max(IT)]

13900K and Raptor Lake's weakness comes at lower power. It scales very poorly compared to Zen 4 in the 35-105W range. 7950X at 35W is more than 2x as fast as 13900K and still have a generous lead at 105W. For mobile applications Intel knows a new architecture is needed. Meteor Lake is the beginning, but next gen doesn't really hit its stride until Arrow Lake.

Absolutely wrong.
At 90W the difference from 7950X and 13900K are quite small

 

[Nhonho]

In this Ryzen 7000 generation, AMD should have made the CPU in a single die, with 2 MB of L2 cache per x86 core and with only 32 MB of L3. When the memory controller is on the same die of the x86 cores, the x86 cores have a very fast memory access (the fastest possible) and thus AMD CPUs would not need to have huge (and expensive) amounts of L3 cache memory. 32 MB of L3 on a CPU that has direct access to RAM is equivalent to ~64 MB of L3 on a CPU that has an intermediate chip between the x86 cores and the memory RAM. The main die (the CPU die that has the memory controller integrated) should be like this and, to increase the amount of x86 cores, it would be enough to add a chiplet of x86 cores, and the operating systems should be optimized to use mainly the cores of the main die.

AMD could even put the SATA controllers, USB controllers and other components on the chipset die and other components that don't need to be made in an expensive lithography.

In my view, AMD made a series of bad decisions on this Ryzen 7000 series (such as the lack of compatibility with DDR4 memories and socket AM4) and Radeons RDNA3 offboard graphics cards with MCM scheme. A GPU die necessarily needs to be a single die to have the lowest possible latencies.

This image represents what I said about what Ryzen CPUs should be:

 

[Wirko]

That is absolutely not true. At 35w infact the 13900k is faster than the 7950x. You probably read anandtechs review, but on anandtech their 35w limit was a 35w TDP on the 7950x. TDP on ryzen CPUS isn't power consumption, with a TDP of 35w the 7950x consumed 45w, with a tdp of 65 it consumed 90w etc.

Yes the 7950x is in fact more efficient, but the difference is way smaller than people think it is. At same wattage it's around 10 to 15% depending on what wattage you run the test.

It's sad that Anandtech didn't actually measure actual system power draw with a real power meter. Instead, the reviewer relied on what the CPUs themselves reported to him.

A comparison with one or two lower power CPUs would also be very welcome here. I mean a 12900T or maybe 12900. The chip is probably better binned and the thread director properly tuned for 35W-65W operation, so the results might be quite different.

 

[Max(IT)]

It's sad that Anandtech didn't actually measure actual system power draw with a real power meter. Instead, the reviewer relied on what the CPUs themselves reported to him.

A comparison with one or two lower power CPUs would also be very welcome here. I mean a 12900T or maybe 12900. The chip is probably better binned and the thread director properly tuned for 35W-65W operation, so the results might be quite different.

Intel T processors are useless: you can buy a normal 12900 and set the PL1 as you wish

 

[Space Lynx]

I probably will sell my Ryzen 5600 when Meteor Lake drops. That should be a proper upgrade.

 

[BoboOOZ]

Chungus refers mostly to the size - The weight's not to bad for a 16" but I mean a Zephyrus G14 is cheaper, faster, smaller, and a similar weight. When you have a small bag, weight is irrelevant because the 16" won't fit in it.

28W is absolutely doable in 1.2Kg laptops because I've already owned two of them - a Lenovo S540 13" Ryzen with a 35W cTDP (dual-fan) and an HP Envy with 28W in a very thin chassis. The S540 was exceptionally small, light, and powerful for under £1000. We seem to be taking steps backwards when it comes to adequately-cooled, reasonably priced AMD ultraportables. My hope with these new Intels is that there will be so many design wins (because Intel) that at least some of them won't suck!

The reason I keep mentioning ultraportables with APUs is because anything approaching 1.5kg is going to lose out to the enormous selection of good, readily-available 13-14" gaming laptops. Not just in size/portability - but also cost, because those small gaming laptops are popular and seem to have economy of scale. So giving up a dGPU and going with integrated graphics isn't just for fun; it's a massive performance sacrifice that isn't worth making unless there's a significant benefit in portability.

Have you considered the ROG Flow x13?
I've been waiting for an affordable 6800U Yoga to pop out in France for a while, hasn't happened yet, so the other day I stumbled upon an as-new open box ASUS ROG Flow with an 6800H and a 3050. I didn't really want the model with the 3050 (there is a model only with the APU), but since it was cheaper than the new APU only model I went for it. This way I still have the 3050 for when I want to play Cyberpunk in Ultra with RT on.
Long story short, with the GPU disabled, the 6800H is awesome, I just played Fortnite at 110 fps, the performance is better than expected, perhaps that is what you need, it's small, light, but it can also go up to 45W TDP if you want it to. This is head and shoulders above the Vega 8, and it's also a 2-in1 if the tablet/tent format is tempting for you

 

[Chrispy_]

Have you considered the ROG Flow x13?
I've been waiting for an affordable 6800U Yoga to pop out in France for a while, hasn't happened yet, so the other day I stumbled upon an as-new open box ASUS ROG Flow with an 6800H and a 3050. I didn't really want the model with the 3050 (there is a model only with the APU), but since it was cheaper than the new APU only model I went for it. This way I still have the 3050 for when I want to play Cyberpunk in Ultra with RT on.
Long story short, with the GPU disabled, the 6800H is awesome, I just played Fortnite at 110 fps, the performance is better than expected, perhaps that is what you need, it's small, light, but it can also go up to 45W TDP if you want it to. This is head and shoulders above the Vega 8, and it's also a 2-in1 if the tablet/tent format is tempting for you

Might just be regional pricing but those are in the £1700-1800 range which is, IMO, easily double the cost of what other options are offering those specs for.

The rule of thumb with laptops in the UK seems to be that if it's nice, it's a rip-off. I have branch offices in Europe so I often buy hardware in France or Germany, but with laptops I really want a UK model for the UK/ISO keyboard.

 

[mkppo]

They're really gunning for Apple silicon efficiency.

You mean they are barely catching up to AMD who are already 50% more efficient.

It's AMD who are competing with Apple at the moment, not Intel.

 

[Imouto]

Absolutely wrong.
At 90W the difference from 7950X and 13900K are quite small

Read what you are replying to again.

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

www.anandtech.com

At 35 W and 65 W the 7950X demolishes the 13900K in efficiency running CB. In other tests the difference is an absolute embarrassment for the 13900K.

 

[JustBenching]

Read what you are replying to again.

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

www.anandtech.com

At 35 W and 65 W the 7950X demolishes the 13900K in efficiency running CB. In other tests the difference is an absolute embarrassment for the 13900K.

No it doesn't. Try to read the review again. The 7950x at 35w wasn't drawing much more wattage than the 13900k. At 65w it was drawing 50% more watts. Go to page 4 and see power draw.

It's pretty common lately, fake numbers and full of amd propaganda about their insane efficiency

You mean they are barely catching up to AMD who are already 50% more efficient.

It's AMD who are competing with Apple at the moment, not Intel.

50% more efficient? Lolno. More like 10 to 15%

 

[Imouto]

No it doesn't. Try to read the review again. The 7950x at 35w wasn't drawing much more wattage than the 13900k. At 65w it was drawing 50% more watts. Go to page 4 and see power draw.

It's pretty common lately, fake numbers and full of amd propaganda about their insane efficiency


50% more efficient? Lolno

That's peak power and has nothing to do with average power.

 

[JustBenching]

That's peak power and has nothing to do with average power.

Yeah, right, how does it peak to 50% higher? Please explain

You think a 7950x at 65w scores 32k in cbr23? Lolk

These are the actual power restricted numbers

https://www.club386.com/wp-content/uploads/2022/10/Wattage-Comparison-02C5Hk9NKZFrktfm-1068x505.png

 

[Imouto]

Yeah, right, how does it peak to 50% higher? Please explain

Same way the 13900K goes 30% over the TDP at stock.

You think a 7950x at 65w scores 32k in cbr23? Lolk

Yes. I provided a review proving so even if the methodology is flawed. You, on the other hand, provided literally nothing but childish remarks.

 

[JustBenching]

Same way the 13900K goes 30% over the TDP at stock.



Yes. I provided a review proving so even if the methodology is flawed. You, on the other hand, provided literally nothing but childish remarks.

As per the reviewr, the 13900K peaks 14% higher. The 7950x peaks 50% higher. Obviously you have not used any of those cpus, else you would realize the 7950x does not score 31k+ at 65watts,lol. Not even close. I already posted a review that actually tests at similar wattages, as you can see the difference in efficiency is small. 10-15%.

I can lead you to water but I can't make you drink, so whatever.

Computerbasede also run at 65w. A stock 7950x scores 55% higher than a 65w 7950x. That means, if the 7950x at 65w scores 31k like Anand tech claims, at stock it should score over 45k!!! Yeah, right, you are obviously wrong my man, admit it and move on.

AMD Ryzen 7000 im Test: So schnell sind 7950X und 7700X: Leistungsaufnahme und Effizienz

Ryzen 9 7950X & Ryzen 7 7700X im Test: Leistungsaufnahme und Effizienz / TDP und PPT steigen deutlich

www.computerbase.de

 

[Imouto]

As per the reviewr, the 13900K peaks 14% higher. The 7950x peaks 50% higher. Obviously you have not used any of those cpus, else you would realize the 7950x does not score 31k+ at 65watts,lol.

Dude, stop with the math. You can't even do a simple percentage right. At 65 W limit the 7950X goes to 90.3 W peak. That's not 50% higher, that's 38.92% higher. Again, that's peak power and can be recorded over less than a microsecond. But you held onto this and you don't seem to let go, wrong math included.

Not even close. I already posted a review that actually tests at similar wattages, as you can see the difference in efficiency is small. 10-15%.

None of them physically measures the power consumption or the energy used like TPU does.

I can lead you to water but I can't make you drink, so whatever.

Going by how you do math that water is probably poisoned.

Computerbasede also run at 65w. A stock 7950x scores 55% higher than a 65w 7950x. That means, if the 7950x at 65w scores 31k like Anand tech claims, at stock it should score over 45k!!! Yeah, right, you are obviously wrong my man, admit it and move on.

Why are you trying so hard to die in this hill? Both of them go over the TDP or PPT. AMD goes up to 40% at peak power. Intel goes up to 30% at peak power. I don't know which cooler any of them used. Maybe Anand's was really good. Maybe CB's was shit.

Stop acting like you are in possession of the absolute and undeniable truth.

Want me to say you chose the right CPU? You did! Are you happy now?

 

[mkppo]

No it doesn't. Try to read the review again. The 7950x at 35w wasn't drawing much more wattage than the 13900k. At 65w it was drawing 50% more watts. Go to page 4 and see power draw.

It's pretty common lately, fake numbers and full of amd propaganda about their insane efficiency


50% more efficient? Lolno. More like 10 to 15%

Not sure how you're trying to even argue about efficiency here with all the numbers in front of you (i'm just going to ignore your fake numbers part, not sure what you're referring to). In the Anandtech power consumption graph, the 13900K and 7950X consume about the same amount of power at 105 W and 125 W respectively, so all your arguments about AMD consuming more power vs what it's set at go out the window. Now let's take a look at performance at that power with the 7950X at 125W and 13900K at 105W.

C-RAY:
No Power limit: 13900K 3.3% faster
105W/125W: 7950X is 12.7% faster

Cinebench:

No Power limit: 13900K is 5.3% faster
105W/125W: 7950X is 13.8% faster

X264 1080p:
No Power limit: 7950X is 8.9% faster
105W/125W: 7950X is 24.4% faster

X264 Bosp 4K:
No Power limit: 7950X is 3% faster
105W/125W: 7950X is 20.5% faster


Keep in mind that those are best case scenarios, as intel loses much more performance as you keep reducing power. Want examples? So a 35W 7950X consumes 6W more than a 35W 13900K but the 7950X is 53.2% faster in CB, 66.1% faster in X264 and the list goes on. Do you see where I get my 50% number from? I was referring to gunning for apple, and the 35W numbers are a lot more relevant there. So rather than asking people to "Admit it and move on" it's best you check your own facts before getting combative.

 

[JustBenching]

Dude, stop with the math. You can't even do a simple percentage right. At 65 W limit the 7950X goes to 90.3 W peak. That's not 50% higher, that's 38.92% higher. Again, that's peak power and can be recorded over less than a microsecond. But you held onto this and you don't seem to let go, wrong math included.



None of them physically measures the power consumption or the energy used like TPU does.



Going by how you do math that water is probably poisoned.



Why are you trying so hard to die in this hill? Both of them go over the TDP or PPT. AMD goes up to 40% at peak power. Intel goes up to 30% at peak power. I don't know which cooler any of them used. Maybe Anand's was really good. Maybe CB's was shit.

Stop acting like you are in possession of the absolute and undeniable truth.

Want me to say you chose the right CPU? You did! Are you happy now?

So we are also going to ignore computerbase.de review as well. Okay

Not sure how you're trying to even argue about efficiency here with all the numbers in front of you (i'm just going to ignore your fake numbers part, not sure what you're referring to). In the Anandtech power consumption graph, the 13900K and 7950X consume about the same amount of power at 105 W and 125 W respectively, so all your arguments about AMD consuming more power vs what it's set at go out the window. Now let's take a look at performance at that power with the 7950X at 125W and 13900K at 105W.

C-RAY:
No Power limit: 13900K 3.3% faster
105W/125W: 7950X is 12.7% faster

Cinebench:

No Power limit: 13900K is 5.3% faster
105W/125W: 7950X is 13.8% faster

X264 1080p:
No Power limit: 7950X is 8.9% faster
105W/125W: 7950X is 24.4% faster

X264 Bosp 4K:
No Power limit: 7950X is 3% faster
105W/125W: 7950X is 20.5% faster


Keep in mind that those are best case scenarios, as intel loses much more performance as you keep reducing power. Want examples? So a 35W 7950X consumes 6W more than a 35W 13900K but the 7950X is 53.2% faster in CB, 66.1% faster in X264 and the list goes on. Do you see where I get my 50% number from? I was referring to gunning for apple, and the 35W numbers are a lot more relevant there. So rather than asking people to "Admit it and move on" it's best you check your own facts before getting combative.

Some calculations from anandtechs numbers.

7950x @ 65w = 90w power draw with a score 31.179
13900k @ 65w = 71.4w power draw with a score of 22.911

Do you know math? You know what the difference in efficiency is between those 2 scores? 7%. You are claiming 50, the actual review says 7.

31179 / 90 = 345 pts/watt
22911 / 71.4 = 320 pts/watt

345/320 = 1.07

Keep up the amd propaganda

 

[Imouto]

So we are also going to ignore computerbase.de review as well. Okay

I'm not ignoring them. Both of them have different methodology. Taking in mind how modern GPUs and CPUs work I believe Anand's is a more real world approach.

Keep up the amd propaganda

Yeah. I'm the one doing propaganda picking a peak figure to do calculations. FFS. You don't seem to have the slightest technical background. Efficiency is calculated with the energy used and the work done so neither CB's or Anand's are a strict indication of efficiency.

 

[mkppo]

So we are also going to ignore computerbase.de review as well. Okay


Some calculations from anandtechs numbers.

7950x @ 65w = 90w power draw with a score 31.179
13900k @ 65w = 71.4w power draw with a score of 22.911

Do you know math? You know what the difference in efficiency is between those 2 scores? 7%. You are claiming 50, the actual review says 7.

31179 / 90 = 345 pts/watt
22911 / 71.4 = 320 pts/watt

345/320 = 1.07

Keep up the amd propaganda

You are still being combative and taking personal shots, really no need for that. Yes I know math, thank you. The percentages that I posted are correct, maybe try to recreate the numbers before asking me if i know math? Read my post again, at no point did I mention 65W. I merely said they consume the same actual power at 105W and 125W respectively. And then, now that power was normalized, I posted some comparison numbers where the math was correct. Instead of refuting those numbers, you bring out a different set of numbers?

Why are you comparing the intel system at 71W power draw with the Ryzen at 90w? You're extrapolating pts/watt at different power levels, which mean they are on a different V/F curve and the one with lower power (13900K in your case) will obviously be higher perf/watt but it's still 7% less efficient according to you. Then you go on about my 50% again and simply ignore that I mentioned it's at 35W, I guess i'll just have to do the math here. Keep in mind that the power isn't exactly the same which I mentioned earlier, but not as wildly different as 71 vs 90W as in your case.

Power at 35W: 7950X/13900K: 45.1W vs 39.3W, so 4.8 watts less for 13900K
Performance in X264: 148.7 pts vs 89.5 pts.

Performance difference: ((148.7-89.5)/89.5)*100 = 66.14%

You may argue that the Intel consumes 4.8 watts less and open another can of worms, so i'll just use your pts/watt approach which is relatively more relevant here as the wattages are pretty close (and not 71 vs 90 in your case - remember different v/f curves). Again, it's not entirely accurate because the CPU with lower power will be in an advantageous position here since they are on a more favourable V/F curve (13900K in this case).

7950X: 148.7/45.1 = 3.297 pts/watt
13900K: 89.5/39.3 = 2.277 pts/watt

Percentage difference: 7950X is 44.79% faster. Sorry I forgot to put the calculations here, it'll be ((3.297 - 2.277) / 2.277) * 100

Clear now? Don't resort to personal attacks, we're not children here.

 

[JustBenching]

You are still being combative and taking personal shots, really no need for that. Yes I know math, thank you. The percentages that I posted are correct, maybe try to recreate the numbers before asking me if i know math? Read my post again, at no point did I mention 65W. I merely said they consume the same actual power at 105W and 125W respectively. And then, now that power was normalized, I posted some comparison numbers where the math was correct. Instead of refuting those numbers, you bring out a different set of numbers?

Why are you comparing the intel system at 71W power draw with the Ryzen at 90w? You're extrapolating pts/watt at different power levels, which mean they are on a different V/F curve and the one with lower power (13900K in your case) will obviously be higher perf/watt but it's still 7% less efficient according to you. Then you go on about my 50% again and simply ignore that I mentioned it's at 35W, I guess i'll just have to do the math here. Keep in mind that the power isn't exactly the same which I mentioned earlier, but not as wildly different as 71 vs 90W as in your case.

Power at 35W: 7950X/13900K: 45.1W vs 39.3W, so 4.8 watts less for 13900K
Performance in X264: 148.7 pts vs 89.5 pts.

Performance difference: ((148.7-89.5)/89.5)*100 = 66.14%

You may argue that the Intel consumes 4.8 watts less and open another can of worms, so i'll just use your pts/watt approach which is relatively more relevant here as the wattages are pretty close (and not 71 vs 90 in your case - remember different v/f curves). Again, it's not entirely accurate because the CPU with lower power will be in an advantageous position here since they are on a more favourable V/F curve (13900K in this case).

7950X: 148.7/45.1 = 3.297 pts/watt
13900K: 89.5/39.3 = 2.277 pts/watt

Percentage difference: 7950X is 44.79% faster. Sorry I forgot to put the calculations here, it'll be ((3.297 - 2.277) / 2.277) * 100

Clear now? Don't resort to personal attacks, we're not children here.

Ah, so you are saying that if you cherrypick specific benches from specific reviews at whatever power limits helps your argument you might come to your preconceived conclusion. Let's totally ignore computerbases review which actually measure average power draw...

So, back to Anand tech.

Povray @35w = 147pts per watt for the 7950x, 124 for the 13900k

Povray at 65w = 115 pts per watt for the 7950x, 118 for the 13900k

Let's go for power normalized then, 83pts per watt for the 7950x at 105w, 78 for the 13900K at 125w. The difference is LESS than 10%.

Keep up the defense

 

[wheresmycar]

Everyone please make note of these efficiency arguments. Its extremely important. The least efficient CPU is possibly the one causing global warming and the best of the 2 is the solution to saving our planet. We must fight tooth and nail to find the efficiency-King or we are all doomed!!!!

"We will not go quietly into the night! We will not vanish without a fight! Efficiency will live onnnnnn!"