Question Raptor Lake - Official Thread

[Hulk]

Since we already have the first Raptor Lake leak I'm thinking it should have it's own thread.
What do we know so far?
From Anandtech's Intel Process Roadmap articles from July:

Built on Intel 7 with upgraded FinFET
10-15% PPW (performance-per-watt)
Last non-tiled consumer CPU as Meteor Lake will be tiled

I'm guessing this will be a minor update to ADL with just a few microarchitecture changes to the cores. The larger change will be the new process refinement allowing 8+16 at the top of the stack.

Will it work with current z690 motherboards? If yes then that could be a major selling point for people to move to ADL rather than wait.

 

[nicalandia]

Raptor Lake has a better memory controller. 13700HX is a re-purposed ADL die so it's limited to DDR5-4800. So the slower memory speed on 13th gen laptops can help identify whether the die is ADL or RPL.

Do You have a list of SKUs that support DDR5-5600? Because up until now I have only seen three the 13900HX, 13950HX and the 13980HX(perhaps missing one). but that's it. Would like to investigate the others.

 

[igor_kavinski]

Do You have a list of SKUs that support DDR5-5600?

I'm not sure what we are disagreeing about. All I re-iterated (and what mikk originally said) was that memory speed may well be an indicator of the die being used. If it's 5600, it's RPL. If lower than that, it's likely ADL.

In other words, Intel produced too many ADL dies and it is now going to knowingly and deceptively punish the ignorant 13th gen laptop buyers with lower memory speeds by making them think they are getting the latest technology when in fact, they will be getting 2021 tech in 2023.

 

[nicalandia]

I'm not sure what we are disagreeing about. All I re-iterated (and what mikk originally said) was that memory speed may well be an indicator of the die being used. If it's 5600, it's RPL. If lower than that, it's likely ADL.

This is what he posted.

Pretty sure all of the DDR5-5600/LPDDR5-6400 models are Raptor based. This isn't unexpected because of the CPU ID change. So much for the ADL rename FUD.

https://twitter.com/x/status/1610276903247663106

So where is the list of all those DDR5-5600/LPDDR5-6400 Models? If it's 3 or 4 then there is no such thing as FUD as he alluded.

 

[Kocicak]

In other words, Intel produced too many ADL dies and it is now going to knowingly and deceptively punish the ignorant 13th gen laptop buyers with lower memory speeds by making them think they are getting the latest technology when in fact, they will be getting 2021 tech in 2023.

Speaking about memory speed in laptops and finished desktops - is it not often that these are shipped with memory clocked way less than what is the memory capable of (or what is the memory controller spec) anyway?

These finished PCs are also often shipped with the slowest (cheapest) RAM sticks available.

I would also argue that most people do not care if they get Alder or Raptor lake CPU, the performance will be from 95% determided by core count and frequency alone.

 

[igor_kavinski]

Speaking about memory speed in laptops

Problem with DDR5 laptops with RAM sockets is that there is nothing better available than DDR5-4800 SODIMMs. So laptop OEMs aren't going to go to the trouble of validating their ADL laptop designs for higher speed memory. I think it may be possible to run higher speed DDR5 SODIMMs in very limited number of ADL gaming laptops that support XMP in their UEFI/BIOS.

 

[jpiniero]

Newegg has the 13400 in stock. It's $239. The 12600K is 12 bucks more.

 

[Carfax83]

Interesting comparison between Raptor Lake and Zen 4 power consumption in creator workloads. The Intel advantage in idle power draw is a huge benefit, and unless all the cores are fully active, Raptor Lake significantly outperforms the 7950x in the Adobe programs in performance per watt.

 

[Kocicak]

The Intel advantage in idle power draw is a huge benefit...

I wonder why he did not test monolithic Ryzens 5600G and 5700G as well, they even include much more powerful graphic card too.

I bet the results would have been MUCH DIFFERENT.

BTW I tested power efficiency of 5700G vs. 12600K a while ago, and in single thread (Cinebench R23) 5700G is more than TWICE AS EFFICIENT as 12600K and in max. multithread load 30% more efficient.

 

[DrMrLordX]

The Intel advantage in idle power draw is a huge benefit, and unless all the cores are fully active, Raptor Lake significantly outperforms the 7950x in the Adobe programs in performance per watt.

Unless you're really worried about your power bill or something, the only thing that matters is whether the cooling solution works and whether it stays within an acceptable noise level. Idle power draw is pretty meaningless outside of a battery-limited environment. That whole video is kind of cringe. 100% utilization power draw is what really matters.

 

[Carfax83]

I wonder why he did not test monolithic Ryzens 5600G and 5700G as well, they even include much more powerful graphic card too.

I bet the results would have been MUCH DIFFERENT.

BTW I tested power efficiency of 5700G vs. 12600K a while ago, and in single thread (Cinebench R23) 5700G is more than TWICE AS EFFICIENT as 12600K and in max. multithread load 30% more efficient.

He's not just testing power efficiency, but performance per watt as well. The Ryzen 5600G and 5700G would have been decimated in performance per watt.

As for Cinebench, rendering is probably the only creator type workload where Ryzen has an advantage in terms of power efficiency and performance per watt.

 

[Carfax83]

Idle power draw is pretty meaningless outside of a battery-limited environment.

I don't see how you can come to that conclusion, because even on desktops, the vast majority of the time your PC is in idle mode. As I write this post, my CPU downclocked to 1100mhz and 0.648V.

That whole video is kind of cringe. 100% utilization power draw is what really matters.

How often is your CPU ever at 100% utilization? It's rare, unless you are rendering or doing heavy encoding both of which most people don't do.

That said, the testing that YouTuber did wasn't just about power efficiency, but performance per watt as well and Intel beat AMD by significant margins.

Raptor Lake is power efficient and has high performance per watt provided the cores aren't maxed out.

 

[Thunder 57]

He's not just testing power efficiency, but performance per watt as well. The Ryzen 5600G and 5700G would have been decimated in performance per watt.

As for Cinebench, rendering is probably the only creator type workload where Ryzen has an advantage in terms of power efficiency and performance per watt.

I'd suspect the 5600G/5700G to do well in efficiency and performance per watt. They are monolithic and there is a reason MD laptop APU's are all monolithic. If I had to guess Intel will lose that advantage with the move to tiles and Meteor Lake. Why else would they cut the number of P cores? Intel was smart to come up with a hybrid design because it is saving them right now.

 

[Kocicak]

The Ryzen 5600G and 5700G would have been decimated in performance per watt.

I wrote that 5700G is more than 200% (ST) to 30% (MT) better than 12600K in performance per watt.

5700G scores 1460 at 12 W and 12600K 1908 at 34 W in Cinebench R23 Single thread.

 

[In2Photos]

I don't see how you can come to that conclusion, because even on desktops, the vast majority of the time your PC is in idle mode. As I write this post, my CPU downclocked to 1100mhz and 0.648V.



How often is your CPU ever at 100% utilization? It's rare, unless you are rendering or doing heavy encoding both of which most people don't do.

That said, the testing that YouTuber did wasn't just about power efficiency, but performance per watt as well and Intel beat AMD by significant margins.

Raptor Lake is power efficient and has high performance per watt provided the cores aren't maxed out.

How many creators are sitting around with their PC at idle? Creators are spending time at their PC creating, not surfing the internet posting in forums.

 

[Carfax83]

I wrote that 5700G is more than 200% (ST) to 30% (MT) better than 12600K in performance per watt.

5700G scores 1460 at 12 W and 12600K 1908 at 34 W in Cinebench R23 Single thread.

Why do you keep bringing up CB? The review focused on Adobe Photoshop and Premiere Pro, neither of which max out the CPU for very long if at all.

I already acknowledged that AMD still has the advantage in rendering efficiency due to rendering maxing out the cores as well as using big cores only as well as a better node.

The whole point of the review was to show the power efficiency and performance per watt of Intel and AMD CPUs is commonly used creator apps. CB isn't a creator app, and most creators don't do rendering. Rendering is even more niche than encoding!

 

[Carfax83]

How many creators are sitting around with their PC at idle? Creators are spending time at their PC creating, not surfing the internet posting in forums.

Well apparently you didn't watch the review (can't say I'm surprised) because even during the load tests, Intel still pulled less power than AMD while outperforming the latter.

The idle power draw is just icing on the cake. And it's silly to imply that creators are using their PCs at full capacity all the time.

 

[In2Photos]

Well apparently you didn't watch the review (can't say I'm surprised) because even during the load tests, Intel still pulled less power than AMD while outperforming the latter.

The idle power draw is just icing on the cake. And it's silly to imply that creators are using their PCs at full capacity all the time.

I didn't watch the video, I was responding to your comment. Didn't realize there was a pre-requisite to responding.

Although it appears you didn't read my post because nowhere did I say or imply that creators would be using their system at full capacity. I simply stated that they wouldn't be at idle or merely posting on an internet forum.

 

[Hans Gruber]

I am not trying to steal the Raptor Lake Thunder but the 4nm (7nm) silicon is what most people are waiting for from Intel. On the other side the non "X" Zen 4 CPU's are coming very soon. Not to mention the 3D-vcache CPU's. If AMD released 24 and 32 core variants of Zen 4 with the reduced CPU TDP of 105-130w. That would give Intel some problems. Instead of drinking a full bottle of wine for power. AMD would be back to a glass or two of wine of power usage.

Meanwhile, Intel has gone from a half gallon of the hard stuff of power consumption down to a more manageable 1/5th or liter of your hard liquor of choice. Either way, Intel is still a drunk with power usage.

By increasing the TDP of Zen4 to offset the modest IPC gains. AMD has eliminated their power sipping advantage that was painfully clear to Intel. It's why most who are sitting on the sidelines look forward to what Intel's 7nm silicon provides in efficiency and performance.

 

[Harry_Wild]

Because they go to OEMs first to be used in office/corporate PCs.

You will keep waiting like this forever. By the time 13600 is available, Intel will reveal juicy details about Raptor Lake Refresh and then you will start dreaming about their non-K SKUs. Just get a i5-12600 and be done with it.

😵‍💫☹️

 

[ondma]

I didn't watch the video, I was responding to your comment. Didn't realize there was a pre-requisite to responding.

Although it appears you didn't read my post because nowhere did I say or imply that creators would be using their system at full capacity. I simply stated that they wouldn't be at idle or merely posting on an internet forum.

Yea, dont watch the video. The facts might confuse you.

 

[Abwx]

Interesting comparison between Raptor Lake and Zen 4 power consumption in creator workloads. The Intel advantage in idle power draw is a huge benefit, and unless all the cores are fully active, Raptor Lake significantly outperforms the 7950x in the Adobe programs in performance per watt.

Adobe softwares use GPU acceleration with Intel CPUs, so that s more of a falsehood than anything else.

Perf/watt wise the 13900K fall flat compared to a 7950X not only in Cinebench but even more in INT code based software.

 

[Carfax83]

Adobe softwares use GPU acceleration with Intel CPUs, so that s more of a falsehood than anything else.

Last time I checked, Zen 4 also has iGPUs and they are used in Adobe Premiere Pro. Without GPU acceleration, Intel is faster in Adobe Premiere Pro. Computerbase.de tested it in software and hardware mode for their 13th gen review and Intel pulls ahead in software mode, while AMD has a slight edge in hardware mode.

Core i9-13900K, i7-13700K & i5-13600K: Gaming Kings Review: Benchmarks in Apps - ComputerBase

Perf/watt wise the 13900K fall flat compared to a 7950X not only in Cinebench but even more in INT code based software.

Apparently not in Adobe Premiere Pro or Photoshop, both of which primarily use the CPU and not the GPU.

 

[Timur Born]

Perf/watt wise the 13900K fall flat compared to a 7950X not only in Cinebench but even more in INT code based software.

How much power does an undervolted 7950X use for a 40-41k Cinebench 23 results? My undervolted 13900K uses about 240 W.

 

[Abwx]

Last time I checked, Zen 4 also has iGPUs and they are used in Adobe Premiere Pro. Without GPU acceleration, Intel is faster in Adobe Premiere Pro. Computerbase.de tested it in software and hardware mode for their 13th gen review and Intel pulls ahead in software mode, while AMD has a slight edge in hardware mode.

Core i9-13900K, i7-13700K & i5-13600K: Gaming Kings Review: Benchmarks in Apps - ComputerBase



Apparently not in Adobe Premiere Pro or Photoshop, both of which primarily use the CPU and not the GPU.

There s no way that the 13900K is more efficient in CPU only tasks, the 7950X efficency is even better in INT based codes, FI the 7950X use only 189W in Handbrake and is 11% faster than the 13900K@253W.

Besides Computerbase tests are GPU accelerated as well, such gaps cant be explained by CPU perfs.

 

[DrMrLordX]

I don't see how you can come to that conclusion, because even on desktops, the vast majority of the time your PC is in idle mode. As I write this post, my CPU downclocked to 1100mhz and 0.648V.

And? How does it help you to use less power under that scenario?