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.

 

[coercitiv]

Possibly. But here is my thinking. Within a power budget if I cap the P's at 5GHz all core then perhaps the E's run at 3.8. But if I let the P's go to 5.2 that might reduce the E's to 2.5. It would require testing but my thought was to try and keep the E's and P's in the more efficient part of their v/f curves.

So in other words you're arguing that Intel engineers don't know how to balance their clock/power curve

Joke aside, on Alder Lake they tend to aim for the frequency combo that takes maximum advantage of the shared voltage (P core, E core, ring bus). The only time I've seen the power management go haywire was when I limited power bellow 35W: at that point something snapped in the internal logic, and IIRC I had to manually limit P core fmax to "help" power managemtn make the correct choices again.

 

[Accord99]

If I do the air cooling route, I will use a Noctua NH-U12A, which is the best one they have at the moment.

I'm curious to see what clock speeds a 13900K will be able to achieve with such a cooler and power limited and undervolted.

A Chinese reviewer used a Thermalright AS 120, and undervolted their 13900K to 5.2/4.3 at 1.23V. Power use went down to 190W in Cinebench R23 but still almost hit 40K. Temperatures were only mid 70s to low 80s running the AIDA Stress FPU test.

 

[controlflow]

Club386 finally did a review comparing 13900k vs 7950x in gaming using a proper RTX 4090. FYI Both systems used DDR5-6000 CL30 here.

Best CPU for Nvidia GeForce RTX 4090: Intel Core i9-13900K vs. AMD Ryzen 9 7950X | Club386

Which flagship CPU can handle the awesome might of GeForce RTX 4090? We identify a clear winner.

www.club386.com

The results at FHD resolution make things quite clear. When we are actually CPU bound in a game, the 13900k puts some pretty solid distance from the 7950x.



You can look through the review, things are a bit closer in some games and a bit further apart in others but the 13900K is a beast at gaming. AMD will certainly be needing the 3D cache models of Zen 4 to compete here. I suspect a Zen 4 3D vs 13900K (or future KS) with DDR5 7000+ will be pretty closely matched in gaming.

 

[Hulk]

So in other words you're arguing that Intel engineers don't know how to balance their clock/power curve

Joke aside, on Alder Lake they tend to aim for the frequency combo that takes maximum advantage of the shared voltage (P core, E core, ring bus). The only time I've seen the power management go haywire was when I limited power bellow 35W: at that point something snapped in the internal logic, and IIRC I had to manually limit P core fmax to "help" power managemtn make the correct choices again.

That is a very good point so it might be best to just set the power limit and leave it alone.

 

[nicalandia]

Club386 finally did a review comparing 13900k vs 7950x in gaming using a proper RTX 4090. FYI Both systems used DDR5-6000 CL30 here.

You an look through the review, things are a bit closer in some games and a bit further apart in others but the 13900K is a beast at gaming. AMD will certainly be needing the 3D cache models of Zen 4 to compete here. I suspect a Zen 4 3D vs 13900K (or future KS) with DDR5 7000+ will be pretty closely matched in gaming.

Either a 7700X or a well tuned 7950X(single CCD) should be able to exceed the 7950X stock performance.

 

[controlflow]

Either a 7700X or a well tuned 7950X(single CCD) should be able to exceed the 7950X stock performance.

Would be interesting to see where such a CPU would fit in the above charts.

Can you enable/disable CCDs without a reboot?

 

[nicalandia]

Would be interesting to see where such a CPU would fit in the above charts.

Can you enable/disable CCDs without a reboot?

I am not sure, but this guy got a 25% gaming boost on a single CCD

 

[exquisitechar]

I am not sure, but this guy got a 25% gaming boost on a single CCD

View attachment 69802

In one game, yes.

 

[nicalandia]

In one game, yes.

I know, so we need to test on the one just posted on a 4090, so any of your Rich Lot are up for this??? PLEASE..!

 

[Carfax83]

Noctua NH-U12A
I have a mATX case so I'm kind of limited. The 13900K is pushing my cooler no doubt.

Yeah that case is definitely holding you back. Large roomy cases help with temps a lot for air cooling. Have you considered getting one of those contact frames? May shave off a few degrees.

 

[nicalandia]

Do we have a 13900K Builder's thread? I believe one could use a 13900K with an Air Cooler and not lose any performance compared to stock if the CPU IHS is lapped to a Copper Mirror Shine, Alphacool Apex Backplate

 

[Asterox]

I LOVE that second video, but would like to see the 13900k and the 7950x compared. In the bottom end one, AMD won gaming by ~6% but Intel won productivity by ~20% due to more cores. (12 threads compared to 20) But on the top end, the thread count is the same, so it would be interesting. But even at the bottom end the Intel took a lot more power.

Again, "not only four games tested as in the case of GamerNexus".

 

[Markfw]

Again, "not only four games tested as in the case of GamerNexus".

OK, I watched the whole thing. The performance in virtually all tests was on average almost a wash, but in power consumption it was like 100 watts (41% percent more he said, like 110 watts) or more for the Raptor Lake. Between that and the upgradeability, the author chose the 7950x. I think for even more multithreaded and longer times for the workloads, the 7950x could also come out farther ahead.

BUT as I said, in performance on short workloads they are almost the same, just takes Raptor lake a lot more power.

The next thing that would be nice to see, is a similar test with both stuck at the same power level, say 150 watts or so, and also very long multi-threaded benchmarks without power constraints.

 

[Hulk]

So in other words you're arguing that Intel engineers don't know how to balance their clock/power curve

Joke aside, on Alder Lake they tend to aim for the frequency combo that takes maximum advantage of the shared voltage (P core, E core, ring bus). The only time I've seen the power management go haywire was when I limited power bellow 35W: at that point something snapped in the internal logic, and IIRC I had to manually limit P core fmax to "help" power managemtn make the correct choices again.

You are right I set it all to "Auto" and just set my power limits. Cinebench scores went up ~800 points vs. my settings.
One less thing to worry about. Thanks.

 

[controlflow]

OK, I watched the whole thing. The performance in virtually all tests was on average almost a wash, but in power consumption it was like 100 watts (41% percent more he said, like 110 watts) or more for the Raptor Lake. Between that and the upgradeability, the author chose the 7950x. I think for even more multithreaded and longer times for the workloads, the 7950x could also come out farther ahead.

BUT as I said, in performance on short workloads they are almost the same, just takes Raptor lake a lot more power.

The next thing that would be nice to see, is a similar test with both stuck at the same power level, say 150 watts or so, and also very long multi-threaded benchmarks without power constraints.

He says at the 11 minute mark that whether we are talking games or applications, the 13900K was the better performer in most cases

He will be getting the 7950x due to him valuing platform longevity and he didn't like that socket 1700 was EOL.

Of course the performance difference is often small even though there are some use cases where the 13900k gains a bit more like in gaming. The power consumption difference is very rarely going to be anywhere near 100 watts unless you are talking about full load MT throughput workloads and even then the difference is only that big when the 13900k is running with those 300w+ power limits. Set to a low 200w power limit the 13900k will lose to a 7950x in MT benchmarks but it won't be by much.

This probably makes the 7950x the better workstation CPU but I think the 13900k is the better choice for more desktop/mixed workloads/gaming. If you want to run a bunch of these at 100% 24/7 I can see why the 7950x would be the clear choice but I don't think that applies to the typical user. IMO for the more typical use cases, the cheaper price of the 13900K,the cheaper platform and the top performance in gaming and applications is much more relevant.

 

[deasd]

First week Raptorlake sales. 5800X3D sales surge at the same time of Raptorlake showcase, looks like the market just clearly tell manufacturer what to do next time...... people suddenly becomes budget gamers when we still argue about E core spam MT workload and efficiency.......

https://twitter.com/x/status/1584311412360646657

 

[Doug S]

Has there been any concrete info on when midrange / lower TDP SKUs of Raptor Lake will launch? I've been looking at building a new PC to replace my i5 Skylake, and since Zen 4 is stupidly going DDR5 only it looks like it'll be Raptor Lake by default. Hoping we see something beyond the high end 'K' parts soon.

 

[TheELF]

Has there been any concrete info on when midrange / lower TDP SKUs of Raptor Lake will launch? I've been looking at building a new PC to replace my i5 Skylake, and since Zen 4 is stupidly going DDR5 only it looks like it'll be Raptor Lake by default. Hoping we see something beyond the high end 'K' parts soon.

It's usually first quarter/early second quarter, so still a while to go.
Although I have a feeling that ddr5 will make a difference even with the lower end CPUs.

 

[Henry swagger]

Club386 finally did a review comparing 13900k vs 7950x in gaming using a proper RTX 4090. FYI Both systems used DDR5-6000 CL30 here.

Best CPU for Nvidia GeForce RTX 4090: Intel Core i9-13900K vs. AMD Ryzen 9 7950X | Club386

Which flagship CPU can handle the awesome might of GeForce RTX 4090? We identify a clear winner.

www.club386.com

The results at FHD resolution make things quite clear. When we are actually CPU bound in a game, the 13900k puts some pretty solid distance from the 7950x.
View attachment 69800
View attachment 69801

You an look through the review, things are a bit closer in some games and a bit further apart in others but the 13900K is a beast at gaming. AMD will certainly be needing the 3D cache models of Zen 4 to compete here. I suspect a Zen 4 3D vs 13900K (or future KS) with DDR5 7000+ will be pretty closely matched in gaming.

The 2mb l2 cahe is like 3dv cache for intel 😏.. zen 4 v cahe will be good too if it doesn't have clock regression like 5800x3d and hit 5.5ghz in gaming

 

[nicalandia]

Battle of Non-X/K 16 Threads parts....

13400 vs 7700

Default string X670E Taichi vs Acer Predator PO5-650 - Geekbench

 

[aigomorla]

Battle of Non-X/K 16 Threads parts....

13400 vs 7700

That is such a bad comparison.. i don't even know why its included.
Nothing has anything called a standard, in that test.

One is a OEM, the other is a high class overclocking board.
One is a upper mid class cpu, the other is a either a upper lower, or a lower mid.

Its really a horrible comparison, which feels like AMD PR trying a bit too hard.

 

[nicalandia]

That is such a bad comparison.. i don't even know why its included.

I see no Issues whatsoever.

Both CPUs are Non-X and Non-K, Both 65 Watt CPUs, both 16 Threaded CPUs, just like the 13900K vs 7950X which are both 32T CPUs.

Oh, I see it's only fair when Intel Wins?

 

[controlflow]

I see no Issues whatsoever.

Both CPUs are Non-X and Non-K, Both 65 Watt CPUs, both 16 Threaded CPUs, just like the 13900K vs 7950X which are both 32T CPUs.

Oh, I see it's only fair when Intel Wins?

Maybe superficially an interesting comparison but I suspect the 13400 isn't really a Raptor Lake and it doesn't have the extra cache and other tweaks made to the 13600k and higher end CPUs. The clocks on the 13400 are probably just too low even though the thread counts are similar. Also isn't the 7700x 105 watts?

Most importantly, the 7700x is going to cost probably close to 2x a 13400 and that's not even considering the more expensive AMD platform. I doubt these two CPUs will ever be cross shopped.

 

[nicalandia]

Also isn't the 7700x 105 watts?

Its The non-X version, The 65W 7700.... So fair comparison at same Thread Count and at same Power Envelope

16T vs 16T, 65W vs 65W

 

[Kocicak]

65W is not 65W and it is another "not 65W" on the other CPU. Numbers mean nothing nowadays.