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.

 

[lobz]

What we know is that a 12900KS running at 5.5 GHz (as evidenced by AdamK47's sample) with 6000 CL30 memory is faster in int but slower in fp than a 7950X with 6000 CL30 memory, but not by a huge amount. Raptor Lake is going to improve on that by a small but noticeable margin.

The fastest ST performance is still going to remain with Intel and Raptor Cove.

Never in my wildest dreams would I have guessed that one day I'm gonna meet a true prophet of the Lord on these forums. I feel honored beyond comprehension!

Please don't stop there, provide us with more of these unquestionable facts about the times to come, to which you seem to have divine access in advance.

Amen.

 

[lobz]

Hehe.

Giiven the fact that AMD is on a far superior node to Intel, and still looking to lose overall performance leadership on the desktop, I believe Intel has done a wonderful job.
Contained in this link is a 13900k at 5500 mhz vs a 7950x at 5738 mhz. Check out the scores.

https://twitter.com/x/status/1565265185064882176

It's so amusing to observe, how you appear to be an actually really smart person, yet you always choose to give in to selective ignorance and disingenuity so easily whenever it benefits the perception of your choice of brand.

 

[Zucker2k]

It's so amusing to observe, how you appear to be an actually really smart person, yet you always choose to give in to selective ignorance and disingenuity so easily whenever it benefits the perception of your choice of brand.

It's funny you say that because I provided a Twitter link to data that had been provided by @nicalandia without a link. Clarity is a bad thing now?
If I'm wrong in my analysis, feel free to point it out. That's what we're all doing here, extrapolating info from data. You have entered two posts this morning and none has contributed in any meaningful way to resolve the differences we all seem to observe between two very different architectures and their performances relative to each other.
So, I suggest you abandon the ad hominem and enlighten us on how this tight battle is going to unravel. You can use AMD slides if you want.
We're all waiting for reviews but before then, members are posting info, from the usual sources on Twitter, shedding light on how these chips are performing. It's strange that the only people accusing others of bias are the people from the AMD camp.

Edited for clarity.

Geekbench FP and INT Scores for 13900K and Ryzen 7950X

View attachment 66971

View attachment 66972

 

[coercitiv]

Folks, relax a bit, we're talking about such low margins that any numbers of factors can tilt the difference. As I said before, get used to the idea that ST performance of this year's chips will be within spitting distance. One wrong turn and somebody will get wet.

 

[exquisitechar]

Raptor Lake seems like it will have better ST performance, but the difference will be very small.

 

[biostud]

Raptor Lake seems like it will have better ST performance, but the difference will be very small.

Exactly, if it is single digit difference and especially low single digit, you will never know the difference in real life.

One thing is desktop and enthusiast builds. But for laptops, Intel is really going to be challenged.

 

[nicalandia]

Cinebench R23 ST and MT at Stock 250W

ST: 2290
MT: 35,693


At 345 Watts

ST: 2288
MT: 40616

 

[tamz_msc]

Cinebench R23 ST and MT at Stock 250W

ST: 2290
MT: 35,693
View attachment 67000

At 345 Watts

ST: 2288
MT: 40616

View attachment 67001

Aren't these like a month old, or even older? What's the point of regurgitating them now?

 

[tamz_msc]

Never in my wildest dreams would I have guessed that one day I'm gonna meet a true prophet of the Lord on these forums. I feel honored beyond comprehension!

Please don't stop there, provide us with more of these unquestionable facts about the times to come, to which you seem to have divine access in advance.

Amen.

You're usually quite funny with your jokes. This was an exception.

 

[Hulk]

It looks like Raptor Lake and Zen 4 are going to be incredibly competitive in much the same way ADL and Zen 3 are. Which to choose for most people will most likely come down to how well each brand performs on the most computationally intensive software the prospective buyer uses, how well a certain part fits the buyers budget, and brand loyalty.

 

[biostud]

Cinebench R23 ST and MT at Stock 250W

ST: 2290
MT: 35,693
View attachment 67000

At 345 Watts

ST: 2288
MT: 40616

View attachment 67001

And in a laptop @ 65W ?

 

[jpiniero]

And in a laptop @ 65W ?

Dragon Range should be pretty good.

 

[Asterox]

I looks like Raptor Lake and Zen 4 are going to be incredibly competitive in much the same way ADL and Zen 3 are. Which to choose for most people will most likely come down to how well each brand performs on the most computationally intensive software the prospective buyer uses, how well a certain part fits the buyers budget, and brand loyalty.

No it won't, CPU power consumption=or performanse per watt Zen 4 is much beeter.

Red is more important today than ever, given the global situation and the rise in electricity prices.The crisis is big, and let's hope that there won't be any much bigger problems.

Alder Lake=Raptor Lake, if you set same TDP as Zen 4 that would be very bad in any CPU benchmark.

 

[Zucker2k]

No it won't, CPU power consumption=or performanse per watt Zen 4 is much beeter.

Red is more important today than ever, given the global situation and the rise in electricity prices.The crisis is big, and let's hope that there won't be any much bigger problems.

Aldar Lake=Raptor Lake, if you set same TDP as Zen 4 that would be very bad in any CPU benchmark.

View attachment 67013

You're a bit late with your argument, don't you think? AMD certainly doesn't agree with you, and the latest consensus in these forums seems to be that it's the performance that matters. 230w vs 253w isn't as dramatic as 142w vs 241w. It seems clear being competitive trumps being power efficient at the cost of performance.

 

[igor_kavinski]

Aldar Lake=Raptor Lake, if you set same TDP as Zen 4 that would be very bad in any CPU benchmark.

I think that would be the fair way to benchmark. I wish all benchmark sites would standardize on that. They can put a section at the end for a "no limits" drag race sort of benchmarking. For people who don't care about safety or crashing into a wall or burning down their house.

 

[nicalandia]

What we know is that a 12900KS running at 5.5 GHz (as evidenced by AdamK47's sample) with 6000 CL30 memory is faster in int but slower in fp than a 7950X with 6000 CL30 memory, but not by a huge amount. Raptor Lake is going to improve on that by a small but noticeable margin.

The fastest ST performance is still going to remain with Intel and Raptor Cove.

We have enough data from at least 10 Runs to average the ST FP and INT On 13900/K/KF.

 

[Zucker2k]

We have enough data from at least 10 Runs to average the ST FP and INT On 13900/K/KF.

That's quite funny since you posted this in the other thread about a chip that's going to be released even earlier than RPL.

Guys. Lets not make a big deal out of unreleased CPU with Beta Bios. Even after release there is room for improvement.

View attachment 67005

Unless "room for improvement" applies only to Zen 4.

 

[lobz]

You're usually quite funny with your jokes. This was an exception.

Have you wondered why? Could it be coz it wasn't a joke after all, but rather a blatant rant trying to let you know how you really shouldn't just STATE things that nobody here could possibly verify and/or know for sure here, and even if they do, it means absolutely nothing without an explanation?

Maybe.

 

[Exist50]

Have you wondered why? Could it be coz it wasn't a joke after all, but rather a blatant rant trying to let you know how you really shouldn't just STATE things that nobody here could possibly verify and/or know for sure here, and even if they do, it means absolutely nothing without an explanation?

Maybe.

You really don't think it's possible to extrapolate results for Raptor Lake given known clocks? Come on now...

 

[nicalandia]

Average Score of 10 runs of 13900/K/KF(Found on GB by looking over 300 pages)

ST: 2,172 Points
Crypto Score: 5,000 Points
INT: 1,894 Points
FP: 2,298 Points

While 10 runs is statistically speaking a low number, but it's enough to average out the outliers(which are included here, but I am on it searching for more)

 

[MarkPost]

You're a bit late with your argument, don't you think? AMD certainly doesn't agree with you, and the latest consensus in these forums seems to be that it's the performance that matters. 230w vs 253w isn't as dramatic as 142w vs 241w. It seems clear being competitive trumps being power efficient at the cost of performance.

So a 13900K will consume the same that 12900K despite higher frec and specially 8 more e-cores? I dont think so.

 

[FangBLade]

So a 13900K will consume the same that 12900K despite higher frec and specially 8 more e-cores? I dont think so.

There is no way that 13900k will consume same or less while being on the same node and basically identical architecture (ok slightly changed). Higher clocks, more cache and 8 more e-cores will increase power consumption significantly, it is not situation like it was with Zen 2 to Zen 3 becase Zen 3 is complete redesign.

 

[Markfw]

There is no way that 13900k will consume same or less while being on the same node and basically identical architecture (ok slightly changed). Higher clocks, more cache and 8 more e-cores will increase power consumption significantly, it is not situation like it was with Zen 2 to Zen 3 becase Zen 3 is complete redesign.

Rather than argue with all these Intel supporters, I am just going to wait for reviews, just like for the 7950x. Its like talking to a brick wall until then, they have their mind made up.

 

[Exist50]

There is no way that 13900k will consume same or less while being on the same node and basically identical architecture (ok slightly changed). Higher clocks, more cache and 8 more e-cores will increase power consumption significantly, it is not situation like it was with Zen 2 to Zen 3 becase Zen 3 is complete redesign.

The higher clocks are for single core. Multicore scaling will come from the extra cores, even if all core speeds are slightly lower. And Intel can set the power limits to basically whatever. Rumors say it's only a few watts higher than the 12900k.

 

[MarkPost]

The higher clocks are for single core. Multicore scaling will come from the extra cores, even if all core speeds are slightly lower. And Intel can set the power limits to basically whatever. Rumors say it's only a few watts higher than the 12900k.

How is that posible if a 13700K (the same core count that 12900K) already consume more power than the aforementioned 12900K?

13700K Intel Core i7-13700K & i5-13600K Tests Reveal Performance Figures

My 12900K, all stock

So a few watts 13900K vs 12900K? yeah sure