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.

 

[Exist50]

Hahaha, that's hilarious. You want to think you are open minded. But then you shut down open debate by blocking users with different opinions and live in an echo chamber. I hope you enjoy it there.

While I'm not going to wholesale support the comparisons made, when certain users repeatedly demonstrate that they're debating in bad faith, the only rational response is to discount them entirely. That's less about conflicting opinions than it is the basic prerequisites for reasonable discussion.

 

[Thunder 57]

While I'm not going to wholesale support the comparisons made, when certain users repeatedly demonstrate that they're debating in bad faith, the only rational response is to discount them entirely. That's less about conflicting opinions than it is the basic prerequisites for reasonable discussion.

So I am discounting them because they are arguing in bad faith? I can see that. What I have yet to see is a reasonable discussion from said user.

 

[Thunder 57]

While I'm not going to wholesale support the comparisons made, when certain users repeatedly demonstrate that they're debating in bad faith, the only rational response is to discount them entirely. That's less about conflicting opinions than it is the basic prerequisites for reasonable discussion.

I know you personally have a beef with Intel which you've spoken of on this forum, and for that reason I normally don't even see your posts. With that kind of overt bias (hatred even) against a brand, I don't see you having anything of value to add.

To clarify, what I meant was them having put people on ignore. I don't put anyone on ignore because I welcome all input, whether I find it faulty or not. In this case this user chooses to put people on ignore that they do not agree with. That is what I was trying to say, though maybe not effectively I will admit.

 

[coercitiv]

On those types of apps, I think Zen 4 will still fall behind Alder Lake by 10% or more when comparing 7600X to 12600K or 7700X to 12700K.

I see. And for the sake of clarity, since you established comparison criteria based on SKU naming, what Intel CPUs will we use to compare against 7900X and 7950X?

Are we going to compare 7900X vs 13900K and leave the 7950X out?

 

[shady28]

To clarify, what I meant was them having put people on ignore. I don't put anyone on ignore because I welcome all input, whether I find it faulty or not. In this case this user chooses to put people on ignore that they do not agree with. That is what I was trying to say, though maybe not effectively I will admit.

I put people on ignore because they constantly thread crap, constantly use multiple logical fallacies (goalpost shifting seems to be the favorite here - first don't compare core counts that are different, now we're comparing on price, one has to wonder what the rules will be when Zen 4 lands) - and the rules are not enforced.

If I were to go into the Zen thread and act like you and a handful of others act, I'd get infractions and eventually banned.

This is contrary to what is evident here, supposed to be a thread about Raptor Lake.

 

[Exist50]

To clarify, what I meant was them having put people on ignore. I don't put anyone on ignore because I welcome all input, whether I find it faulty or not. In this case this user chooses to put people on ignore that they do not agree with. That is what I was trying to say, though maybe not effectively I will admit.

I tend to have the same approach, but I can understand why someone would be justified in outright ignoring a user. It's less a statement about opinions or alignment expressed, and more an expression of the belief that the user in question does not contribute anything of net value. It's the same reason why one might ignore any rumors from certain sources. Sure, there's nothing inherently preventing any source from contributing something of value, but some have demonstrated such a poor track record that it's simply not worth the time of day.

 

[Thunder 57]

I put people on ignore because they constantly thread crap, constantly use multiple logical fallacies (goalpost shifting seems to be the favorite here - first don't compare core counts that are different, now we're comparing on price, one has to wonder what the rules will be when Zen 4 lands) - and the rules are not enforced.

If I were to go into the Zen thread and act like you and a handful of others act, I'd get infractions and eventually banned.

This is contrary to what is evident here, supposed to be a thread about Raptor Lake.

I seriously doubt that. And remember, you are the one who crapped on Zen 3/4 in this thread and got where we are. I don't see any warnings or infractions.

 

[Markfw]

I put people on ignore because they constantly thread crap, constantly use multiple logical fallacies (goalpost shifting seems to be the favorite here - first don't compare core counts that are different, now we're comparing on price, one has to wonder what the rules will be when Zen 4 lands) - and the rules are not enforced.

If I were to go into the Zen thread and act like you and a handful of others act, I'd get infractions and eventually banned.

This is contrary to what is evident here, supposed to be a thread about Raptor Lake.

This all started because you said this "Frankly, Alder Lake is so far ahead of Zen 3 that Zen 4 is unlikely to do anything more than catch up to that 2021 chip. "

So, first you were talking about Alder lake in a raptor lake thread, but since you are inferring in that sentence that Raptor lake will be even faster, its "sort of" on topic. Everyone else was just responding to what is inherently WRONG with that sentence. And they are providing benchmarks to prove their point, including the below post. And even you have not argued the perf/watt issue I said Zen 3 wins in. It appears to be you against everyone else in this thread, at least the last 2 pages worth.

Talk about the pot calling the kettle black......

 

[Det0x]

The only thing Zen 3 beats Alder Lake at is rendering.

I wonder how many people reading this forum have drawn a wire frame image and created their own render outside of running benchmarks.

I use 7-zip quite often, how about you ?
Just a few pages back we went over this comparison..

 

[ZGR]

On the PugetSystems website for they include their score weightings for Photoshop:

https://www.pugetsystems.com/pic_disp.php?id=71544

If I were to cover up the CPU names, I am not sure if I could tell the difference. The competition is fierce right now. This is awesome for the consumer! But I am not a fan of using one source for this kind of data... Especially one using DDR5 4400, lol.

I put people on ignore because they constantly thread crap, constantly use multiple logical fallacies (goalpost shifting seems to be the favorite here - first don't compare core counts that are different, now we're comparing on price, one has to wonder what the rules will be when Zen 4 lands) - and the rules are not enforced.

If I were to go into the Zen thread and act like you and a handful of others act, I'd get infractions and eventually banned.

This is contrary to what is evident here, supposed to be a thread about Raptor Lake.

So why post video screenshots and say weird things like this?

"Everyone here talks about games it seems, but Alder Lake just annihilates Zen 3 on these kind of apps. IT's not even close. So a little 15% bump on Zen 4 won't even catch Alder Lake and we already know that Raptor is going to be able to throw more cores at these kinds of apps. That's to say nothing of cache, turbo, and microcode performance improvements. "

This statement would be true if you replace Alder Lake with Sandy Bridge and Zen 3 with FX! By all data we have seen from leaks and doing napkin math, Zen 4 and Raptor Lake will be quite competitive.

I am excited to see Raptor Lake's performance once undervolted and at a power limit below 200W.

 

[shady28]

I see. And for the sake of clarity, since you established comparison criteria based on SKU naming, what Intel CPUs will we use to compare against 7900X and 7950X?

Are we going to compare 7900X vs 13900K and leave the 7950X out?

As I said before, I don't care about the 7590X or 13900K. I'm not going to be buying an $650/$800 CPU from either vendor.

Do you plan to buy one of those? If so great, if not why do you care?

I also said I expect Zen 4 to do nothing but normalize against Intel SKU for SKU Alder Lake. Let's keep in mind, AMD has two high end SKUs, 7900X and 7950X. Intel has never really responded to the 7950X, and it's likely to be $200 more than the 13900K.

IMO, the more typical 7600X and 7700X will still lose not only to Alder Lake, but even more so to Raptor Lake.

 

[Hulk]

This is just flat WRONG. While Alder lake may have its strengths, they trade blows, except perf/watt, Then Zen 3 wins by a mile.

I have a 12700K but I agree 100% with this statement.

 

[Furious_Styles]

As I said before, I don't care about the 7590X or 13900K. I'm not going to be buying an $650/$800 CPU from either vendor.

Do you plan to buy one of those? If so great, if not why do you care?

I also said I expect Zen 4 to do nothing but normalize against Intel SKU for SKU Alder Lake. Let's keep in mind, AMD has two high end SKUs, 7900X and 7950X. Intel has never really responded to the 7950X, and it's likely to be $200 more than the 13900K.

IMO, the more typical 7600X and 7700X will still lose not only to Alder Lake, but even more so to Raptor Lake.

How can you know this when we don't have any credible reviews of the new AMD chips?

 

[shady28]

How can you know this when we don't have any credible reviews of the new AMD chips?

I said "IMO" = In my opinion. I never said "I know"

Having pointed that out, it is clearly implausible that the 7600X, a 6/12 CPU, will make up performance delta's like the one in the image below between the 5600X and its SKU competitor the 12600K. It would need a more fundamental overhaul, which it isn't getting.

In fact, it's highly unlikely that the 7600X and 7700X will even match the 12600K and 12700K. After the initial buying frenzy, and assuming AMD doesn't pull another stunt like they did in 2020 with the paper launch of Zen 3 and only making 1M CPUs into a market that consumes 300-350M cpus/year, 7600X/7700X will have to compete on price (by lowering theirs).

 

[Gideon]

AMD has two high end SKUs, 7900X and 7950X. Intel has never really responded to the 7950X, and it's likely to be $200 more than the 13900K.

The 5950X went down to 12900K level pretty soon after it launched. Rumored prices indicate 7950X will stay in the ballpark.

Intel mentioned during Q2 earnings that they will up prices by 20%. And judging by the performance leaks they can (and they have to, to improve their financials)

If the rumored AMD prices for the 7xxx series are true, the direct competitor to 13600K be 7800X, 7900X for the 13700K, etc.

Yes, sure, AMD wouldn't mind charging more, as the do on every opportunity, but let's not forget that for the majority of Ryzen's existence AMD was fine charging 200-250$ for 6 cores and 300-350$ for 8 cores. Discounting 5 series certainly hasn't hit their margins much.

 

[Markfw]

I said "IMO" = In my opinion. I never said "I know"

Having pointed that out, it is clearly implausible that the 7600X, a 6/12 CPU, will make up performance delta's like the one in the image below between the 5600X and its SKU competitor the 12600K. It would need a more fundamental overhaul, which it isn't getting.

In fact, it's highly unlikely that the 7600X and 7700X will even match the 12600K and 12700K. After the initial buying frenzy, and assuming AMD doesn't pull another stunt like they did in 2020 with the paper launch of Zen 3 and only making 1M CPUs into a market that consumes 300-350M cpus/year, 7600X/7700X will have to compete on price (by lowering theirs).



View attachment 66101

Why are you pitting a 6 core CPU (7600x) against an equivalent 8 core CPU 12600k (6+4, the 4 ~ 2 P cores). In your own chart the 5800x (8 core) beats the 12600k. Is that why ? to skew the results so Intel wins ?

 

[Abwx]

In fact, it's highly unlikely that the 7600X and 7700X will even match the 12600K and 12700K.



View attachment 66101

If we apply the 35% better perf claim from AMD then in this test the 7600X should be a few % slower than a 12600K while a 7700X should be 18% faster than a 12700K.

 

[shady28]

The 5950X went down to 12900K level pretty soon after it launched. Rumored prices indicate 7950X will stay in the ballpark.

Intel mentioned during Q2 earnings that they will up prices by 20%. And judging by the performance leaks they can (and they have to, to improve their financials)

If the rumored AMD prices for the 7xxx series are true, the direct competitor to 13600K be 7800X, 7900X for the 13700K, etc.

Yes, sure, AMD wouldn't mind charging more, as the do on every opportunity, but let's not forget that for the majority of Ryzen's existence AMD was fine charging 200-250$ for 6 cores and 300-350$ for 8 cores. Discounting 5 series certainly hasn't hit their margins much.

Right, most of that I agree with. I don't think we will see desktop SKUs going up 20% from MSRP though. I expect 13600K to be ~$350.

That means a 5700X would need to be ~$300. Except you can do this with DDR4 on Rocket Lake.

If we apply the 35% better perf claim from AMD then in this test the 7600X should be a few % slower than a 12600K while a 7700X should be 18% faster than a 12700K.

See chart above. There's been nothing to suggest Zen 4 will get 35% better overall performance boost whatsoever.

It's a lot more likely they are saying "Up to 35%" and are likely referencing something that gets a big boost from DDR5 vs Zen 3 DDR4. Not exactly a lie, just a marketing technique for unobservant consumers.

 

[DrMrLordX]

But Cinebench is about rendering, isnt it..?.

He's threadcrapping a Raptor Lake thread with Vermeer vs. Alder Lake comparisons. Don't expect too much out of this one.

Meanwhile, in the real world, Vermeer wins plenty of benchmarks vs. Alder Lake, especially at the same power consumption. And don't you dare take focus off that application to use something like a web browser while you're running Handbrake or something similar, or Win11 will put focus on your foreground application and effectively idle all your P cores. Woops. Intel better fix that behavior for Raptor Lake.

 

[Abwx]

It's a lot more likely they are saying "Up to 35%"

They say "more than 35%" in their slide and they give the precision that it s in Cinebench R23, so oddities like AES in GB are excluded as being the reason for the perf uplift..

 

[shady28]

They say "more than 35%" in their slide and they give the precision that it s in Cinebench R23, so oddities like AES in GB are excluded as being the reason for the perf uplift..

Right, so I assume you were all over that when Zen 3 came out and was compared to Intel Gen 10 correct.

Grey bar is my 10850K, blue is a 5800X, single thread. Zen 3 is barely faster in single thread without AES. The red ones, 10850K wins single thread :

 

[Kosusko]

If you enjoy playing Ashes of the Singularity, sure.

Intel Core i9-13900K Raptor Lake CPU Overclocked Up To 6.2 GHz, Over 65% Faster Than 12900K & 5950X

New Intel Core i9-13900K Raptor Lake CPU benchmarks show an insane 65% performance uplift over 12900K and Ryzen 9 5950X.

wccftech.com

The paradigm shift continues... is performance of big.LITTLE x86 processors.

 

[Gideon]

They say "more than 35%" in their slide and they give the precision that it s in Cinebench R23, so oddities like AES in GB are excluded as being the reason for the perf uplift..

Zen 3 real-world all-core clocks are quite low for the 5950X and 5900X (3.8 Ghz and 4.1 GHz respectfully, though lower models get around 4.4 Ghz).

Judging from leaked base-clocks (4.7 Ghz for worst-case AVX-512 loads), all-core clocks in the 5GHz range should be nothing special for 12-16 core 170W TDP AMD designs done on TSMC 5nm. For the upper end models that's a 20-25% uplift alone. Add the 8-10% IPC and you'll pretty much get the advertised 30-35% uplift.

For 7600X and 7700X it will be more problematic (since their 1T turbo is around 5.2-5.3 GHz it's not reasonable to expect much more than ~5GHz, and that only gives a 10-15% uplift in MT tasks)

I agree that the 6 core and 8 core Ryzen models will have a hard time against Raptor Lake.

They will all-but-guaranteed be behind in MT workloads (not catastrophically but noticeably). At best around the margin of error for ST task (but still lose stuff favouring large front-ends). To make matters worse Raptor Lake can still be bought with DDR4 for more budget-oriented rigs.

AMD really needs V-cache 7xxx series to carve a unique niche for them.

 

[Gideon]

Grey bar is my 10850K, blue is a 5800X, single thread. Zen 3 is barely faster in single thread without AES. The red ones, 10850K wins single thread :

I wouldn't read too much into these results. IMO these subtests really depend on the exact rig and even benchmark run. At least in earlier versions I managed to influence some of these by more than 20% just by playing around with memory timings on my 3700X.

I particularly wouldn't equal these very specific micro-benches to overall web use experience. Web is complex, there are no true passable benchmarks for it. Speedometer 2.0 is probably the best available, as it at least runs an actual (rather simple but still) SPA on different frontend frameworks, doing multiple runs, in an actual web browser. It still isn't nowhere near good enough to make any overarching claims, but it certainly paints a different picture and is somehat relevant (unlike octane, kraken, which measures nothing similar to actual web code):

And this is the area I do feel comfortable talking about, being a full stack developer for 15 years, out of which a good chunk I have optimised front-end code in different real-world web apps, to run well.

 

[shady28]

I wouldn't read too much into these results. IMO these subtests really depend on the exact rig and even benchmark run. At least in earlier versions I managed to influence some of these by more than 20% just by playing around with memory timings on my 3700X.

I particularly wouldn't equal these very specific micro-benches to overall web use experience. Web is complex, there are no true passable benchmarks for it. Speedometer 2.0 is probably the best available, as it at least runs an actual (rather simple but still) SPA on different frontend frameworks, doing multiple runs, in an actual web browser. It still isn't nowhere near good enough to make any overarching claims, but it certainly paints a different picture and is somehat relevant (unlike octane, kraken, which measures nothing similar to actual web code):

And this is the area I do feel comfortable talking about, being a full stack developer for 15 years, out of which a good chunk I have optimised front-end code in different real-world web apps, to run well.

Well, I got you beat. I've been a full stack developer for 30 years, started with DOS and Delphi and I learned Modula-2 in college. And a lot of what I do includes machine control.

And yes you can tune a system to do disproportionately well in any benchmark, but my chart shows that most of the Geekbench win was due to a 224% win in AES, with mixed results in other areas. And I was responding to someone implying that somehow this made it a poor benchmark to compare Raptor Lake to Zen 4. That's either ignorance on their part, or another goal post shifting excercise.

But yes, you and I are both using AES right now because you know, we're both using web browsers. That greatly impacts a users feeling of snappiness, which is why for example a Tiger Lake laptop can feel snappier than a 9900K or 10850K. This simple fact goes right over most people's head.

i.e. it was not me who was saying that AES was not important, it was implied by someone else.