Discussion Intel current and future Lakes & Rapids thread

[mikk]

I'd prefer to see all these done locked at 6000 mt/s.

Nobody cares what you prefer. They tested with the officiall supported RAM speeds which is fine. And why locked at 6000, because AMD doesn't work well above 6000? I can see you intention there. You are a little biased little guy right? You sound like a crybaby.

 

[A///]

Nobody cares what you prefer. They tested with the officiall supported RAM speeds which is fine. And why locked at 6000, because AMD doesn't work well above 6000? I can see you intention there. You are a little biased little guy right? You sound like a crybaby.

Locked at 6000 because 12th gen, 13th gen and Zen 4 can run those speeds. Even the worst Z690 board can handle 6000. By going with official speeds you're introducing biases whereas a flat memory speed across the plane levels the playing field.

 

[eek2121]

Not controlling for every variable, including memory speed and timings, leads to poor benchmarking results. Either test both AMD and Intel chips at supported (JEDEC) speeds/timings, or use comparable speeds/timings across both platforms. If you don’t, you end up testing the memory or the test rather than the CPU.

 

[A///]

Not controlling for every variable, including memory speed and timings, leads to poor benchmarking results. Either test bothAMD and Intel chips at supported (JEDEC) speeds/timings, or use comparable speeds/timings across both platforms. If you don’t, you end up testing the memory or the test rather than the CPU.

The issue with comparing at jedec speeds is that you're still delivering a bias. If the Intel IMC and it's L2 is better than it should be better even on an evened out plane. If the Zen processors are better it'll show up in that testing. You want to remove as many variables as you can and let the processors do their work. When you have obviously worse processors destroying Intel processors or having similar results regardless of generation you're fuzzying up the test results.

Set it all to 6000cl30 and let the processors speak for themselves. Test DDR4 separately from DDR5. Only then bring up how a DDR4 setup might be just as good as a DDR5 setup. Nobody in the enthusiast crowd in their right mind is running jedec speeds. If they are then someone ought to slip a banana under their feet.

 

[Abwx]

Nobody cares what you prefer. They tested with the officiall supported RAM speeds which is fine. And why locked at 6000, because AMD doesn't work well above 6000? I can see you intention there. You are a little biased little guy right? You sound like a crybaby.

From their remarks it s likely a scheduling issue, the game use few threads but they are balanced from one core to another, so there s surely threads that are switched from CCX1 cache to CCX2 cache, hence a slow down due to the inherent latency, once the issue is solved things will get back at their usual place.

 

[DrMrLordX]

Showing everything with default supported memory speeds may seem fair until you realize that practically nobody except maybe really unfortunate OEM buyers would run those speeds. Office PCs? Sure, but are you playing Starfield on those? Probably not.

 

[A///]

Showing everything with default supported memory speeds may seem fair until you realize that practically nobody except maybe really unfortunate OEM buyers would run those speeds. Office PCs? Sure, but are you playing Starfield on those? Probably not.

masochists

 

[Henry swagger]

Superior ipc always shows in demanding games.. not all games love a cache spamming

 

[deasd]

GameGPU result of Starfield shows a very different picture:

 

[itsmydamnation]

Superior ipc always shows in demanding games.. not all games love a cache spamming

this comment has aged well

 

[TESKATLIPOKA]

Superior ipc always shows in demanding games.. not all games love a cache spamming

Some games like more cache and some faster memory controller.
BTW, Raptor also received more cache compared to Alder.

 

[TESKATLIPOKA]

this comment has aged well

No, @Henry swagger aged well.
Even after 20 years, he will be remembered for his undying loyalty to Intel.

 

[A///]

He is aware Raptor has double the L2 $ over Zen 4, right?

Imagine in 2023 arguing against cache spamming when Intel spammed their L2 cache from Alder Lake and spams e-cores each major generation like someone's nan trying to shove werther's originals down her brassiere.

 

[A///]

Some games like more cache and some faster memory controller.
BTW, Raptor also received more cache compared to Alder.

Yep. 1.2 to 2 MB L2 $ for Alder to Raptor. Faster memory controller which shines above 8000 mt/s.

 

[biostud]

GameGPU result of Starfield shows a very different picture:

View attachment 85215

Neither 5800X3D or 7800X3D present?

 

[SiliconFly]

Great to hear you work street corners. It explains a lot.

Gameplay does not equal game performance. No storyline and mechanics makes up for buggy releases.

I don't. But I've seen you there plenty of times. You need help

(Note: Strongly recommend you avoid personal attacks in a public forum).


That member was already warned yet you retailiated. Insults are not allowed here.


esquared
Anandtech Forum Director



Stop removing mod edits from your infracted posts.


esquared
Anandtech Forum Director

 

[DrMrLordX]

GameGPU result of Starfield shows a very different picture:

That is peculiar. The previous test showing Raptor Lake with a healthy lead made very little sense, while this test seems bottlenecked by the GPU or the game engine. Hard to tell which.

 

[TESKATLIPOKA]

That is peculiar. The previous test showing Raptor Lake with a healthy lead made very little sense, while this test seems bottlenecked by the GPU or the game engine. Hard to tell which.

Why? If It's a different location, It could happen, no?

 

[mikk]

Kitguru confirmed that Intel unveils Meteor Lake mobile at Intel innovation and they can talk in detail about it including sample laptops in a consumer product, the press got briefed under NDA already. Unrelated to this news it seems Meteor Lake Ultra 9 tops out at 5.1 Ghz. This is a 300 Mhz/6% deficit to Raptor Lake based i9-13900H ST boost. It makes me wonder if the 8-9% RWC IPC uplift rumor is true because it would be enough to match or slightly beat ST from Raptor even with the 300 Mhz clock speed deficit.

 

[Tup3x]

Kitguru confirmed that Intel unveils Meteor Lake mobile at Intel innovation and they can talk in detail about it including sample laptops in a consumer product, the press got briefed under NDA already. Unrelated to this news it seems Meteor Lake Ultra 9 tops out at 5.1 Ghz. This is a 300 Mhz/6% deficit to Raptor Lake based i9-13900H ST boost. It makes me wonder if the 8-9% RWC IPC uplift rumor is true because it would be enough to match or slightly beat ST from Raptor even with the 300 Mhz clock speed deficit.

To be honest, ST performance never really was an issue so it's fine if that doesn't improve much. However, if efficiency would improve noticeably that would be very welcome.

 

[DrMrLordX]

Why? If It's a different location, It could happen, no?

Possibly. Would be nice if there were some uniformity in testing but we know review sites don't often do that.

 

[lightisgood]

Intel Core i9-14900K 6.0 GHz ES CPU shows up in first CPU-Z test, 8-10% faster than Core i9-13900K - VideoCardz.com

Intel Core i9-14900K tested in CPU-Z Intel’s fastest LGA1700 CPU makes its first appearance. The upcoming Core i9-14900K processor, part of Intel’s Raptor Lake Refresh lineup, is set to maintain the same core count as the existing Core i9-13900K with 24 cores, consisting of 8 P-cores and 16...

videocardz.com

14900K beats 13900KS.
Possibly, RPL-R is made by Intel 6 (not Intel 7 Ultra).

 

[Henry swagger]

He is aware Raptor has double the L2 $ over Zen 4, right?

Imagine in 2023 arguing against cache spamming when Intel spammed their L2 cache from Alder Lake and spams e-cores each major generation like someone's nan trying to shove werther's originals down her brassiere.

Its not l2 cache only .. raptor is on a better process node plus 512+ rob size

 

[eek2121]

Raptor Lake refresh looks to increase performance by 5-10% for the top SKUs, and even more for lower SKUs. We won’t discuss power, however. 😬

Curious as to how the competition reacts.

Kitguru confirmed that Intel unveils Meteor Lake mobile at Intel innovation and they can talk in detail about it including sample laptops in a consumer product, the press got briefed under NDA already. Unrelated to this news it seems Meteor Lake Ultra 9 tops out at 5.1 Ghz. This is a 300 Mhz/6% deficit to Raptor Lake based i9-13900H ST boost. It makes me wonder if the 8-9% RWC IPC uplift rumor is true because it would be enough to match or slightly beat ST from Raptor even with the 300 Mhz clock speed deficit.

The are “P” dies. Even if Intel represents them as “H” dies (unclear), the cores were designed for a much lower power limit (55W vs 28W). The i7 1370p had a 5.2 ghz clock, which is 100mhz higher.

Meteor Lake likely has an IPC increase at play to more than compensate for this.

 

[FangBLade]

It's a bit funny for me to read this story about the superior IPC at RPL, mostly pushed by characters known for trolling, not changing their style. Look at this benchmark, Baldur's Gate 3, a highly acclaimed game, received top scores, and the 5800x3d outperforms the 13900k, not to mention the 7800x3d: https://www.pcgameshardware.de/Bald...rk-Test-Review-Grafik-Ersteindruck-1425826/2/ By the way, I've already written to you about Henry, he's just trolling you, he's not an Intel fan, keep that in mind, and adapt, he'll change his style very soon when he realizes it's pointless.