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.

 

[dullard]

Kinda miffed that I bought into LGA1700 partially on the hope of future platform updates, but it looks like they're pretty meaningless. Not sure if even going from 12700k to 14900k would make any meaningful difference for mostly gaming.

Am I missing anything?

You are missing that future platform updates are generally a bad idea for people to factor into their decision. Either (A) the CPUs aren't worth the update or (B) you massively hamper the new CPU with old memory speeds, old PCI lanes, complete lack of new technology, etc. This is especially true for gaming, where upgrading the GPU is usually a far better deal than putting an expensive new CPU into an old system.

 

[H433x0n]

Once all the dust is settled and people are no longer emotionally or financially-invested in Intel 10nm/Intel 7 releases, it would be interesting to see just how well a fully-tuned 12900k on Z690 fares against a 14900k on Z790 in a suite of benchmarks.

I agree that the difference between 13900K/13900KS/14900K is negligible and not worth bothering with.

However, there's no doubt that a 14900K would absolutely outperform a well tuned 12900K, this is true even at iso core counts (comparing 14900K w/ only 8P+8E enabled). There's a real difference between Alder Lake & Raptor Lake.

 

[DrMrLordX]

However, there's no doubt that a 14900K would absolutely outperform a well tuned 12900K, this is true even at iso core counts (comparing 14900K w/ only 8P+8E enabled). There's a real difference between Alder Lake & Raptor Lake.

To be honest I'd just like to see by how much, using the same software (rather than scraping up benchmarks from 1-2 years ago).

 

[Hulk]

Sorry but that's bunk, on multiple levels. Two main ones:

1). Nobody asked Intel "for years" to release the same part over and over again, just at a higher TDP.
2). These parts aren't so much unlocked as they are overclocked-from-the-factory. These things are nearly pushed to their limits out-of-the-box, and many users are being forced to learn to lower their power consumption by hand rather than being forced to learn to overclock them.

Their competition seems to be doing just that, so . . .

If that were really true, shouldn't they just cut the price on the 13900ks and be done with it? Don't lie to people and pretend like it's a fundamentally new part.

You seem to have confused my statement with someone else's.
Point 1 - I never said/wrote that they did.
Point 2 - Yes they are. So? Again what does my post have to do with that?

Unfortunately Intel doesn't make money from sales of their competition. Intel has to offer a better product than it's CURRENT product to persuade people to upgrade. ST on the 14900K is a high bar.

You seem to be more concerned about the fact that the Refresh is a relatively insignificant roll out. I agree. But that's not what I was referring to in my post.


I don't know if you were overclocking in the early '90's when chips were unlocked but the community was pissed off when Intel began locking parts and all we could do was change FSB frequencies. I like having full control with the K parts but with great power comes great responsibility!

 

[Hulk]

To be honest I'd just like to see by how much, using the same software (rather than scraping up benchmarks from 1-2 years ago).

Benchmarks aren't "scraped up." They are tightly written code that is very repeatable believe it or not. There is no "magic" 12900K that is going to outperform one from 3 years ago (beyond normal statistical variation). A stock 14900K will beat a stock 12900K every time in every benchmark with equal power and cooling. That goes for both ST and MT. That is absolutely irrefutable.

If you want to get into a pointless debate about the meaning of the word "beat" then sorry I don't have the time. Point is a 14900K is more performant than a 12900K. Period. There is no walking back your quote above. It is complete and absolute bunk!

 

[H433x0n]

Their competition seems to be doing just that, so . . .

We don’t know the performance of Zen 5.

While I’m sure it’ll outperform RPL we’d be comparing a product that debuted in 2022 against AMD’s latest offering from 2024 that’s launching 18+ months after RPL’s debut.

This isn’t actually a new product release.

 

[CropDuster]

You are missing that future platform updates are generally a bad idea for people to factor into their decision. Either (A) the CPUs aren't worth the update or (B) you massively hamper the new CPU with old memory speeds, old PCI lanes, complete lack of new technology, etc. This is especially true for gaming, where upgrading the GPU is usually a far better deal than putting an expensive new CPU into an old system.

No doubt, guess I was just hoping 1700 would be Intel's AM4 moment. I should've known better.
My old X58 system going to a Xeon on the cheap was pretty sweet though.

I agree that the difference between 13900K/13900KS/14900K is negligible and not worth bothering with.

However, there's no doubt that a 14900K would absolutely outperform a well tuned 12900K, this is true even at iso core counts (comparing 14900K w/ only 8P+8E enabled). There's a real difference between Alder Lake & Raptor Lake.

I wish more reviews were comparing 12th gen to 14th gen, most seem to just be focused on 13 vs 14. Maybe a 14900k will be a worthwhile upgrade to my 12700k once next gen releases. Or say screw it and go X3D. Will have been 20 years since I had an AMD rig at that point.

 

[Det0x]

A stock 14900K will beat a stock 12900K every time in every benchmark with equal power and cooling. That goes for both ST and MT. That is absolutely irrefutable.

If you want to get into a pointless debate about the meaning of the word "beat" then sorry I don't have the time. Point is a 14900K is more performant than a 12900K. Period. There is no walking back your quote above. It is complete and absolute bunk!

Dont want to throw gasoline on the fire but a older 12900k can beat all newer gen 13 and "14" raptor lake cpus in y-cruncher thanks to its ability to enable avx512

y-cruncher - Pi-2.5b overclocking records @ HWBOT

Overclocking records

hwbot.org

 

[blckgrffn]

No doubt, guess I was just hoping 1700 would be Intel's AM4 moment. I should've known better.
My old X58 system going to a Xeon on the cheap was pretty sweet though.


I wish more reviews were comparing 12th gen to 14th gen, most seem to just be focused on 13 vs 14. Maybe a 14900k will be a worthwhile upgrade to my 12700k once next gen releases. Or say screw it and go X3D. Will have been 20 years since I had an AMD rig at that point.

I dunno, pretty similar in stepping up from a 8700K (which was solid) to a 10700k (which was also fine) if they were on the same chipset. Doesn’t seem that exciting to me but yeah, some benchmark numbers go up. Seat of the pants feel with modest tuning of both systems? Probably extremely minimal difference.

And I say that having had a soft spot for 10th gen CPUs. My dad’s 10850k build that happened due to Zen 3 launch shenanigans that at least Pepperidge Farm remembers is a solid rig even now, but part of that was the availability of fast DDR4 and nvme storage compared to previous generations.

 

[DrMrLordX]

Benchmarks aren't "scraped up." They are tightly written code that is very repeatable believe it or not.

Let me try again: I would like someone to run the benchmark software on a 12900k or ks system, fully tuned, in 2024 or later. Not using benchmarks results from 2022. Is that more clear?

There is no "magic" 12900K

Let me just stop you right there and tell you that I am curious about the results only for curiosity's sake. In no way am I predicting results. I just want to see them. Calm down, Beavis.

We don’t know the performance of Zen 5.

Depending on whom you believe, it's anywhere from 15%-32% faster than Zen4 1T. Kind of a wide margin but the point is, one company is making advancements @isopower while another one isn't.

This isn’t actually a new product release.

And whose fault is that?

 

[igor_kavinski]

I wish more reviews were comparing 12th gen to 14th gen

To be honest I'd just like to see by how much, using the same software (rather than scraping up benchmarks from 1-2 years ago).

*Arnold Shwarzenburger voice*

Buy the 14900K. Go. NOW!

 

[dullard]

View attachment 87515

At stock, that is a whopping 14.4% gaming boost from the 12900K to 14900K. That is why platform CPU upgradability is so laughable of a concept--especially for gaming. Buy a platform that will last you several years with the original CPU you buy and then put the upgrade money into a GPU instead.

 

[igor_kavinski]

Intel Core i9-14900K

This 'Raptor Lake Refresh' is mostly a reheated i9

me.pcmag.com

Just be sure to rein in the power limits or 14900K will give PSUs a serious workout.

 

[igor_kavinski]

Intel Core i9-14900K review

It might be the end of the road for LGA1700, but the Intel Core i9-14900K certainly packs a mighty punch, taking the fight to the best of AMD Zen 4.

www.custompc.com

Look, 14900K consumes a whole 3W less than 13900K!

 

[igor_kavinski]

Almost no advantage to using DDR5-7800 with 14900K.

Intel Core i9-14900K CPU Review

The Intel Core i9-14900K offers a value increase with faster performance than the 13900K but at the same price level, it's fast, but that power demand...

www.thefpsreview.com

Anyone got benchmarks with DDR5-8800?

 

[DrMrLordX]

Buy the 14900K. Go. NOW!

1). That's a stock 12900k, which as we all ought to know by now, defaults to stupidly-low RAM speeds compared to the 13900k and also sports lower clockspeeds.
2). That includes some MT scores which hurts the 12900k due to its lower core count

I have my reasons for wanting to see fully-tuned 12900k vs 13900k, 13900ks, and/or 14900k but again people are really emotionally invested in loving/hating current-gen Intel products right now. Once Arrow Lake is on the scene we can do some actual juxtaposition without people immediately becoming upset about everything.

 

[Hulk]

1). That's a stock 12900k, which as we all ought to know by now, defaults to stupidly-low RAM speeds compared to the 13900k and also sports lower clockspeeds.
2). That includes some MT scores which hurts the 12900k due to its lower core count

I have my reasons for wanting to see fully-tuned 12900k vs 13900k, 13900ks, and/or 14900k but again people are really emotionally invested in loving/hating current-gen Intel products right now. Once Arrow Lake is on the scene we can do some actual juxtaposition without people immediately becoming upset about everything.

Only thing I'm emotionally invested in regarding this top is logic and truth. 13900K is more performant than the 12900K

 

[DrMrLordX]

Only thing I'm emotionally invested in regarding this top is logic and truth. 13900K is more performant than the 12900K

Then you shouldn't have anything to worry about, should you?

 

[Abwx]

Intel Core i9-14900K review

It might be the end of the road for LGA1700, but the Intel Core i9-14900K certainly packs a mighty punch, taking the fight to the best of AMD Zen 4.

www.custompc.com

Look, 14900K consumes a whole 3W less than 13900K!

View attachment 87527

8-10W less at Computerbase, and remember, max TDP is 125W...

Intel Core i9-14900K, Core i7-14700K and Core i5-14600K Review: Raptor Lake Refreshed

www.anandtech.com

 

[Markfw]

8-10W less at Computerbase, and remember, max TDP is 125W...

Intel Core i9-14900K, Core i7-14700K and Core i5-14600K Review: Raptor Lake Refreshed

www.anandtech.com

do we have real benchmarks with actual power usage ?

 

[Abwx]

do we have real benchmarks with actual power usage ?

There is at Computerbase, they added a few measurements for the 14700K at different power levels since 8 + 12 is new configuration.

Nothing for the 14900K set apart stock and power unlocked, either they will add them later or they think that it would provide the same results as the 13900K wich was tested extensively power wise.

On the graphs click on +1xx Eintrage to display all CPUs tests.

Intel Core i9-14900K, i7-14700K & i5-14600K im Test: Anwendungs-Benchmarks

i9-14900K, i7-14700K & i5-14600K im Test: Anwendungs-Benchmarks / Testmethodik / Multi-Core-Leistung / Single-Core-Leistung

www.computerbase.de

 

[igor_kavinski]

https://openbenchmarking.org/result/2310102-NE-2303111SA73

14900K has a pretty low context switching overhead. So it should feel "smoother" and maybe be a bit better at multitasking but I'm sure it would get bogged down whenever the task switches to an E-core. Really wanna see how much context switching latency is improved with E-cores/HT disabled.

And will someone please tell Ganesh to do his multitasking overhead benchmarks with desktop CPUs??? It's like he has some kind of contract clause that he will ONLY do that benchmark on mobile CPUs and if his services are required for desktop CPU benchmarking, he will charge an arm and a leg or something. Annoys the hell out of me

 

[nicalandia]

As I mentioned long ago, the lower SKUs of 14th Gen will be Alder Lake dies

ASRock Z790 Taichi - Geekbench

Benchmark results for an ASRock Z790 Taichi with an Intel Core i3-14100 processor.

browser.geekbench.com

 

[DrMrLordX]

As I mentioned long ago, the lower SKUs of 14th Gen will be Alder Lake dies

That is sad.

 

[coercitiv]

That is sad.

It's sad but doesn't change the outcome of what is happening in that price class. The 4-core i3 is in a very awkward place where purchase recommendations would steer towards either 12400F for gaming or 4600G for a general purpose build. (this applies to OEM build options too)

I assume this tier will receive a bump in computing power next year. From what I understand, Arrow Lake will lack SMT, so I hope they'll bump the i3 tier to 6c/6t or give it some freE-cores.