Intel Planning P-core Only "Bartlett" LGA1700 Processor for 2025

[FoulOnWhite]

I wonder if Bartlett will end up coming before 25. After the 13/14th gen troubles as a kinda look at the hand thing

 

[Wolverine2349]

I wonder if Bartlett will end up coming before 25. After the 13/14th gen troubles as a kinda look at the hand thing

I wonder too. Would be great if they release the 12 P core version soon. Only if it can be stable and not degrade. I will be happy with 5GHz all core and ring 4.6GHz and 7000 DDR5 rock stable all work loads no degradation and lasts a long time continuing to be rock stable if they can do that and release it in 2024?

 

[Wolverine2349]

Oh you absolutely do need more than 8 for gaming. Check the new game once human for example, it absolutely hammers 8core chips

Is that really true in any areas of that game Human?

Its weird given Once Human only minimum requires an i5 4460 and recommends an i7-7700 both of which are only 4 cores one with better IPC and HT being the 7700. So it really struggles with 8 and HT??

What about other games. Any need or big benefit from more than 8 cores even in 4K?

I have heard so many conflicting stories on this?

 

[Wolverine2349]

Based on some very surface level googling it appears that the game has a problem with real-time rendering of shaders. That would, aside being a massive fail for the devs, work just as fine on the E-cores.

I think fevgatos was just making a general assumption that you need more than 8 cores for gaming. Someone stating they did not know what the point of the chips were. I mean yes the e-cores can handle games that need more cores and do not have the scheduling issues just fine.

But many gamers do not want hybrid or dual CCD setups so stick to 8 P core only models with SMT/HT (7800X3D, 7700X, 5800X3D) or Intel 12th to 14th Gen with e-cores disabled) and if a 12 P core came about, many gamers who want no scheduling or heterogenous or dual CCD would jump on it with such a game example of desiring more than 8 cores.

But I have not heard anything since about this Bartlett Lake so not sure how much legs this rumor has.

 

[OneSorcerer]

I wonder if Bartlett will end up coming before 25. After the 13/14th gen troubles as a kinda look at the hand thing

10 p core would be a nice uplift from my 12600kf *.*

 

[iameatingjam]

Now if they release BTL with at least 48MB L3 Cache, they're aiming at the upcoming 9800X3D.

They might wanna increase that amount to 64MB if they'd rather be safe than sorry, although they've always acted skimpy when it comes to the Cache capacity. They needed AMD for the push!

I... Don't see how that would be possible. Don't forget... This is the same node. There's even variants with ecores that use that same die (b0) at least according to the rumours ( So essentially another refresh). In all likelihood they are just switching out the ecore clusters for pcores as they are about the same size. How are they gonna double the l3 cache? I mean hardware unboxed even showed that with Intel's architecture, cache isn't even the bottleneck in games.

 

[OneSorcerer]

I... Don't see how that would be possible. Don't forget... This is the same node. There's even variants with ecores that use that same die (b0) at least according to the rumours ( So essentially another refresh). In all likelihood they are just switching out the ecore clusters for pcores as they are about the same size. How are they gonna double the l3 cache? I mean hardware unboxed even showed that with Intel's architecture, cache isn't even the bottleneck in games.

Any room for cache increase/improvements?

 

[Dr. Dro]

Any room for cache increase/improvements?

Looks like it's based on Raptor Cove (improved Golden Cove design, so very unlikely. The extra four P-cores occupy a space similar to the sixteen E-cores on the current i9's, so there is not much room for improvement there.

Still, if Intel releases a 12P CPU with all AVX-512 extensions enabled, it'll make for a fine Core non-Ultra Series 2 processor line.

 

[OneSorcerer]

Looks like it's based on Raptor Cove (improved Golden Cove design, so very unlikely. The extra four P-cores occupy a space similar to the sixteen E-cores on the current i9's, so there is not much room for improvement there.

Still, if Intel releases a 12P CPU with all AVX-512 extensions enabled, it'll make for a fine Core non-Ultra Series 2 processor line.

ty, it will still be a worthy upgrade

 

[iameatingjam]

Looks like it's based on Raptor Cove (improved Golden Cove design, so very unlikely. The extra four P-cores occupy a space similar to the sixteen E-cores on the current i9's, so there is not much room for improvement there.

Still, if Intel releases a 12P CPU with all AVX-512 extensions enabled, it'll make for a fine Core non-Ultra Series 2 processor line.

I totally didn't even think about that. Maybe this will end up rocketing to the top of the RPCS3 cpu tier list.

 

[Wolverine2349]

Is there any confirmed report that a 12 P core variant on a ring bus and/or Bartlett Lake in general is actually even coming. There has been like no news at all since this. Its so quiet??

 

[Wirko]

Is there any confirmed report that a 12 P core variant on a ring bus and/or Bartlett Lake in general is actually even coming. There has been like no news at all since this. Its so quiet??

Intel isn't going to develop a new chip just for a small part of the DIY market, it's not profitable enough. But... Maybe they see a market niche in low end servers and think they can fill it with a 12 P-core CPU. They send the chip to enthusiasts first as a Core (packaged in black) for a few months of field testing, then if everything is OK, launch it as a Xeon too.

 

[Wolverine2349]

Intel isn't going to develop a new chip just for a small part of the DIY market, it's not profitable enough. But... Maybe they see a market niche in low end servers and think they can fill it with a 12 P-core CPU. They send the chip to enthusiasts first as a Core (packaged in black) for a few months of field testing, then if everything is OK, launch it as a Xeon too.

Is that why we are not hearing more news on it. Is it that they do not intend to do it. It was only rumor.

Or maybe the small server segment that wants a ring bus and AVX512 than package it as enthusiast :LGA 1700 chip as a beta test run for us gamers niche who want more than 8 P cores and all such cores on one die of homogenous arch and if it works good produce more for small server and network edge market.

Maybe they intend to use us as a beta test run which could mean even worse problems when they milk us as we run out to buy it the chip we so waited for that never was made and they scrap it and hang up on our RMA requests when it degrades like a paper tiger and has stability issues like the current 8 + 16 chips do. And I am skeptical even of the microcode update.

But a new 12 + 0 die on a ring bus maybe its good. I may be wiling to take that risk.

 

[Dr. Dro]

Intel isn't going to develop a new chip just for a small part of the DIY market, it's not profitable enough. But... Maybe they see a market niche in low end servers and think they can fill it with a 12 P-core CPU. They send the chip to enthusiasts first as a Core (packaged in black) for a few months of field testing, then if everything is OK, launch it as a Xeon too.

Bartlett is supposed to be a chip meant for edge computing, developed by Intel NEX - but since there is clearly interest in LGA 1700, I believe Intel should make the effort to release Bartlett - particularly considering they could use the profits from shifting more chips and as long as that is executed right - they'll be very easily pardoned by the Raptor blunders

 

[64K]

My take is that Intel is scrambling to recover right now. This is just another thing that's up in the air. It would be interesting to me at least to see how these CPUs could perform in gaming but I admit my faith in Intel is a bit shaken right now and the 7800X3D is a great CPU for gaming anyway if nothing comes from this all P-core chip in the end.

 

[iameatingjam]

My take is that Intel is scrambling to recover right now. This is just another thing that's up in the air. It would be interesting to me at least to see how these CPUs could perform in gaming but I admit my faith in Intel is a bit shaken right now and the 7800X3D is a great CPU for gaming anyway if nothing comes from this all P-core chip in the end.

I don't even get why it would be beneficial for gaming, I mean maybe a little in some cases where the game scales really well with cores.

But we know most games are made for consoles with 8 cores and the best PC gaming cpu has 8 cores. It will certainly do worse in multi-threaded performance. Still... A lot of people seem to really not like ecores so that alone might sell a good amount of chips.

I'm sure there's some niche use cases where it would be beneficial, otherwise why would they make it? I heard networking, but not sure about that.

 

[irsh]

I actually applaud this if it gives me an upgrade path from a possibly prematurely degraded 14th gen i7. Now add X3D style mega cache please.

 

[iameatingjam]

I actually applaud this if it gives me an upgrade path from a possibly prematurely degraded 14th gen i7. Now add X3D style mega cache please.

Apparently intel's architecture doesn't benefit as much from L3 cache as AMD. Dunno why. I mean one reason is the better memory controller but that can't be all of it, otherwise, they would be as fast as the 7800x3d in gaming just from the ram. I'm thinking perhaps the whole ecore thing is a way to slowly redesign a main core from the ground up while keeping the old cores around to keep performance up while the new cores are incubating. And if thats true, maybe they will be more receptive to cache increases? IDK. But intel's gotta do something to respond to x3d.

 

[JustBenching]

they would be as fast as the 7800x3d in gaming just from the ram.

They are, tuned ram 14900k is faster than the 7800x 3d.

 

[iameatingjam]

They are, tuned ram 14900k is faster than the 7800x 3d.

In gaming? Consistently? Like you can show a random 20 - 40 game mean where the 14900k does better? I know RL has a few games that prefer its architecture. And sure it can get some benefit out of better memory speed/timings. But to surpass the 7800x3d in gaming on the whole from memory tuning? Color me.... skeptical.

 

[JustBenching]

In gaming? Consistently? Like you can show a random 20 - 40 game mean where the 14900k does better? I know RL has a few games that prefer its architecture. And sure it can get some benefit out of better memory speed/timings. But to surpass the 7800x3d in gaming on the whole from memory tuning? Color me.... skeptical.

Against a 7800x 3d, yes, against a tuned 9800x 3d, not sure, I think the 9800x 3d is faster but not by much.

 

[Wolverine2349]

Against a 7800x 3d, yes, against a tuned 9800x 3d, not sure, I think the 9800x 3d is faster but not by much.

Like what kind of memory tuning? And HT on or off?

I like Raptor Lake arch. Problem is the B0 8 + 16 stepping seems to degrade so so fast and/or is not stable. The microcode update not sure if its a real fix or a band aide. And if it is a fix here and now how long will they last without degrading in 6 months or 1 year.

And I have seen you absolutely need a non Asus 2 DIMM board to run XMP RAM reliable or otherwise OCCT Large Dataset variable will randomly throw WHEA errors or CPU core errors with XMP turned on on any 4 DIMM board even with onk1 1DPC (2 DIMMs total).

But to my other comment above I have had other stability problems even on 2 DIMM board not related to RAM like weeks down the road running CIbnebench a random WHEA error or a random BSOD or WHEA shader compilation even with an undervolted underclocked CPU

I see you have a 12900K as your system in your profile. That is a more stable chip and can that be tuned to match or beat 7800X3D in gaming. You get extra cores though only 8, but with HT on 24 threads and it does not have the degradation/stability issues 13th/14th Gen have. Benchmarks seem to suggest no, but wonder if that is because ring clock is crippled to 3600MHz with e-cores on?? With e-cores off you can push ring to close to 5GHz stable I think. But then e-cores off is a waste on that chip for gaming and miswell go 7800X3D/9800X3D if gonna do that which I would not.

Though I mention that because I think you can manually get the ring clock to 4200MHz or 4300MHz (But I doubt any more lol maybe if super lucky 4400MHz) stable on 12900K with e-cores on and wonder if that would shoot gaming performance up a lot with well tuned RAM while having those extra cores handy as well without having to go the 13th/14th Gen route and deal with the stability degradation fiasco.

I would like more than 8 cpores, but the options for it suck. Raptor Lake would be best bet if it were not for degradation and stability, but its there so its out. Maybe tuned Alder Lake???

Arrow Lake is an absolute flop.

7950X3D and 9950X3D both only have 1 CCD with 3D vcache and scheduling issues are bad as the XBOX Game bar parks all non 3D cores so the games that can use more threads cannot because only 8 3D cache cores can be used while gaming so less important threads cannot be sent of to other CCD that route. Intel 12th to 14th Gen do not have this problem as thread director is good at hardware level. So 12th Gen maybe given 13th and 14th Gen degradation/stability fiasco.

or otherwise be stuck with only 8 cores 7800X3D/9800X3D and get hampered in Cities Skylines 2 with large populations.

Sucks options are so limited.

And AMD no not gonna have 3D cache on both CCDs on 9950X3D unfortunately:

AMD Revealing Ryzen 9 9950X3D And 9900X3D Second Week In January

There's a huge amount of interest in AMD's Ryzen 9000X3D models with the Ryzen 7 9800X3D selling out in many areas and the highly anticipated Ryzen 9 9900X3D and 9950X...

www.forbes.com

 

[iameatingjam]

I see you have a 12900K as your system in your profile. That is a more stable chip and can that be tuned to match or beat 7800X3D in gaming.

WHAT?
Seriously am I in upside down land right now....? Why is everybody talking about alder lake and raptor lake beating the 7800x3d? I know in some select titles there is an advantage, but with the 12900k, even if its more, stable, its definitely even slower than the 14900k, by not an insignificant amount, and even the 14900k cannot beat the 7800x3d unless we are cherry picking games.

 

[Wolverine2349]

WHAT?
Seriously am I in upside down land right now....? Why is everybody talking about alder lake and raptor lake beating the 7800x3d? I know in some select titles there is an advantage, but with the 12900k, even if its more, stable, its definitely even slower than the 14900k, by not an insignificant amount, and even the 14900k cannot beat the 7800x3d unless we are cherry picking games.

Not talking about it being faster. Certainly not in lightly threaded games.

Its about options of wanting more than 8 cores for a gaming rig for now and future (without upgraidng core system, but just GPUs for go to options)

AMD has the dual CCD cross latency penalty and the weird scheduling issues esopeically with dual CCD chips with one CCD with 3D cache.

Intel has a very good thread diretcor 12th to 14th Gen, but still a hybrid arch. But their thread director hardware level is very goiod and works well.

So 13th and 14th Gen seem like best choice if you wnat more than 8 cores and tune them right?? One problem. Degradation/Stability fiasco.

So that leaves Intel 12th Gen which does not have degradation and stability issues. And 13th Gen was just a refined 12th Gen so maybe 12th Gen can be configured to be almost or just as fast? Or if not no good go to options of more than 8 cores that are reliable and will age well.

And yeah I know HW Unboxed did a video showing how some thought 3950X would age better than 5800X3D because it had 16 cores and future videos showed that was so false in gamning benchmarks as 5800X3D smokes it in todays games. However lets not forget the 3950X is like 4 4 core CPUs glued together and has those weak IPC Zen 2 cores. Zen 2 to Zen 3 was such a big IPC and latency improvement.

Its not unreasonable (though only time will tell) to think Intel 12th Gen, 13th and 14th Gen (if only they were reliable and did not degrade) and 16 core Zen 4 and Zen 5 and maybe even Zen 3 parts will age better than even 8 core Zen 4 and Zen 5 X3D counterparts?

Of course more than 8 homogenous cores on a single die would be best with at least Zen4 X3D/Golden/Raptor Cove IPC, but no such option exists and appears will not a long while if ever.

That means no CPU best for all around games like Cities Skylines 2 and others

 

[chrcoluk]

WHAT?
Seriously am I in upside down land right now....? Why is everybody talking about alder lake and raptor lake beating the 7800x3d? I know in some select titles there is an advantage, but with the 12900k, even if its more, stable, its definitely even slower than the 14900k, by not an insignificant amount, and even the 14900k cannot beat the 7800x3d unless we are cherry picking games.

I would assume the same applies to AMD though, tune the ram and it also gets gains, @ir_cow can offer insight on that one.

My 13700k gets better figures than what reviewers got and not within error of margin, typically 10-20% better. Assuming its my RAM config. Also could be down to that I tuned the power scheduler, and have custom affinities.

The scheduler stuff thats happening affects p-cores as well, on p-cores with default setup Windows will overload the preferred cores when it has a lot of threads, they can end up saturated whilst other p-cores have very low load, I fixed it by making all my p-cores run at same speed.