News Intel Bartlett Lake-S: up to 12P-Core or up to 8P-Core +16E-core

[Kosusko]

LGA 1700 is still alive

Intel Bartlett Lake-S Desktop CPUs Launching In 2025: Up To 8+16 Hybrid & Up To 12 P-Core Only Flavors​

Hybrid version up to 65W TDP and only P-core version up to 125W TDP

• Core i9 - 12P-Core
• Core i7 - 10P-Core alebo 8P-Core + 16E-core
• Core i5 - 8P-Core alebo 6P-Core + 8E-core ako aj 6P-Core+ 4E-core
• Core i3 - 4P Core
• Core processor - 2P Core

source: https://wccftech.com/intel-bartlett...nch-2025-up-to-8-16-hybrid-12-p-core-flavors/

 

[zir_blazer]

Based on how complex the logistics to make a Processor with chiplets coming from multiple different sources should be, it seems like some kind of Plan B to use preinstalled manufacturing for simpler monolithic dies in case the next generation can't come close to satisfy market demand. However, I don't see Intel lowering prices of a previous generation platform like AMD seems to be doing with AM4 to let it fit the budget niche, I don't recall Intel having ever done so.
Oh, and because the upcoming class action lawsuit for the Raptor Lake degradation issues will hurt. Free upgrades to socket compatible 15th gen to all affected customers? If I was Intel, I would be preparing my butt for it, just in case the boot kicks too hard. Plus it will be easier to identify fixed Raptor Lake parts if they have a new name, heh.

 

[Tech Junky]

Got it right the first time with ADL but got greedy and added more E cores and upped the naming.

 

[DavidC1]

So they have problems on 13 and 14 generation?

Got it right the first time with ADL but got greedy and added more E cores and upped the naming.

Yes but that's not why they have the problem. The issue is pushing it too far to previously exotic cooling levels of MHz.

Repeat of Pentium III 1.13GHz. So the peak CPU should have been the 13700K. Doesn't sound too good though.

Everything is mini-Netburst. Not insane clocks, but still extremely high. Uop cache instead of Trace Cache to make up for Willamette-like pipeline stages.

 

[Tech Junky]

@DavidC1

the problem is they're using the same tech for the past 10 years aka 10m,+++++++++ and cramming more stuff into the same footprint aka LGA1700 and then charging for more E cores w/ no real increase in performance and more crashes. ADL though doesn't / didn't have any issues like these subsequent releases.

If you're ok with dropping $500+ on a CPU/board and rolling the dice on old tech feel free. I wouldn't bother with anything Intel though until ARL comes out which will be a complete redesign from the ground up instead of repackaging the SOS.

Personally I ditched Intel and went AMD as a lateral when I had issues with the setup and adding U.x drive to the system and having 2 fail spectacularly within less than a week of adding them. Switching over though wasn't a huge thing as I tend to rebuild / upgrade often anyway just to stay up to date with tech and sell off / recoup the costs before things get too old. Now, looking back at some of the gremlins with that setup it seems to all come back to the MOBO I was using vs the CPU itself. I had a card that wouldn't work 100% and figured maybe it was the platform but, I pulled the card out of the drawer and put it in the AMD setup and worked just fine like new. Some other little things as well improved by switching things up. In between the two I did switch MOBOs to accommodate 5*M2 before finding the U.x drives to be more budget friendly for higher capacity/unit. All in all though for the base AMD setup it was only $800 for a 12/24 CPU / 32GB / 850W PSU / misc parts (case/fans/etc.) It's been rock solid for the past 11 months.

 

[Kosusko]

The essential thing is that LGA 1700 (DDR4 / DDR5 platform) is still alive in 2025 with Intel Bartlett Lake-S and even with really big cores up to 12C, but also with x86 paradigm shift big.LITTLE up to 24 cores.

 

[Wolverine2349]

This is so exciting if true. I have waited and badly wanted for a long while more than 8 P core son a homogenous arch on the same die/node/ring bus/CCD-CCX.

Bartlett Lake finally it.

Though still a year wait for Raptor Cove arch. And given stability issues with 13th and 14th Gen, it may all be for naught!!

But if they are fixed and it does not degrade easily, this is the dream 12 P cores on a single die chip I have wanted.

 

[MangoX]

If these are still on 10nm, it's a hard pass. Why the excitement for 12P cores when AMD has been offering 16 cores for three (soon 4) generations now?

 

[Hulk]

It's July, not April.

 

[IEC]

*IF* it resolves instability issues and *IF* it is on a leading edge node and *IF* it is competitive with Zen 5 chips then I might actually have an upgrade path for my Z690 board in 2025-2026.

But that's a lot of ifs.

 

[Wolverine2349]

If these are still on 10nm, it's a hard pass. Why the excitement for 12P cores when AMD has been offering 16 cores for three (soon 4) generations now?

More than 8 big cores on a single node instead of dual nodes (CCX-CCDs)

I imagine all 12 of those would be on a single ring bus would they not?

10nm, yeah maybe a hard pass due to stability issues, though 10nm Alder Lake was fine unlike Raptor Lake. But if this is based on Raptor Lake on 10nm, gives me reason to pause and fearful of easy to degrade/stability issues like 13th and 14th Gen.

If it is fixed and they can prove it was fixed and it does not degrade easily, I am all on board and will be a buyer.

Easier said than done without months of concrete proof and stability testing in the wild if only the words come form Intel just themselves.

 

[Wolverine2349]

*IF* it resolves instability issues and *IF* it is on a leading edge node and *IF* it is competitive with Zen 5 chips then I might actually have an upgrade path for my Z690 board in 2025-2026.

But that's a lot of ifs.

Yes a lot of ifs. I like 12 P cores on a single ring bus/CCX-CCD/Zen 5 will not have it and Zen 6 likely will not either.

Though stability issues if if if if they are fixed like you stated. Without extensive testing and research posted to prove stability is for real a hard pass despite it being the only option of a CPU I have desired for more than 2 years now (10-12 Big cores on a single node/ring bus/CCX-CCD/tile).

 

[vanplayer]

It would be very, very interesting to see comparison of 12 P cores to any 8P+NE variant when BartlettLake is launched.
Looks like Intel realize something is wrong with ADL/RPL hybrid method and make hard decision to switch back to the way of CometLake which is 10C/20T.

And it seems AVX512 is still have to disabled.

 

[hemedans]

It would be very, very interesting to see comparison of 12 P cores to any 8P+NE variant when BartlettLake is launched.
Looks like Intel realize something is wrong with ADL/RPL hybrid method and make hard decision to switch back to the way of CometLake which is 10C/20T.

And it seems AVX512 is still have to disabled.

Or they don't have capacity for Arrow lake and they need something to cover volume, probably these are going to be really cheap to compete with Zen 5.

 

[coercitiv]

It's July, not April.

We can switch to February if you like

Intel's Bartlett Lake-S is incoming, but not to desktop PCs — chips will debut for Intel's NEX group

Intel preps socketed Bartlett Lake-S for network and edge systems.

www.tomshardware.com

 

[SteinFG]

So Intel will offer more 2 more P cores, per tier, than AMD? We're coming full circle. I realy like this P-only chip

 

[coercitiv]

So Intel will offer more 2 more P cores, per tier, than AMD? We're coming full circle. I realy like this P-only chip

The only catch is they're going to be Raptor Cove most likely. We'll have to see prices and availability though, as the original rumors said these products were meant for the professional market.

 

[ondma]

I got really excited when I saw this, but then saw it was Raptor Cove and lost interest immediately. Hot, power hungry, stability issues, and behind in IPC compared to Zen 5 and hopefully ARL. With all the trouble 13 and 14th gen is having, I suspect we will never see this product for the consumer market anyway. Besides, if you want more than 8 "big" cores, AMD already can provide it with Zen 5 on a better process and with better IPC and much better power efficiency. Yea, there is a CCD penalty, but better core design and process will probably still put it ahead of a hypothetical Bartlett Lake.

 

[poke01]

If it’s still on 10nm, what’s the point? The power consumption for a 12P core will be too high.

 

[NTMBK]

More than 8 big cores on a single node instead of dual nodes (CCX-CCDs)

I imagine all 12 of those would be on a single ring bus would they not?

10nm, yeah maybe a hard pass due to stability issues, though 10nm Alder Lake was fine unlike Raptor Lake. But if this is based on Raptor Lake on 10nm, gives me reason to pause and fearful of easy to degrade/stability issues like 13th and 14th Gen.

If it is fixed and they can prove it was fixed and it does not degrade easily, I am all on board and will be a buyer.

Easier said than done without months of concrete proof and stability testing in the wild if only the words come form Intel just themselves.

Broadwell-E had 10 cores on a single ring back in 2016:

But beyond 10 cores they switched to dual ring, presumably because a ringbus scales poorly beyond that point. And Skylake-E moved to a mesh.

 

[coercitiv]

12P and 8P+16E would use the same number of ring stops, so there's nothing new under the sun from this point of view. IIRC Intel doubled the ring bus with Tiger Lake for the twice the coherent fabric bandwidth, I would imagine they kept it and improved on it for further designs, so the "ring bus" of today is no longer the ring bus of the last decade.

 

[Det0x]

Sounds interesting until i read:

Hybrid targeting early Jan’25, P-core only targeting Q3’25

 

[poke01]

Sounds interesting until i read:

Yeah, what a mess

 

[coercitiv]

Yeah, what a mess

The original rumor from February said these chips were meant for Intel's Networking and Edge group (NEX). They were not primarily intended for the consumer platform.

People need to keep in mind everything we are discussing here is based on superficial leaks, with important aspects being gleaned from one sentence tweets. We don't know anything for certain.

 

[TheELF]

Oh, and because the upcoming class action lawsuit for the Raptor Lake degradation issues will hurt.

They would have to prove that this happens on a CPU that never ran outside of intel specs, so it will never happen.
You can't claim damages for something the company never promised.
A class action could be had against mobo makers that didn't clearly state that they where using out of spec values.