Intel Advancing 13th Gen Core "Raptor Lake-S" Launch to Q3-2022?

[btarunr]

Intel is allegedly advancing the launch of its 13th Gen Core "Raptor Lake-S" desktop processors to some time in Q3-2022, according to a report by Moore's Law is Dead. It was earlier believed to be a Q4 launch, much like "Alder Lake" was, in 2021. The report predicts the debut of "Raptor Lake" in the desktop segment in Q3-2022 (between July and September), with certain mobile SKUs expected toward the end of the year, in Q4. The Core "Raptor Lake-S" processor is built in the existing Socket LGA1700 package, and is being designed for compatibility with existing Intel 600-series chipset motherboards with a firmware update.

The "Raptor Lake-S" silicon is built on the existing Intel 7 (10 nm Enhanced SuperFin) node, and physically features eight "Raptor Cove" P-cores, along with sixteen "Gracemont" E-cores that are spread across four clusters. The chip has additional cache memory, too. Moore's Law is Dead predicts that the "Raptor Cove" P-core could introduce an IPC uplift in the region of 8 to 15 percent over the "Golden Cove" core, while the chip's overall multi-threaded performance could be anywhere between 30 to 40 percent over "Alder Lake-S," on account of not just increased IPC of the P-cores, but also eight additional E-cores.



View at TechPowerUp Main Site | Source

 

[londiste]

MLID accuracy has not exactly been on a high level.
It is a good enough guesstimate though. Both Zen4 and Raptor Lake have been said to be 2022 H2 with the implication that this means 2022 Q4. With AMD stating Zen4 will be out in Q3, Intel is likely to try and counter that somehow.

 

[AusWolf]

Why so many E-cores?

 

[Cutechri]

Why so many E-cores?

1. The universally known die constraints, many P-cores on the same die is difficult for them to achieve and most definitely bad for cooling.
2. Why not? They're proven to be quite strong and efficient in multi-threaded workloads. Improve the P-core architecture keeping the numbers the same, while piling on more E-cores. I see nothing wrong with that.

 

[TheLostSwede]

Why so many E-cores?

Because they're cheap.

 

[Tomorrow]

Am i the only one who's a little worried about bugs and reliability in light of the new arms race? If both try and release their platforms ASAP then i can only guess what corners are being cut.
We already have games industry that releases half baked products. We dont need the same happening in the hardware side. I rather both take an extra quarter to make sure it's properly tested. Rather than releasing with potential major bugs and poor availability.

 

[TheLostSwede]

Am i the only one who's a little worried about bugs and reliability in light of the new arms race? If both try and release their platforms ASAP then i can only guess what corners are being cut.
We already have games industry that releases half baked products. We dont need the same happening in the hardware side. I rather both take an extra quarter to make sure it's properly tested. Rather than releasing with potential major bugs and poor availability.

Sorry, but haven't you noticed that early adopters have been beta testers for what, the last decade at least?
Intel used to be the reliable hardware maker, but apparently not so much any more.
Based on my own experience with the X370 and X570 platforms from AMD, I'm not jumping on something new until it's been in the market for at least six months.
Routers, phones and just about all other tech out there is the same these days. It's a sad state of affairs driven by trying to appease the shareholders.

 

[napata]

Why so many E-cores?

More performance/mm². If everything scaled perfectly you'd probably want an E-core only CPU.

 

[londiste]

More performance/mm². If everything scaled perfectly you'd probably want an E-core only CPU.

Also performance/W at lower power levels.
Didn't Intel already say there will be Xeons with only E-cores? Similarly, AMD has said the 128-core EPYCs will be Zen4e/Zen4c cores only (whatever the optimizations on those end up being).

 

[ThrashZone]

Hi,
Wicked

 

[AusWolf]

1. The universally known die constraints, many P-cores on the same die is difficult for them to achieve and most definitely bad for cooling.
2. Why not? They're proven to be quite strong and efficient in multi-threaded workloads. Improve the P-core architecture keeping the numbers the same, while piling on more E-cores. I see nothing wrong with that.

More performance/mm². If everything scaled perfectly you'd probably want an E-core only CPU.

I'm not so sure. I'll have to dig deeper into how powerful E-cores are. I've been under the assumption that E-cores are mostly responsible for background stuff, while P-cores do the heavy lifting. I might be wrong.

 

[Cutechri]

I'm not so sure. I'll have to dig deeper into how powerful E-cores are. I've been under the assumption that E-cores are mostly responsible for background stuff, while P-cores do the heavy lifting. I might be wrong.

There was a thread the other day that showcased the significant efficiency and performance multiple E-cores have in MT tasks, though I can't find it. I think they're a welcome addition to the desktop space and the way forward, even on MCM. Those 8 E-cores allowed the 12900K to surpass the 5950X in MT.

 

[ThrashZone]

Those 8 E-cores allowed the 12900K to surpass the 5950X in MT

Exactamundo

 

[btarunr]

Why so many E-cores?

 

[Dr_b_]

View attachment 237428

lol.

I'm not so sure. I'll have to dig deeper into how powerful E-cores are. I've been under the assumption that E-cores are mostly responsible for background stuff, while P-cores do the heavy lifting. I might be wrong.

dont think you are wrong at all, the use case for them is heavy MT workloads

 

[AusWolf]

View attachment 237428

I saw this meme during the AMD FX era. It's still funny, though.

 

[DeathtoGnomes]

So Intel is trying to pull an AMD so that any PR reads as having 24 cores, which is off course more cores than any amds desktop offerings ( not including threadripper), again which Intel will likely use as a marketing point.

When is AMDs next cpu(6xxx? 7?) launch? The advancement could push AMD into a premature launch, or not.

 

[Deleted member 24505]

Nice to see people actually understanding the E cores.

 

[mama]

Amazing what an incentive a little competition is for Intel.

 

[ModEl4]

Although as already been said, MLID accuracy has not exactly been on a high level and i would add that this is not the worst thing of his videos (I haven't heard the word "sources" so many times outside some wizard build guide videos for WRPG style of games, lol)
But this "report" i think is correct and i would speculate regarding Meteor Lake iGPU (arc tile based) that we will have at least doubling of the performance vs Alder Lake.
If Intel wants to sell at $159-199 bracket the 128EU card:

https://www.notebookcheck.net/fileadmin/Notebooks/News/_nc3/Intel_DG2_arc_alchemist_rx_6700xt_rtx_3070.jpg

then this comes with certain performance requirements which translates well for this kind of advancement in performance for the iGPU sector (also on top of that there is the possibility that Meteor Lake iGPU will be based on Battlemage core given the timeframe) which leads to another speculation that maybe AMD's 680M successor of that time period Q3 2023-Q1 2024 will go from 12CUs to 16CUs +RDNA3 advancements +clock advancement to retain the performance iGPU lead, which translates well for us, after all competition is good

 

[Cutechri]

Amazing what an incentive a little competition is for Intel.

And I hope they keep going like this. Alder Lake genuinely impressed me, especially the performance of the E-cores.

 

[Minus Infinity]

So Intel is trying to pull an AMD so that any PR reads as having 24 cores, which is off course more cores than any amds desktop offerings ( not including threadripper), again which Intel will likely use as a marketing point.

When is AMDs next cpu(6xxx? 7?) launch? The advancement could push AMD into a premature launch, or not.

AMD is reportedly releasing Zen4 Q3, I think Intel panicked when they saw Su had brought forward the Zen 4 release and now magically they have brought RL forward a quarter to Q3. Ain't competition wonderful.

AMD's answer to E cores is 4c cores, they will lose some features over regular 4 cores, less cache, lower IPC uplift, but they are reportedly releasing 4c only CPU's with possibly 24/32 cores. 4c cores will be faster than Zen3+ cores and so will demolish Gracemont cores. If you need a MT monster and 24/32 core Zen 4c will be the answer. Zen 4 is the largest architectural change in the Zen series, muvh bigger than Zen 2 -> Zen 3 and will get 25-30% IPC uplift, 2x the L2 cache, possibly v-cache. Come late this year it's going to be a great contest between Raptor Lake and Zen 4. I just hope RL improves efficiency a lot, especially for P versions.

 

[Deleted member 24505]

From what i can find, 4c is for cloud called bergamo, there may be a 24/20 zen4 i9 though. Can't see bergamo on desktop.
https://www.club386.com/amd-teases-128-core-bergamo-zen-4c-epyc-chips/

As for Zen 4
We will see 24 core Zen4/5 CPUs on AM5 socket. Probably the listed SKUs will be:
Ryzen 9 = 24/20 core
Ryzen 7 = 16/12 core
Ryzen 5 = 8/6 core
Ryzen 3 = 4/2 core

https://www.techadvisor.com/news/pc-components/amd-zen-4-3801795/

 

[Vader]

I wonder how much advancement will these bring to the i5s and i3s. A 8-15 ipc increase for the p-cores won't be a big jump like comet lake -> alder lake. Unless they bring p-cores to them, that'll be interesting.

 

[watzupken]

Why so many E-cores?

The E-cores are there to prop up the core/ thread numbers to match competition. It is rumored that AMD's Zen 4 is going to have a higher core count, so Intel is trying to dull the more core/ thread threat with more cores so that their multithreaded numbers won't fall behind. They are not there from an efficiency standpoint, at least not the primary reason.