Question Raptor Lake - Official Thread

Page 136 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

Hulk

Diamond Member
Oct 9, 1999
4,525
2,518
136
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.
 
  • Like
Reactions: vstar

In2Photos

Platinum Member
Mar 21, 2007
2,026
2,054
136
Can't remember where I saw it, maybe it was der8auer, but I've seen some bracket thingy for LGA1700. It supposedly helps get a more even contact between cooler and IHS.

Is it worth getting one of those things? I intend to get the 13600K and overclock it. For cooling I have the Corsair H150i (older version) with Noctua fans.

Not sure what to expect from 13600K in terms of OC and temperatures. My (delided) 8086K at 5Ghz 1.35V can get pretty hot during stress testing, but it's rather comfortable in normal use.
It's called a contact frame. I believe it can provide up to 7C lower temps on the CPU. The Thermal Grizzly one is the original.
 

Kocicak

Golden Member
Jan 17, 2019
1,090
1,141
136
How well a cooler sits on on a cpu can be a matter of a half of turn of one of the cooler mounting screws. You can screw up mounting a cooler on a cpu with that frame as well.

In my opinion the issue is largely overblown and fed by the frame manufacturers and sellers.

You can check yourself if you have the problem by removing the cooler and inspecting, how is the paste spread on the the IHS. You may first try to adjust the mounting screws of your cooler differently by using common sense, if you have some problem. If you have a thicker paste layer on one side of the cpu, try loosening the screws on the opposite side, etc.
 
  • Like
Reactions: ftt

In2Photos

Platinum Member
Mar 21, 2007
2,026
2,054
136
How well a cooler sits on on a cpu can be a matter of a half of turn of one of the cooler mounting screws. You can screw up mounting a cooler on a cpu with that frame as well.

In my opinion the issue is largely overblown and fed by the frame manufacturers and sellers.

You can check yourself if you have the problem by removing the cooler and inspecting, how is the paste spread on the the IHS. You may first try to adjust the mounting screws of your cooler differently by using common sense, if you have some problem. If you have a thicker paste layer on one side of the cpu, try loosening the screws on the opposite side, etc.
Sounds like you admit there's an issue. If one has to remove the cooler to check for paste coverage the first time, how do you know you got it right the second time, or the third time without pulling the cooler again? Every review of these frames shows a benefit in temperatures. Even reviews by people with no stake in the frames themselves.
 

DAPUNISHER

Super Moderator CPU Forum Mod and Elite Member
Super Moderator
Aug 22, 2001
29,620
24,612
146
Intel is more efficient than AMD in gaming workloads.
Good stuff. I will post up if I see a test comparing power usage in Spiderman with ray tracing. That stresses way more cores than Valorant, which will gives us a look at how power usage scales.
 

DrMrLordX

Lifer
Apr 27, 2000
22,065
11,693
136
at production the 13th will edge out the 7950x at a higher power draw.

Eh, not really? But if you want to believe that go ahead, that wasn't what I was replying to anyway.

Going back to the idea of a hyptothetical 12p+16e Meteor Lake, which is what I was replying to . . . with 12 P cores, there's no need to try to push them to 5 GHz or beyond for MT workloads ala Raptor Lake to try for a benchmark win. From what has been implied by leakers, Intel 4 may bring with it clockspeed regressions anyway. You can't point to Raptor Lake as an example of how power-hungry a 12P Meteor Lake would be, when Intel would be certainly capable of dropping clocks to the 4 GHz range (or so) in an MT scenario.

That being said, the likelihood of such a chip coming to fruition seems very low at this point. We probably would have seen a mockup of such a chip or heard a leak by now.
 
Last edited:

Hulk

Diamond Member
Oct 9, 1999
4,525
2,518
136
In terms of die area, 12P + 16E = 8P + 28.4E using current RPL dimensions.

There wouldn't be much performance difference in lightly threaded loads and in heavily threaded ones 8P + 28.4E would be more performant and more efficient.

Such a part running at a relatively sedate 5GHz/4GHz would score almost 50,000 in CB R23 MT.
 

DrMrLordX

Lifer
Apr 27, 2000
22,065
11,693
136
In terms of die area, 12P + 16E = 8P + 28.4E using current RPL dimensions.

Intel doesn't have any Meteor Lake chiplets that they've demoed that would allow them to produce such a core configuration. They do have a 6P + 8e chiplet they've already shown to the public.
 

Kocicak

Golden Member
Jan 17, 2019
1,090
1,141
136
What it the core numbering on Raptor lake CPUs, by that I mean where is physically located core of specific number on the die?

I would like to try enable one or two more cores to run at higher frequency, and would like to start with the cores close to the two designated cores, hoping that they could be higher quality too.

Using distant cores would be good for heat dissipation though.
 

nicalandia

Diamond Member
Jan 10, 2019
3,331
5,282
136
What it the core numbering on Raptor lake CPUs, by that I mean where is physically located core of specific number on the die?

Core 0 to 7 are P Cores. 8 to 15 are e Cores, Cores 16 to 31 are just virtual cores residing on P cores, for example, number 16th core is actually a Virtual Core residing on P core 0
 

Hulk

Diamond Member
Oct 9, 1999
4,525
2,518
136
I did a stupid thing. For some reason my system was resetting anytime the CPU pulled about 250 or more Watts. My "old" power supply was a 460W Seasonic fanless, surely capable of this much power with iGPU right? So I figured there was something wrong with it and bought a Seasonic TX-750 Platinum. As you can see in my new updated signature.

As I was installing the new PS I noticed only one 8 pin power connector installed! My old 12700K never pulled enough power for the second cable so to have less cable clutter I never installed it! Stupid me! I'm too old for mistakes like this.

Thought about returning the TX-750 but it's just so nice, runs fanless in hybrid mode, 12 year warranty, and of course is more efficient than my old "gold" rated one that I've decided to keep it. I'll keep the FL460 around as a spare I guess.

Just posting in case someone else has this issue. But I don't think anyone here is dumb enough to miss that like I did!
 

Grimnir

Member
Jun 8, 2020
27
10
51
Sounds like you admit there's an issue. If one has to remove the cooler to check for paste coverage the first time, how do you know you got it right the second time, or the third time without pulling the cooler again? Every review of these frames shows a benefit in temperatures. Even reviews by people with no stake in the frames themselves.
Yeah, it's a bit tricky to check without removing the cooler... I'm also thinking that if the frame has been shown to improve temps, then I'd rather just install it straight away. It's not that fun to take the system apart and install the frame later on.

Also saw GN's test of Thermaltake's variant of the frame. Much cheaper, seems to work as well, and I can actually find it in stock. I think I'll go with that one.
 

IntelUser2000

Elite Member
Oct 14, 2003
8,686
3,786
136
In terms of die area, 12P + 16E = 8P + 28.4E using current RPL dimensions.

The E cores are relatively smaller with Meteorlake so 1P = 4E is roughly true. Besides, they wouldn't go with the oddball configuration and go all 32E cores if they wanted it for a very small increase in die size.

Being the first generation, Alderlake is pretty much the worst as the hybrid implementation goes.
 

Hulk

Diamond Member
Oct 9, 1999
4,525
2,518
136
The E cores are relatively smaller with Meteorlake so 1P = 4E is roughly true. Besides, they wouldn't go with the oddball configuration and go all 32E cores if they wanted it for a very small increase in die size.

Being the first generation, Alderlake is pretty much the worst as the hybrid implementation goes.

If the P's and E's are fabricated on the same process how have they increased the E to P ratio from the current 3.1:1 to 4:1? I assume either the P's are getting larger or the E's are getting smaller or a combination of both?
 

A///

Diamond Member
Feb 24, 2017
4,351
3,158
136
So tell us, if AMD weren't caught in the dark, what would they have released?
Provide a time machine to go back 3-4 years when zen 4 development began. Intel's e core development likely began in 2014-2016. Not after zen 2 like some young people believe. Intel has smart engineers. lots of them. but even they couldn't get out a product like alderlake in under 2 years.

in regard to your 10%, intel releases a product almost every 12 months while amd's development schedule is longer. buy every other generation for intel.
 

RTX2080

Senior member
Jul 2, 2018
325
512
136
Good question and actually that's the beauty of Intel's strategy. If you need more than 8 cores then you have an application this is very effectively multithreaded so it will do well with 16E cores or more. In fact as the bulk of applications become better multithreaded we could see the P's be reduced to 4 and the E's to 32 or 40.

The reason why Intel keep 8 P core is gaming and daily productivity usage, these usage would never being well threaded especially in latency sensitive situation. Reducing P core is a suicide decision. I think more than 16E won't happen, it would even possible to be reduced in the future, because using E core is just the reflection of process node disavantage.
 
  • Like
Reactions: coercitiv

TheELF

Diamond Member
Dec 22, 2012
4,027
753
126
Provide a time machine to go back 3-4 years when zen 4 development began. Intel's e core development likely began in 2014-2016. Not after zen 2 like some young people believe. Intel has smart engineers. lots of them. but even they couldn't get out a product like alderlake in under 2 years.

in regard to your 10%, intel releases a product almost every 12 months while amd's development schedule is longer. buy every other generation for intel.
So we are saying the same thing, amd wasn't caught by surprise, zen4 is all amd could have released without a new node.
 

A///

Diamond Member
Feb 24, 2017
4,351
3,158
136
The reason why Intel keep 8 P core is gaming and daily productivity usage, these usage would never being well threaded especially in latency sensitive situation. Reducing P core is a suicide decision. I think more than 16E won't happen, it would even possible to be reduced in the future, because using E core is just the reflection of process node disavantage.
Nothing about that leak made sense.
 

OneEng2

Member
Sep 19, 2022
78
95
61
Now that we've had over a year to test and discuss Intel's hybrid approach with both Alder and Raptor Lake I think we can come to some reasonable conclusions as to why they went down this path. Here's my version of this "story."

First, we have to assume they had a pretty good idea of their performance targets Zen 2, 3, 4.. at the time these parts were on the drawing board and they also had a pretty good idea of the transistor density they could achieve. It would have been obvious to them that they were not going to be able to compete on a core-for-core basis given the node deficit without creating huge parts and impacting the financial bottom like significantly.

With a die the size of the current 8+16 Raptor Lake die they could fit about 13.1 P cores, which would score about 35,000 in Cinebench R23, MT. Good but not enough to beat Zen 4. Furthermore using all E cores they could achieve a score of over 47,000, which would win the MT Cinebench war but of course lose in so many applications requiring ST performance.

This is where I think the Intel designers got it right. They figured that most applications currently that are not highly parallel don't use more than 8 or so cores at once so they went with the hybrid approach to mitigate their node deficiency. Whether or not you like it, given their fabrication limitations it is a good solution and has kept them competitive with AMD, who have equally good big core architecture and better process.

There is no doubt that in order to be competitive they not only had to employ the hybrid approach but also do the best they could from an efficiency point of view because AMD (TSMC) had them there as well. The final piece of the puzzle is because Intel has their own fabs and a significant war chest they could also confront AMD on price. When you put all of this together you see how we've come to the point where we have to very good brand choices at a variety of price points.

The data in the attachment were recorded with max power PL1 and PL2 set at 175W, which I think is reasonable, especially on air cooling. You can see that the E's do quite a bit for efficiency in terms of both area for the compute they generate as well as compute for the power they consume.
I mostly agree.

In the desktop, Intel did a great job of obtaining parity and even eclipsing Zen 4 performance in many cases by the use of "Big/Little" and by throwing power budget out the window. These trade-offs make sense in the desktop.

In the laptop market, my feeling is that Zen 4 and RDNA2 on TSMC N5 will be a much more trouble to Intel's Raptor Lake design (time will tell). Considering the market makeup of laptop vs desktop, this will likely be the more important battle between the 2 companies.

In the server market, AMD's design and scaling appear to dominate anything that Intel can put on the table throughout 2023. I suspect AMD would be perfectly willing to sacrifice market share in the PC Desktop to Intel in exchange for even more server market share.

For me, this comes down to Intel being able to meet its process roadmap. On paper, Intel is poised to dominate transistor density by 2024. My personal engineering opinion is that Intel's process roadmap is dangerously optimistic. Every process step involves big changes and advancements. Every process step is scheduled to coincide with a new CPU architecture. While this looks good on paper, will Intel be able to pull this off? Intel's transition to 10nm was a disaster. They are in the process of laying off a ton of engineers. To me, it seems like an impossibly difficult task that Intel executives are asking of Engineering. It took Intel 10 years to dig the hole they are in. I find it very hard to see a way for them to climb out of it in 3.

FWIW, I could easily be wrong. When Intel nose dived with Itanium and P4 I didn't see the Israel team bailing the company out with Conroe either ;). Of course, Intel still had the process advantage then.

from here:

https://www.granitefirm.com/blog/us/2021/12/28/tsmc-process-roadmap/
 

Hulk

Diamond Member
Oct 9, 1999
4,525
2,518
136
There is a continuum between single threaded applications and multithreaded applications. Today, most applications will utilize more than one thread. More so than 5 years ago. 5 years in the future even more applications will be better multithreaded. If all applications were like Cinebench, or nearly 100% multithreaded we would have no use for the P cores at all and the most efficient, performant, and cost effective CPU would consist of a swarm of E cores.

We are in transition to better multithreading usage in applications. Intel picked 8 as the magic number when they hatched this plan probably 5 or 6 years ago. That was actually pretty forward thinking of them.

Looking ahead they could be thinking that 6 P's may be enough and the remaining resources should be spent on increasing the IPC of the E's.

Like it or not, as far as the hybrid approach goes on the desktop TSW!