Discussion Intel current and future Lakes & Rapids thread

Page 578 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

Markfw

CPU Moderator, VC&G Moderator, Elite Member
Super Moderator
May 16, 2002
22,784
11,194
136
It isn't a rendering benchmark. Rendering benchmarks are essentially just a measure of how many cores in a system. CrossMark uses more features of the chips. See pages 21 to 31 to see what affects the scores more and what affects it less (frequency > core count > drive speed > memory channels > graphics capabilities > memory capacity).

a 12600k beats a 12700k and a 12900kf ?? There is something wrong with these benchmarks IMO
 
  • Like
Reactions: Tlh97 and Drazick

dullard

Elite Member
May 21, 2001
23,809
1,994
126
a 12600k beats a 12700k and a 12900kf ?? There is something wrong with these benchmarks IMO
The 12600K benchmarks were often with DDR5 4800 memory:

All six 12700K benchmarks were with DDR4 memory and 5 of those used quite slow DDR4 memory (these are OEM computers, not computers pushed to their limits):

And it appears that the graph sorts by maximum score, not by average score. If you sort by average score, the 12600K average of 1943 is well slower than the 2308 average of the 12900KF. It does seem that one single 12600K test threw off the maximum and caused the graph to sort it above the 12900KF.
 

Markfw

CPU Moderator, VC&G Moderator, Elite Member
Super Moderator
May 16, 2002
22,784
11,194
136
The 12600K benchmarks were often with DDR5 4800 memory:

All six 12700K benchmarks were with DDR4 memory and 5 of those used quite slow DDR4 memory (these are OEM computers, not computers pushed to their limits):

And it appears that the graph sorts by maximum score, not by average score. If you sort by average score, the 12600K average of 1943 is well slower than the 2308 average of the 12900KF. It does seem that one single 12600K test threw off the maximum and caused the graph to sort it above the 12900KF.
So, these are useless, since they are user provided, and no standards in place to compare real hardware.

I will take real reviews.
 

eek2121

Golden Member
Aug 2, 2005
1,779
1,953
136
Thanks ! But thats not good news for Raptor lake. With the same thread count and more real cores than a 5950x, its 3/4 of the performance of a 5950x. 2082(5950x) vs 1591(Raptor)

Except something is wrong with these charts or benchmark. How can a 12600k beat a 5950x by 25% !!!!
You are looking at an early ES chip. Clocks are likely less than half of what they should be.
 

Joe NYC

Senior member
Jun 26, 2021
687
618
96
Don't think it went that far. But AMD stopped cooperating with them ten years ago, along with Nvidia and VIA.
I don't remember exactly what it was, and I can't find anything good, but it may have been that IP addresses of BapCo came from block owned by Intel....

In any case, I came across some articles where AMD, NVidia and Via all quit BAPCo membership because of Intel's cheating (Intel controlled BapCo and rigged the benchmarks in favor of Intel).
 

Zucker2k

Golden Member
Feb 15, 2006
1,717
1,035
136
Intel was always faster all those years. The AMD-Zoners just couldn't handle the truth.


As you are VERY aware of by this point, trolling is not allowed.
This post offers nothing to this discussion/debate of all other users in this thread.

AT Mod Usandthem
 
Last edited by a moderator:
  • Like
Reactions: CHADBOGA

mikk

Diamond Member
May 15, 2012
3,449
1,197
136
You are looking at an early ES chip. Clocks are likely less than half of what they should be.

Any 0000 ES is meaningless performance related, but I guess he is trolling anyways. Any user with some knowledge should be aware of this, even though every time it's the same as usual some people behave as if it were the final performance.
 

DrMrLordX

Lifer
Apr 27, 2000
19,136
7,906
136
Yay, let's trigger people and derail once more! Merry trolling Christmas!
Got you a sweater:


Anyway we can't get much info out of a Raptor Lake release this early.
 

Ajay

Lifer
Jan 8, 2001
11,206
5,020
136
2024 ish. It is the supposed new core architecture.
Well, hopefully that will be a desktop chip, as opposed to Meteor Lake being mobile only. Wait, wasn’t there an Arrow Lake in there somewhere??
 

Exist50

Senior member
Aug 18, 2016
610
543
136
Well, hopefully that will be a desktop chip, as opposed to Meteor Lake being mobile only. Wait, wasn’t there an Arrow Lake in there somewhere??
Alder -> Raptor -> Meteor -> Arrow -> Lunar -> ??? -> Nova

Something like that, iirc. Probably some solid overlap between generations in different categories.
 
  • Like
Reactions: Tlh97 and uzzi38

DrMrLordX

Lifer
Apr 27, 2000
19,136
7,906
136
Didn't AMD do that after they knew that Bulldozer was a dud?
That was the straw that broke the camel's back. They had issues with them before that, and the BAPCO committee actually censured AMD at some point for all their complaining.

Alder -> Raptor -> Meteor -> Arrow -> Lunar -> ??? -> Nova

Something like that, iirc. Probably some solid overlap between generations in different categories.
There's gonna have to be a lot of overlap if Intel plans to get Nova out by 2024.
 

dullard

Elite Member
May 21, 2001
23,809
1,994
126
There's gonna have to be a lot of overlap if Intel plans to get Nova out by 2024.
Below is the latest roadmap leak that I've seen (posted earlier in this thread).
1638800013791.png

A few rumors other rumors help give that image credibility, but as always rumors can be wrong and plans can change. I personally think up to Arrow Lake seems likely to be achievable with a very rapid succession of products. Past that I have doubts. And yes, this is a planned short overlap of products.
  • Alder Lake. Just over 7 months after Rocket Lake. This the start of many short overlaps.
  • Raptor Lake. May be late 3Q 2022 or early 4Q 2022. Maybe the September/October time frame that Intel used to do somewhat regularly.
  • Meteor Lake. Intel's EUV is supposed to be up and running Q3 2022, so Meteor Lake 1H 2023 seems correct.
  • Arrow Lake. Depends on TSMC's N3 capability. The last that I saw, their N3 node is supposed to be having revenue production in 1Q 2023, so Arrow Lake 2H 2023 should be doable.
  • Lunar Lake. Intel claims that Intel 3 node will be used in manufactured products in 2H 2023. So, Lunar Lake in 2024 is possible
  • Nova Lake. Seems the hardest to achieve. It involves multiple innovations that haven't been perfected yet. They have time on their hands, but Intel is possibly biting off more than they can chew again.
  • Beyond. High-NA EUV is supposed to be ready Q3 2025, and may very well slip, so I think the chip after Nova Lake is 2026 at the earliest.
 

repoman27

Senior member
Dec 17, 2018
219
275
106
Well, if we use our critical thinking skills for just a second, we'd recognize that it isn't a "roadmap leak" at all. It's merely a table 3DCenter cobbled together based on a handful of sketchy rumors, similar to those found on VideoCardz and WCCFTech.

Rocket Lake was released 6 months ago. 5% of Alder Lake SKUs have already been released and the other 95%, which are presumably forthcoming in H1'22, are all but known quantities at this point.

On the other end of the spectrum, the only thing we know about Lunar Lake and Nova Lake are potential codenames and further speculation by MLID and AdoredTV. By all appearances, the mooreslawisnotdead account on reddit and Curmudgeon666 account here on the Anandtech forums were sock-puppet accounts used to promote these rumors and leak additional codenames so that MLID and AdoredTV could further discuss them and drive traffic to their YouTube channels. Arrow Lake is close to being in the same boat, except for some additional sketchy rumors regarding possible CPU core and GPU configurations.

Intel has disclosed a fair amount about Meteor Lake at this point, but a lot of posters here seem intent on disregarding that information and promoting leaks from sources with mixed track records instead. I realize "real desktop" CPUs from Intel require at least 241 W these days, but I still think 125 W is a tad on the high side for a notebook PC. All of Intel's public disclosure thus far has indicated that 125 W Meteor Lake-S will be a thing. Meanwhile, MTL-M certainly looks like it is on track for launch as early as Q1'23. We know it will use Foveros with separate CPU, SOC, and GPU tiles. The base tile will be manufactured using Intel 7 FOVEROS, and the CPU tile using Intel 4. The GPU tile will almost certainly be made by TSMC on N3, for which the timing and all of the rumors line up perfectly, despite the protests from various posters here for... reasons? That leaves just the SoC tile, which could be made on either Intel 7, TSMC N5P, or TSMC N4. Seeing as Meteor Lake is a Foveros design, I highly doubt it will use the LGA1700 package. It also looks like like mobile will come first, with MTL-M and MTL-P launching in H1'23 and MTL-S possibly following in Q4'23.

Raptor Lake is rumored to be a minor "moar cores" update to Alder Lake, adding two more 4-core Gracemont clusters for an 8+16+1 configuration, which should help keep Intel's head above water on multi-threaded benchmarks. This configuration only makes sense on the desktop and probably will continue to use the LGA1700 socket. RPL-S will likely launch around Q4'22 to fill the hole in the desktop schedule. This could be the last monolithic processor family from Intel.

If any of the configurations rumored for Arrow Lake are correct, it sounds like an extension of Meteor Lake targeting the more niche qausi-HEDT and premium ultra-mobile platforms. Basically MTL-S with four 2+8 CPU tiles for an 8+32+1 config, and MTL-P with two GPU tiles for a 320 EU 6+8+3 config. Arrow Lake is also rumored to use updated cores (Lion Cove + Skymont) in the CPU tiles, which might be the case? I could see Arrow Lake launching anywhere in the Q4'23-H1'24 window.

Lunar Lake will probably be the 2024 refresh of Meteor Lake but with Lion Cove and Skymont CPU cores and next gen GPU.
 

dullard

Elite Member
May 21, 2001
23,809
1,994
126
Well, if we use our critical thinking skills for just a second, we'd recognize that it isn't a "roadmap leak" at all. It's merely a table 3DCenter cobbled together based on a handful of sketchy rumors, similar to those found on VideoCardz and WCCFTech.
I am a bit confused by you bashing rumors in your first paragraph and then mentioning rumors throughout the rest of your post.

We can put all of what Intel has publicly mentioned into a single post. Most of it is contained here: https://www.intel.com/content/www/us/en/newsroom/resources/press-kit-accelerated-event-2021.html#gs.inqbq5 These bullet points are direct from Intel:
  • Intel 7 products such as Alder Lake for client in 2021
  • Intel 4 will be ready for production in the second half of 2022 for products shipping in 2023, including Meteor Lake
  • Intel 3 will be ready to begin manufacturing products in the second half of 2023.
  • Intel 20A It is expected to ramp in 2024.
  • 2025 and Beyond: Intel 18A is already in development for early 2025
Sure, their plans may go astray and products may slip, but these are their publicly stated goals.

From there, we do need external rumors. Yes, they are rumors, but many do have some strong basis in reality, such as leaked benchmarks, updated drivers mentioning products, etc. There is some faith needed, such as assuming that Intel will produce a CPU on Intel 3. Thus, I assume that when Intel ramp's up Intel 3 in H2'2023 that there is likely to be a CPU using it in 2024. I don't think it is that much of a stretch of what Intel has stated publicly to assume that. If Intel's timeline slips, the assumption will be wrong, but we can't know until closer to the date.

If we go fully into rumor mode, then we can speculate about more precise details. For example, this tweet hints at Raptor lake in Q3'22. But that is just pure rumor and thus, should not be taken seriously yet. It does fit somewhat within Intel's annual cadence goal, but it is just rumor.
 
Last edited:

repoman27

Senior member
Dec 17, 2018
219
275
106
I am a bit confused by you bashing rumors in your first paragraph and then mentioning rumors throughout the rest of your post.

We can put all of what Intel has publicly mentioned into a single post. Most of it is contained here: https://www.intel.com/content/www/us/en/newsroom/resources/press-kit-accelerated-event-2021.html#gs.inqbq5 These bullet points are direct from Intel:
  • Intel 7 products such as Alder Lake for client in 2021
  • Intel 4 will be ready for production in the second half of 2022 for products shipping in 2023, including Meteor Lake
  • Intel 3 will be ready to begin manufacturing products in the second half of 2023.
  • Intel 20A It is expected to ramp in 2024.
  • 2025 and Beyond: Intel 18A is already in development for early 2025
Sure, their plans may go astray and products may slip, but these are their publicly stated goals.

From there, we do need external rumors. Yes, they are rumors, but many do have some strong basis in reality, such as leaked benchmarks, updated drivers mentioning products, etc. There is some faith needed, such as assuming that Intel will produce a CPU on Intel 3. Thus, I assume that when Intel ramp's up Intel 3 in H2'2023 that there is likely to be a CPU using it in 2024. I don't think it is that much of a stretch of what Intel has stated publicly to assume that. If Intel's timeline slips, the assumption will be wrong, but we can't know until closer to the date.

If we go fully into rumor mode, then we can speculate about more precise details. For example, this tweet hints at Raptor lake in Q3'22. But that is just pure rumor and thus, should not be taken seriously yet. It does fit somewhat within Intel's annual cadence goal, but it is just rumor.
Sorry, I didn't mean to be a jerk. I mean, most of why I follow this thread is for the leaks and rumors after all.

I guess I'm just skeptical by nature, so I always try to follow each new piece of information back to its original source the best I can before evaluating its veracity. For instance, I generally trust information passed on by @momomo_us, but that account is rarely the primary source. Fortunately they almost always cite their sources, so I can see that the SPR-X and RPL-S in Q3'22 rumor goes back to a September 16, 2021 post by 热心市民描边怪 (Enthusiastic Citizen Stroke Monster) on Chiphell: https://www.chiphell.com/thread-2355542-1-1.html

So that is both plausible and appears to originate from someone who may have seen an actual Intel roadmap. I just figured October for RPL-S based on annual cadence and the ADL-S launch date, but September certainly wouldn't be out of the question.

Intel hasn't really said anything about Raptor Lake yet, I'm guessing that's because it's not very interesting and there's not much to say about it. They've been teasing Meteor Lake pretty hard though. They told us when they taped in the compute tile, what process it was on, showed us a test wafer, told us it was using 2nd generation Foveros, gave us a TDP range, showed us that there are three top tiles with labels indicating how the platform is disaggregated and possible EU counts for the GPU, let us see prototype MTL-M packages being assembled... That's a lot more color than we've gotten from the rumor mill thus far.
 

dullard

Elite Member
May 21, 2001
23,809
1,994
126
Intel hasn't really said anything about Raptor Lake yet, I'm guessing that's because it's not very interesting and there's not much to say about it. They've been teasing Meteor Lake pretty hard though. They told us when they taped in the compute tile, what process it was on, showed us a test wafer, told us it was using 2nd generation Foveros, gave us a TDP range, showed us that there are three top tiles with labels indicating how the platform is disaggregated and possible EU counts for the GPU, let us see prototype MTL-M packages being assembled... That's a lot more color than we've gotten from the rumor mill thus far.
I can speculate on two possible reasons for that.

1) Raptor Lake might be a stop-gap chip that is Intel's top chip for just a short time--similar to Rocket Lake. If Meteor Lake does make it to desktop, then it will replace Raptor Lake only ~6 to 9 months after Raptor Lake is launched. Rumors go both ways as to whether Meteor Lake is a desktop chip or not. But if it is a desktop chip, then Raptor Lake will only be a blip in history.

2) Intel wants to show off its capabilities. There is nothing really special about Raptor Lake. A bit higher frequency, a bit more energy efficient, a bit better memory controller, a bit more cache, and another set of E cores. Nothing really Earth-shattering. Intel 7 is still a slightly behind the greatest from TSMC. But, Intel 4 is likely to be quite competitive to TSMC. Meteor Lake shows off Intel 4, Foveros, Intel's first EUV use, etc. If Intel is trying to drum up new business for their foundry, Meteor Lake leaks are by far the way to do it.
 

IntelUser2000

Elite Member
Oct 14, 2003
8,054
2,856
136
Meanwhile, MTL-M certainly looks like it is on track for launch as early as Q1'23.
I don't see how that's possible since they've never made early mobile launches and Raptor Lake mobile seems to be a thing. We don't even have Alderlake mobile yet. If Raptor Lake mobile is going to have a short lifespan, most manufacturers are going to say no, considering we are still seeing Tigerlake based platforms being launched.

If they want overlapping launches, it would mean either Raptor or Meteor is a sidegrade in many aspects. Maybe it's possible with Meteor configurations having less cores but beefier GPU and Raptor the other way around.

The rest of Raptorlake seem "ok" but DLVR seems like a big deal for mobile.

Unless something like that happens I see Raptor mobile in Q1 2023 and Meteor in Q1 2024.

One disconnect I see is that Meteorlake doesn't sound very exciting from the rumor mill but Intel is putting it up front like there's nothing else. I have a suspicion that its a marketing tactic and Raptorlake will be a decent chip and Meteorlake, despite all it's breakthroughs are really another incremental improvement and nothing special except for maybe the iGPU.

Kinda like Icelake. Sure their first 10nm implementation but other than the GPU what was special? Extra battery life? Nope. Better performance? Not really.

Basically MTL-S with four 2+8 CPU tiles for an 8+32+1 config, and MTL-P with two GPU tiles for a 320 EU 6+8+3 config. Arrow Lake is also rumored to use updated cores (Lion Cove + Skymont) in the CPU tiles, which might be the case? I could see Arrow Lake launching anywhere in the Q4'23-H1'24 window.
Raichu has also stated there are Meteorlake variants with greater than 192EUs. That means a 320EU Arrow Lake is really not an upgrade.

Also, unless that 320EU is a Kaby-G like part I see a Arrowlake mobile on the horizon.
 
Last edited:

ASK THE COMMUNITY