News Intel's Arrow Lake-S won't be an AI powerhouse — 13 TOPS NPU is only slightly better than Meteor Lake, much less than Lunar Lake

[Admin]

Arrow Lake-S is reportedly coming with an NPU, but it isn't the same NPU equipped in Lunar Lake. The desktop architecture will arrive with just 13 TOPS of NPU performance, making Arrow Lake-S' NPU only 3 TOPS quicker than Meteor Lake's NPU.

Intel's Arrow Lake-S won't be an AI powerhouse — 13 TOPS NPU is only slightly better than Meteor Lake, much less than Lunar Lake : Read more

 

[Deleted member 2731765]

It is a mystery why Intel has decided to go with an NPU at all in Arrow Lake-S and use one that is not at least at the same performance level as Lunar Lake's NPU (after all, Lunar Lake is Arrow Lake's ultra-low power mobile counterpart — architecturally they are both similar).

It's no mystery actually, because Arrow Lake-S lineup is most likely using an older NPU architecture, whereas INTEL used the NPU 4 arch version for the Lunar Lake mobile lineup (see image below). That's why there is a lot of difference in TOPS value between these two.

Nonetheless, Arrow Lake S processors will have a combined 37 AI TOPS which puts it close to the Ryzen 8000G "Hawk Point" APUs which have a combined TOPS of 39.

 

[edzieba]

Oh no! Anyway...

 

[Deleted member 2731765]

At this point though, none of these predictions and leaks can be fully trusted, and this poster has been tweeting almost every day with some new info, which in itself cannot be verified because there is no proof whatsoever to backup his claim.

So, at this point I think these are just clickbait posts, and nothing else.

Sorry, that's what one has to assume when some random user floods the internet each day with a new leaked info. For each tweet, we now get a new article which makes it even more awkward and funny.

I wouldn't take this TOPS data value seriously though. Not unless confirmed via some official Intel slide/document.

 

[usertests]

This is a good thing, if businesses throw the dGPU-less mini desktops away faster to get to 40 TOPs for Copilot+... if there is demand for that. Then I'll buy them for cheep.

 

[thestryker]

The real question is why there would be any NPU in ARL-S given that it is a part of the SoC tile. They're certainly not just reusing the MTL tile since it had low power E-cores in it. It seems like a complete waste of silicon since they had to design the SoC tile for ARL-S in the first place. This would almost make sense if we were talking the ARL-P as Intel may not be putting the latest NPU in there and the IGP will be much more powerful.

 

[Deleted member 2731765]

Well, on some other news, CPU-Z tool has just confirmed the nomenclature of the entire Arrow Lake S desktop processor lineup, and also Lunar Lake series.

At least we can fully trust on this info since this is coming directly from the official CPU-Z change log.

https://www.cpuid.com/softwares/cpu-z.html

So for Arrow Lake-S, we have the following SKUs confirmed, but without any core counts though. If one would hazard a guess, then most probably, the Core Ultra 9 285K processor might be the flagship 24 cores part, sporting the "8 P + 16 E" core config:

Core Ultra 9 285K
Core Ultra 9 275,
Core Ultra 7 265K
Core Ultra 7 255,
Core Ultra 5 245K
Core Ultra 5 240.


These are the Lunar Lake SKUs:

Core Ultra 9 288V ;
Core Ultra 7 268V,
Core Ultra 7 266V,
Core Ultra 7 258V,
Core Ultra 7 256V ;
Core Ultra 5 236V,
Core Ultra 5 228V,
Core Ultra 5 226V (this specific part has a typo error in it's name)

Version 2.10 for windows® x86/x64

• AMD Ryzen 9 9950X (16C/32T), 9900X (12C/24T), Ryzen 7 9700X (8C/16T) and Ryzen 5 9600X (6C/12T) Granite Ridge (Zen 5).
• AMD Ryzen AI 9 HX 375 (4x Zen 5 + 8x Zen 5c), Ryzen AI 9 365 (4x Zen 5 + 6x Zen 5c) Strix Point APUs.
• AMD Ryzen 9 8945H, Ryzen 7 8845HS (Hawk Point).
• Intel Core Ultra 9 285K & 275, Core Ultra 7 265K & 255, Core Ultra 5 245K & 240 (Arrow Lake).
• Intel Core Ultra 9 288V ; Ultra 7 268V, 266V, 258V, 256V ; Ultra 5 236V, 228V, 2266V (Lunar Lake).
• AMD Radeon RX 7600 XT (Navi 33 XT).
• New Chinese translation.

 

[thestryker]

There are Lunar Lake SKUs:

Core Ultra 9 288V ;
Core Ultra 7 268V,
Core Ultra 7 266V,
Core Ultra 7 258V,
Core Ultra 7 256V ;
Core Ultra 5 236V,
Core Ultra 5 228V,
Core Ultra 5 226V (this specifics part has a typo error in it's name)

It's looking like you were spot on before about the 6/8 at the end signifying DRAM amount.

Core Ultra 9 285K
Core Ultra 9 275,
Core Ultra 7 265K
Core Ultra 7 255,

Kind of curious what the significance of the second number of the model being different is if anything. It could just be a differentiation to make it less confusing for customers (eg thinking a 13900 is as fast as a 13900K), but companies aren't really known for caring a whole lot about that.

I'm also curious if the tiled architecture will mean the death of the F SKU or if Intel will release some without an IGP. I cannot imagine the packaging failure rates on just the IGP will be high enough on their own to make for a whole F SKU line.

 

[Deleted member 2731765]

The real question is why there would be any NPU in ARL-S given that it is a part of the SoC tile. They're certainly not just reusing the MTL tile since it had low power E-cores in it.

Exactly my thoughts. I was also about to point this out.

It makes little sense to put a SoC tile on a desktop part. So having an NPU seems a bit off to me. Like I mentioned above, I have doubts on these tweets being legit or not. I'm not buying all this leaked Arrow Lake info as of now.

Will wait for some official leaked slide or document instead.

But in any case, FWIW, the same leaker also claimed that Arrow Lake "refresh" CPUs which might incorporate a bigger and faster NPU within the SOC tile. The leaker states that this NPU will be an upgraded version from the one featured on the upcoming Arrow Lake-S.

So with the upgraded NPU, Intel Arrow Lake refresh chips could see a larger die size, expanding by 2.8mm compared to current-gen Arrow Lake chips.

While the die size will be larger, the package size will be the same. So the chips won't require major socket changes (BGA and LGA). They will retain the compatibility but motherboard vendors will have to enable Fast Voltage Mode (FVM) on the VccSA rails.

Makes sense, eh ? To me, it doesn't. But here is the tweet anyway.

https://twitter.com/x/status/1808621251965169774

View: https://x.com/jaykihn0/status/1808621251965169774


EDIT:

Ouch, Break Time. Head spinning fast. Grabs a cappuccino ! 🥱

 

[thestryker]

Makes sense, eh ? To me, it doesn't. But here is the tweet.

I saw that one earlier and almost laughed out loud. There's a zero percent chance that Intel would put out a socketed CPU that may or may not work in existing motherboards. If it was dependent on something implemented at the board level it would be a mandatory part of the specification.

 

[usertests]

It makes little sense to put a SoC tile on a desktop part. So having an NPU seems a bit off to me. Like I mentioned above, I have doubts on these tweets being legit or not. I'm not buying all this leaked Arrow Lake info as of now.

Intel's tile strategy is all over the place, just look at Lunar Lake reducing the tiles and ditching LP E-cores.

I'm not sure it doesn't make sense to have a "SoC tile" on desktop? Can you expand on that? Because if they are moving to tiles on desktop from now on, they could subdivide the chip in all sorts of ways. All we have to go on for now are Meteor Lake and Lunar Lake.

NPU on desktop is going to be a thing simply to make sure it's everywhere as a minimum spec, and for power efficiency as long as it can get better TOPS/Watt than the iGPU. Seems absurd but it could be an "ESG" type of thing that governments want, and Intel will surely have 35W TDP 'T' variants of these chips being stuffed into OptiPlex, ThinkCentre, etc. business systems, giving them all NPUs.

They also may end up with some SKUs that have no iGPU but do have an NPU. You'd almost certainly want to use the discrete GPU for AI in that case, but as long as it's there, it might see some use somewhere. You could also use it in the background while leaving the GPU completely for something else like gaming.

So with the upgraded NPU, Intel Arrow Lake refresh chips could see a larger die size, expanding by 2.8mm compared to current-gen Arrow Lake chips.

Arrow Lake Refreshin' the NPU?! What happened to 8 P-cores + 32 E-cores?

 

[thestryker]

I'm not sure it doesn't make sense to have a "SoC tile" on desktop?

Given that as you say they are moving to a tiled architecture across the board there will likely be a SoC tile, but there doesn't necessarily need to be.

The MTL SoC tile included LP E-cores, NPU, media engine, display controller, memory controller and the IO similar to a desktop chipset. So realistically the only thing a desktop chip needs out of there is the media engine, display controller and memory controller. These could potentially be integrated into the IO tile instead and now they're down to 3 tiles from 4.

 

[watzupken]

Good riddance. Happy to get a processor without all these meaningless NPUs. I mean if AI is a priority, just add another dedicated GPU from say Nvidia, and its going to be way more powerful than what any CPU or SOC can provide.

 

[TechyIT223]

We also don't need any AI hardware processing unit in a gaming processor.

Useless AI TOPS crap for the masses should be kept best to the mobile chips certified for copilot + PC hype.

 

[TechyIT223]

My take on the performance bracket 😜

Ultra 9 285K = i9-14900K
Ultra 9 275 = i9-14900
Ultra 7 265K = i7-14700K
Ultra 7 255 = i7-14700
Ultra 5 245K = i5-14600K
Ultra 5 240 = i5-14400

 

[Deleted member 2731765]

Kind of curious what the significance of the second number of the model being different is if anything. It could just be a differentiation to make it less confusing for customers (eg thinking a 13900 is as fast as a 13900K), but companies aren't really known for caring a whole lot about that.

I'm also curious if the tiled architecture will mean the death of the F SKU or if Intel will release some without an IGP. I cannot imagine the packaging failure rates on just the IGP will be high enough on their own to make for a whole F SKU line.

A bit tricky to guess what the second number might denote in the hierarchy and the lineup. But it could be directly related to the P "core count" value of that particular SKU.

The number 8 could mean the total P core count value. In the case of 285K, assuming this is indeed the flagship part having an "8 P+16 E" core count configuration (24 cores), then it matches the second number which is 8.

For the Core Ultra 9 275, it's kind of hard to make a match. "7 P+ 13 E" core config for a 20 core SKU ? I'm not aware of any processor having 7 P cores though.

But, if we take the Core Ultra 7 265K, then it should be the "6 P+ 14 E" core count config part, sporting a total of 20 cores. Makes kind of sense here though. So again the number "6" might denote the P core count value. What do you say ?

Assuming the Core Ultra 5 245K is 14 core part, then the number 4 implies the number of P cores, so that would give a "4 P+ 10 E" core config part.

Few more theories are there, but they again make things confusing, because the nomenclature is entirely different than previous gen Raptor Lake desktop parts.

For example, it might be possible that both the Core Ultra 9 285K and Core Ultra 9 275 are 24 core SKUs, sporting the "8 P+16 E" core count configuration. One is a 125 Watts part, and the other is a 65 Watts SKU. So the number 8 will denote a 125W SKU.


Btw, yes I have also been hearing whispers about the death of F-class SKUs with Arrow Lake lineup. Can't say with surety whether a tilde architecture will lead to this though.

Core Ultra 9 285K
Core Ultra 9 275,
Core Ultra 7 265K
Core Ultra 7 255,

 

[Deleted member 2731765]

Core Ultra 9 285K
Core Ultra 9 275,

It seems probable that both the Core Ultra 9 285K and Core Ultra 9 275 are 24 core SKUs, sporting the "8 P+16 E" core count configuration. The K variant has a 125 Watts TDP, and the other is a 65 W SKU.

 

[Deleted member 2731765]

Arrow Lake Refreshin' the NPU?! What happened to 8 P-cores + 32 E-cores?

A 40 core count SKU was never really fully confirmed, but maybe it is still in the pipeline ? But wasn't MLID the only leaker to claim this SKU though ?

Not all of his leaks and predictions come out true btw. So what would be the best use case for a double Skymont E core count SKU, up from a 16 E core part ?

 

[thestryker]

A bit tricky to guess what the second number might denote in the hierarchy and the lineup. But it could be directly related to the P "core count" value of that particular SKU.

The number 8 could mean the total P core count value. In the case of 285K, assuming this is indeed the flagship part having an "8 P+16 E" core count configuration (24 cores), then it matches the second number which is 8.

For the Core Ultra 9 275, it's kind of hard to make a match. "7 P+ 13 E" core config for a 20 core SKU ? I'm not aware of any processor having 7 P cores though.

But, if we take the Core Ultra 7 265K, then it should be the "6 P+ 14 E" core count config part, sporting a total of 20 cores. Makes kind of sense here though. So again the number "6" might denote the P core count value. What do you say ?

Well I think it's fairly safe to say that with E-cores coming in 4 core clusters and there having been no partial cluster releases since their inception that E-cores are always going to be multiples of 4 until that changes. So that would imply 4/8/12/16 will be the only number of E-cores that we'll see. I've only seen what I'd consider solid evidence of ES in 24 and 20 core configuration which should be 8p/16e and 8p/12e.

I don't think the second number is directly related to cores, but maybe it's something along the lines of expected performance tier. Just based on the Videocardz rumor specs chart it seems like the last number might be what signifies a core count change within a specific branding tier. In that chart the only that carries a different core count is the 245K vs 240. Of course that chart is still rumor based and outside of the top SKUs we really don't have anything I'd consider to be solid evidence of.

So long as it doesn't need a wheel to determine what part is what I'll be happy though. 🤣

Another thing I was pondering today is regarding the CPU tile, because all RPL parts were the same 8p/16e die. I'm sure that was done for overall yield reasons, but with the tiles I could see mobile and desktop potentially sharing CPU tiles which might lead to some different low-mid tier SKUs.

It seems probable that both the Core Ultra 9 285K and Core Ultra 9 275 are 24 core SKUs, sporting the "8 P+16 E" core count configuration. The K variant has a 125 Watts TDP, and the other is a 65 W SKU.

I really hope this is accurate, because for all the stupid naming tech companies do this would actually be a positive for consumers. I always thought the non K SKUs should have been slightly lower to differentiate that they don't boost as high (ex: 13900K and then have a 13850) while maintaining their series branding.

 

[usertests]

So what would be the best use case for a double Skymont E core count SKU, up from a 16 E core part ?

That's a tough question. Other than benchmarks and bragging rights, there are some people who can use as many cores as they can get. Most people don't need them, but most people don't need a "Core Ultra 9" CPU either.

One consideration is that the thread count is apparently slashed. You get 32 threads with a 14900K, but if HT is disabled for the next 8+16 part, it's down to 24, less than AMD. With 8+32, they're back up to 40. Of course, thread count could become less meaningful if "rentable units" redefine how their CPUs manage work.

If Intel sticks with something similar to their current hybrid strategy several generations from now, their consumer desktop CPUs could end up with 12-16 P-cores and 64+ E-cores. Next-generation consoles could dictate the amount of P-cores.

Well I think it's fairly safe to say that with E-cores coming in 4 core clusters and there having been no partial cluster releases since their inception that E-cores are always going to be multiples of 4 until that changes. So that would imply 4/8/12/16 will be the only number of E-cores that we'll see.

I think they should be able to disable cores inside the E-core clusters if that's what you mean. For example, see the bottom of the barrel Alder Lake-N SKUs with only 2 cores enabled out of 8 (N50, x7213E, x7211E).

Why not on desktop? I assume it's part of their segmentation strategy. They disable an entire underperforming cluster, and its slice of the L3 cache at the same time.

 

[thestryker]

I think they should be able to disable cores inside the E-core clusters if that's what you mean. For example, see the bottom of the barrel Alder Lake-N SKUs with only 2 cores enabled out of 8 (N50, x7213E, x7211E).

Why not on desktop? I assume it's part of their segmentation strategy. They disable an entire underperforming cluster, and its slice of the L3 cache at the same time.

Seeing as they've never done it for any hybrid CPUs I assume there's a very good reason (could be technical or could be yield). Consider that there are 1p/4e, 2p/4e, 2p/8e, 4p/4e, 4p/8e, 6p/4e, 6p/8e, 8p/4e, 8p/8e, 8p/12e, and 8p/16e SKUs if it was viable they'd have done it already.

 

[TechyIT223]

It seems probable that both the Core Ultra 9 285K and Core Ultra 9 275 are 24 core SKUs, sporting the "8 P+16 E" core count configuration. The K variant has a 125 Watts TDP, and the other is a 65 W SKU.

This one would be a strange lineup though.

Two different CPUs having the same core counts?

 

[Deleted member 2731765]

Intel's Core Ultra 7 258V "Lunar Lake" CPU spotted in Geekbench 6 !! 🤠

Geekbench 6 Vulkan test The Arc 140V iGPU scored 34181 points. Same ballpark as the GTX 1650 laptop dGPU, and more faster than the 780M .

 

[Deleted member 2731765]

It's looking like you were spot on before about the 6/8 at the end signifying DRAM amount.

BTW, I was just wondering what would be the memory on package amount for SKU's ending in 4 though ? As you know, there are some SKUs ending with 4 as well.

Like e.g, 234V, 264V and 254V. There might be more parts in the pipeline as well. They just haven't been included in any early Linux support patches though.

 

[thestryker]

BTW, I was just wondering what would be the memory on package amount for SKU's ending in 4 though ? As you know, there are some SKUs ending with 4 as well.

Like e.g, 234V, 264V and 254V. There might be more parts in the pipeline as well. They just haven't been included in any early Linux support patches though.

Since the packages have been 64Gb/128Gb I really don't have a clue since the leaks have pointed towards those being bottom of the stack. It may even have just been a naming shift by Intel rather than different SKUs. To me 8 SKUs seems like a reasonable number since there's only 16/32GB 4p/4e parts, but if they had 24GB parts coming I could see more. Of course it wouldn't make sense to have 24GB parts using a "lower" number than 16GB parts. So I remain confused by anything ending in 4 at this point in time.