Intel's 8P+32E "Arrow Lake Refresh" Rumored Cancelled

[btarunr]

Intel's new 8P+16E "Arrow Lake" silicon powers not just the company's upcoming Socket LGA1851 desktop processors as the "Arrow Lake-S," but also its enthusiast mobile processors, as the "Arrow Lake-HX." The company is debuting this silicon later this year with the Core Ultra 200-series, which will see a ramp through 2025 across its desktop and mobile variants, but Intel is (or rather, was) slated to introduce an "Arrow Lake Refresh" microarchitecture as a follow-up, with a late-2025 debut and 2026 ramp. According to Olrak29, a reliable source with Intel leaks, this refresh has been cancelled.

The "Arrow Lake Refresh" microarchitecture would have been unlike "Raptor Lake Refresh," which was essentially the same silicon, but powering speed bumped SKUs under a newer 14th Gen Core branding, it would have been a physically larger chip, with more cores. Specifically, Intel would give the "Arrow Lake Refresh" an 8P+32E configuration, with eight "Lion Cove" performance cores, and eight "Skymont" E-core clusters for 32 efficiency cores. Intel claimed to have made huge strides with increasing the IPC of its E-cores with "Skymont," and the 8P+32E "Arrow Lake Refresh" would have been a multithreaded performance monstrosity.



There are many reasons Intel could have cancelled the 8P+32E "Arrow Lake Refresh." It probably finds its maxed out Core Ultra 9 285K competitive with the Ryzen 9 9950X in multithreaded productivity workloads; or it's simply cutting costs from having a new silicon release cycle, and will probably just release speed bumps for the 8P+16E silicon, like it did with the "Raptor Lake Refresh," regardless of how the 285K is competitive against the 9950X, or the possible 9950X3D.

Intel's public or leaked roadmaps see the "Nova Lake-S" succeed "Arrow Lake-S" on desktop, with its introduction slated for 2026, which should mean that the upcoming "Arrow Lake-S," and its possible refresh with an unchanged core-count, has to face AMD for at least 2 years. AMD recently released the "Zen 5" microarchitecture, which was received with mixed reviews, but could claw back with variants of the chip that feature 3D V-cache, under the Ryzen 9000X3D series. AMD confirmed that its succeeding "Zen 6" microarchitecture is on-track for a 2025 unveiling, with product launches expected either within 2025 or in 2026.

Meanwhile, Intel has bigger problems. Its underperformance in the stock markets means that the company is exposed to either amicable or hostile takeover moves, with the most recent utterance of the A-word coming from Qualcomm, according to a WSJ report.

View at TechPowerUp Main Site

 

[nguyen]

Yeah Intel is getting smarter, nobody like refreshes that bring nothing to the table except accelerated degradation

 

[TheinsanegamerN]

32 E cores,a multi threading "Beast" with all the performance of a rusty Yugo. Really, intel, push 10 P cores! If Arrow lake is as efficient as they claim 10 P cores shouldn't be hard to manage.

 

[phanbuey]

It looks like they're cutting down the release cycles and consolidating, which imo is a good thing.

14th gen was a total waste of time that could have been used for ARL-S or the successor.

 

[TheEndIsNear]

No more waste of sand

 

[Dristun]

Agent 47, you have a new task... The leakers have gone rampant - they are now cancelling the rumours they had previously created themselves. This has got to end, their meddling is no good for the sanity of our Agency. The methods are up to your choice. Good luck, 47.

 

[InVasMani]

32 E cores,a multi threading "Beast" with all the performance of a rusty Yugo. Really, intel, push 10 P cores! If Arrow lake is as efficient as they claim 10 P cores shouldn't be hard to manage.

Yeah, but until Intel has a answer to AMD's X3D variant CPU's a pure P core design isn't going to be very competitive. The E cores for better or worse at least give them flexibility to offer higher MT than AMD's been willing or able to offer within consumer level market of CPU's. A traditional P core design isn't enough with that few cores. It had a lot of P cores it might be enough to offset the X3D advantage AMD is able to offer. The question is how many P cores does it take for that to make sense and what's the cooling, power, and cost association of that for Intel relative to performance difference of what AMD is offering!?

Beyond that the platform itself has been holding Intel back. They switch sockets too frequently and their MB designs in certain regards these days have become inferior AMD offers more PCIE lanes and that's pretty important for future expand ability of discrete add in components for NVME devices to GPU's to network adapters and so on.

If they were to offer a 8P+32E chip on LGA1700 later on after node die shrink later on they'd probably have people on the platform interested by it quite a bit. In fact even a 12P+ 24E or a 8P + 24E would garner a lot of interest. That is of course is if Intel is winning to do some of the things AMD has done with AM4 to increase socket longevity.

Intel's in hole it's dug itself in the end. This mess is their own doing. If their unable to learn from their mistakes and change they can bury themselves inside of it. Intel wasted so much resources to tread water at 14nm trying desperate to look like it was competing more earnestly against AMD's Ryzen architecture and what it offered consumers in terms of value and productivity relative to the dumpster fire that Intel was offering at the time. Intel scrambled to add more cores and did so, but created hot power hungry chips in the process. Their still very much stuck in that scenario to some extent.

AMD's pretty much cornered the sweet spot of the market that the general consumer cares most about. Not everyone needs a hyper multi-core workstation beast and for just gaming beyond 16 cores or threads isn't very relevant given everything is ported from consoles primarily because why target just PC gamer's!!?

 

[RogueSix]

That would be kinda sad if true and another testament to how deep in shit Intel is. Intel Arrow Lake was originally supposed to be produced on Intel's 20A node before that node was scrapped/cancelled. Intel were forced to go with TSMC N3B for ARL because their own IFS is too incompetent to produce Arrow Lake.

I would have expected that Intel would bring the ARL refresh back home, i.e. ARL-R on the Intel 18A process node next year to demonstrate that the node is at least equal or more advanced than TSMC N3B.

So, my theory is that ARL-R has been cancelled specifically for that reason. Intel want to shy away from that comparison because they know exactly that TSMC N3B is kicking their 18A node's ass to hell and back. So much for catching up, or even outperforming, TSMC next year. Not gonna happen.

 

[Tomgang]

Forget all these E-cores. They fine, but 32 of them. Come on intel.

No give us more p-cores, that's what i want at least. 12 to 16 p-core would be fitting i think. 8 p-cores has been the same for like 3 gen now. Time to ad more.

 

[Vayra86]

That would be kinda sad if true and another testament to how deep in shit Intel is. Intel Arrow Lake was originally supposed to be produced on Intel's 20A node before that node was scrapped/cancelled. Intel were forced to go with TSMC N3B for ARL because their own IFS is too incompetent to produce Arrow Lake.

I would have expected that Intel would bring the ARL refresh back home, i.e. ARL-R on the Intel 18A process node next year to demonstrate that the node is at least equal or more advanced than TSMC N3B.

So, my theory is that ARL-R has been cancelled specifically for that reason. Intel want to shy away from that comparison because they know exactly that TSMC N3B is kicking their 18A node's ass to hell and back. So much for catching up, or even outperforming, TSMC next year. Not gonna happen.

You might be on the money here

 

[RogueSix]

14th gen was a total waste of time that could have been used for ARL-S or the successor.

?

I don't think that it took significant amounts of time to relabel the boxes . Intel 14th gen required zero engineering effort but only binning the 13th gen parts that could do 6GHz+. I don't really see a loss of time or resources with regard to Arrow Lake due to 14th gen RPL-R. They probably didn't even assign a B team but really the bottom of the barrel to RPL-R. It was one of the laziest "refreshes" ever.

I would even say it is quite the contrary: Intel 14th gen bought them (i.e. the real engineering teams and not the box labeling department) the time to improve upon Arrow Lake while they were fumbling to get it made on the now failed Intel 20A node.

 

[chrcoluk]

Yeah, but until Intel has a answer to AMD's X3D variant CPU's a pure P core design isn't going to be very competitive. The E cores for better or worse at least give them flexibility to offer higher MT than AMD's been willing or able to offer within consumer level market of CPU's. A traditional P core design isn't enough with that few cores. It had a lot of P cores it might be enough to offset the X3D advantage AMD is able to offer. The question is how many P cores does it take for that to make sense and what's the cooling, power, and cost association of that for Intel relative to performance difference of what AMD is offering!?

Beyond that the platform itself has been holding Intel back. They switch sockets too frequently and their MB designs in certain regards these days have become inferior AMD offers more PCIE lanes and that's pretty important for future expand ability of discrete add in components for NVME devices to GPU's to network adapters and so on.

If they were to offer a 8P+32E chip on LGA1700 later on after node die shrink later on they'd probably have people on the platform interested by it quite a bit. In fact even a 12P+ 24E or a 8P + 24E would garner a lot of interest. That is of course is if Intel is winning to do some of the things AMD has done with AM4 to increase socket longevity.

Intel's in hole it's dug itself in the end. This mess is their own doing. If their unable to learn from their mistakes and change they can bury themselves inside of it. Intel wasted so much resources to tread water at 14nm trying desperate to look like it was competing more earnestly against AMD's Ryzen architecture and what it offered consumers in terms of value and productivity relative to the dumpster fire that Intel was offering at the time. Intel scrambled to add more cores and did so, but created hot power hungry chips in the process. Their still very much stuck in that scenario to some extent.

AMD's pretty much cornered the sweet spot of the market that the general consumer cares most about. Not everyone needs a hyper multi-core workstation beast and for just gaming beyond 16 cores or threads isn't very relevant given everything is ported from consoles primarily because why target just PC gamer's!!?

I disagree on the platform, the platform situation is what led me to Z690.

Intel has 8 x gen 4 lanes on chipset, whilst AMD only has 4, and looking at the recent X870 boards, they horrible compared to my board, lots of things like use feature A, then feature B is disabled, and barely any PCIe slots. On top of that more expensive boards.

AMD does have more CPU lanes, but I think thats an inferior choice to having more chipset connectivity.

Also something else interesting about your post, is the U turn on productivity.

I remember when Zen 1 was released, in that era, AMD was better for productivity as it had more cores, especially with threadripper, whilst Intel was still the clear leader on games, that also applied to Zen 2. Back in those days, we had people going on about how important productivity is and per core performance (which is what gaming needs) is less important, I kid you not I read these posts often during Zen 1 and Zen 2 era. Now the two companies have basically done a switcheroo, the same people are saying productivity is pointless and gaming performance is king, it is these inconsistencies I puck up on. Now for what its worth I do prefer better gaming performance over better productivity, but my view has always been consistent on that, I said that 10 years ago, 5 years ago and am saying it again now. I dont flip flop based on what my favourite corporation is doing.

I think ultimately both vendors offer too many cores for gamers, no need for more than 8 cores, although there is benefits in offloading background tasks to spare cores to avoid scheduling bottlenecks, whether thats e-cores or extra performant cores.

Also what are these past mistakes that you refer to? I dont think Intel deliberately chose to use larger nodes, its been in the news enough times they have tried to get smaller nodes to work but hit problems. I think Intel's strategical mistakes have been wanted to use their own fab's instead of out sourcing which seems to be the better strategy and also releasing too many products, they need a longer product cycle.

Ultimately both Intel and AMD have had good and bad periods, AMD bulldozer was trash, they did the same thing, clocked these things up as high as mount everest to compensate, they were so aggressively clocked I had to downclock a friends FX chip to make it stable. It had degraded to instability at stock clocks within 2 years of purchase. Unlike Intel AMD didnt extend any warranties or do any investigation, it was just how the chips were.

But yeah Intel's issues are not the chipset in my view, in that area they better than AMD, although AMD sockets are currently lasting longer, the boards cost more to compensate.

I think Intel need to let go of these fab's and that might take some time, as no one likes doing a U turn on an investment.

 

[SOAREVERSOR]

Yeah, but until Intel has a answer to AMD's X3D variant CPU's a pure P core design isn't going to be very competitive. The E cores for better or worse at least give them flexibility to offer higher MT than AMD's been willing or able to offer within consumer level market of CPU's. A traditional P core design isn't enough with that few cores. It had a lot of P cores it might be enough to offset the X3D advantage AMD is able to offer. The question is how many P cores does it take for that to make sense and what's the cooling, power, and cost association of that for Intel relative to performance difference of what AMD is offering!?

Beyond that the platform itself has been holding Intel back. They switch sockets too frequently and their MB designs in certain regards these days have become inferior AMD offers more PCIE lanes and that's pretty important for future expand ability of discrete add in components for NVME devices to GPU's to network adapters and so on.

If they were to offer a 8P+32E chip on LGA1700 later on after node die shrink later on they'd probably have people on the platform interested by it quite a bit. In fact even a 12P+ 24E or a 8P + 24E would garner a lot of interest. That is of course is if Intel is winning to do some of the things AMD has done with AM4 to increase socket longevity.

Intel's in hole it's dug itself in the end. This mess is their own doing. If their unable to learn from their mistakes and change they can bury themselves inside of it. Intel wasted so much resources to tread water at 14nm trying desperate to look like it was competing more earnestly against AMD's Ryzen architecture and what it offered consumers in terms of value and productivity relative to the dumpster fire that Intel was offering at the time. Intel scrambled to add more cores and did so, but created hot power hungry chips in the process. Their still very much stuck in that scenario to some extent.

AMD's pretty much cornered the sweet spot of the market that the general consumer cares most about. Not everyone needs a hyper multi-core workstation beast and for just gaming beyond 16 cores or threads isn't very relevant given everything is ported from consoles primarily because why target just PC gamer's!!?

X3D is gaming only. PC gaming isn't serious. It's not what most computers are used for. It's not worth caring about compared to anything else you can use a computer for when you are trying to sell chips.

Consoles are targeted because PC gaming frankly sucks from the developer, publisher, and funding standpoint. Windows comes with all sorts of issues, the vast varriation of hardware makes it a pain in the ass, and you get more sales and money off consoles. So from a business standpoint you'd have to be a brain dead ferret on ketamine to prioritize the PC and you should be fired into to the sun when it comes to AAA titles. Games are made for consoles because that's what matters.

Intel's problem is not PC gaming. It's Xeon vs EPYC.

 

[chrcoluk]

Forget all these E-cores. They fine, but 32 of them. Come on intel.

No give us more p-cores, that's what i want at least. 12 to 16 p-core would be fitting i think. 8 p-cores has been the same for like 3 gen now. Time to ad more.

Yeah dont need 32, that is excessive.

 

[TheinsanegamerN]

X3D is gaming only. PC gaming isn't serious. It's not what most computers are used for. It's not worth caring about compared to anything else you can use a computer for when you are trying to sell chips.

Consoles are targeted because PC gaming frankly sucks from the developer, publisher, and funding standpoint. Windows comes with all sorts of issues, the vast varriation of hardware makes it a pain in the ass, and you get more sales and money off consoles. So from a business standpoint you'd have to be a brain dead ferret on ketamine to prioritize the PC and you should be fired into to the sun when it comes to AAA titles. Games are made for consoles because that's what matters.

Intel's problem is not PC gaming. It's Xeon vs EPYC.

Valve has over 200 million active users. PC versions of games frequently outsell the individual console version (playstation or xbox version). Gaming is a multi billion dollar per quarter industry.



But sure. Gaming isnt serious. That's why intel releases multiple K SKUs every generation and poured money into gaming GPUs.

Yeah, but until Intel has a answer to AMD's X3D variant CPU's a pure P core design isn't going to be very competitive. The E cores for better or worse at least give them flexibility to offer higher MT than AMD's been willing or able to offer within consumer level market of CPU's. A traditional P core design isn't enough with that few cores. It had a lot of P cores it might be enough to offset the X3D advantage AMD is able to offer. The question is how many P cores does it take for that to make sense and what's the cooling, power, and cost association of that for Intel relative to performance difference of what AMD is offering!?

Beyond that the platform itself has been holding Intel back. They switch sockets too frequently and their MB designs in certain regards these days have become inferior AMD offers more PCIE lanes and that's pretty important for future expand ability of discrete add in components for NVME devices to GPU's to network adapters and so on.

If they were to offer a 8P+32E chip on LGA1700 later on after node die shrink later on they'd probably have people on the platform interested by it quite a bit. In fact even a 12P+ 24E or a 8P + 24E would garner a lot of interest. That is of course is if Intel is winning to do some of the things AMD has done with AM4 to increase socket longevity.

Intel's in hole it's dug itself in the end. This mess is their own doing. If their unable to learn from their mistakes and change they can bury themselves inside of it. Intel wasted so much resources to tread water at 14nm trying desperate to look like it was competing more earnestly against AMD's Ryzen architecture and what it offered consumers in terms of value and productivity relative to the dumpster fire that Intel was offering at the time. Intel scrambled to add more cores and did so, but created hot power hungry chips in the process. Their still very much stuck in that scenario to some extent.

AMD's pretty much cornered the sweet spot of the market that the general consumer cares most about. Not everyone needs a hyper multi-core workstation beast and for just gaming beyond 16 cores or threads isn't very relevant given everything is ported from consoles primarily because why target just PC gamer's!!?

I never said it had to be a pure P variant. Intel COULD do one, their Xeons are pure P core. If Arrow lake is as efficient as they claim, they could do a 10 P core+ 4-8 E core design, and it'd likely win out ont he consumer end. The K series are already relatively competitive.

Yeah dont need 32, that is excessive.

Yeah, but it makes those benchmarks look really good, and cinebench is what we all run all day right : )

 

[Tomgang]

Yeah dont need 32, that is excessive.

Yeah. If any thing as i said, intel should ad more P-cores in stead. Specially now that they drop hyper-threading on p-cores.

 

[InVasMani]

I disagree on the platform, the platform situation is what led me to Z690.

Intel has 8 x gen 4 lanes on chipset, whilst AMD only has 4, and looking at the recent X870 boards, they horrible compared to my board, lots of things like use feature A, then feature B is disabled, and barely any PCIe slots. On top of that more expensive boards.

AMD does have more CPU lanes, but I think thats an inferior choice to having more chipset connectivity.

Also something else interesting about your post, is the U turn on productivity.

I remember when Zen 1 was released, in that era, AMD was better for productivity as it had more cores, especially with threadripper, whilst Intel was still the clear leader on games, that also applied to Zen 2. Back in those days, we had people going on about how important productivity is and per core performance (which is what gaming needs) is less important, I kid you not I read these posts often during Zen 1 and Zen 2 era. Now the two companies have basically done a switcheroo, the same people are saying productivity is pointless and gaming performance is king, it is these inconsistencies I puck up on. Now for what its worth I do prefer better gaming performance over better productivity, but my view has always been consistent on that, I said that 10 years ago, 5 years ago and am saying it again now. I dont flip flop based on what my favourite corporation is doing.

I think ultimately both vendors offer too many cores for gamers, no need for more than 8 cores, although there is benefits in offloading background tasks to spare cores to avoid scheduling bottlenecks, whether thats e-cores or extra performant cores.

Also what are these past mistakes that you refer to? I dont think Intel deliberately chose to use larger nodes, its been in the news enough times they have tried to get smaller nodes to work but hit problems. I think Intel's strategical mistakes have been wanted to use their own fab's instead of out sourcing which seems to be the better strategy and also releasing too many products, they need a longer product cycle.

Ultimately both Intel and AMD have had good and bad periods, AMD bulldozer was trash, they did the same thing, clocked these things up as high as mount everest to compensate, they were so aggressively clocked I had to downclock a friends FX chip to make it stable. It had degraded to instability at stock clocks within 2 years of purchase. Unlike Intel AMD didnt extend any warranties or do any investigation, it was just how the chips were.

But yeah Intel's issues are not the chipset in my view, in that area they better than AMD, although AMD sockets are currently lasting longer, the boards cost more to compensate.

I think Intel need to let go of these fab's and that might take some time, as no one likes doing a U turn on an investment.

I'd say in that era of early Ryzen productivity was a bigger matter than it is now for PC technology where as today's CPU productivity has grown exponentially. From a consumer standpoint I don't see it as nearly as pivotal to the average consumer than it was when Ryzen arrived and we'd been stuck on a stagnated 4c paradigm for eternity as mostly the only realistic option. Intel had some workstation stuff, but it was was both a lot more expensive than Ryzen and while it had additional cores they were clocked appallingly low.

I guess I was a bit incorrect on the platform side of things or partially/impartially it's mixed baggage with platforms of both. It's particularly complex with the more recently launched AMD chipsets that are more cut down, but more affordable likewise. Socket longevity still certainly benefits AMD consumers though in a way that's hard to refute. It's something Intel has the option to do something about, but a issue of is management willing to do so. Based on past history it's unlikely especially when they've already moved on yet again with a new socket replacement.

I'm not saying only Intel's made mistakes, but they've had their sets of blunders in the last decade to half decade. Remember Comet Lake and Rocket Lake for example where they had to go from 10c to 8c!!? That was on top of both being underwhelming on efficiency and sockets with no longevity. There are the host of vulnerabilities they got hit with as well though AMD's had a handful of as well, but doesn't seem like they've gotten as hammered by those and mitigating them has worse it seems on Intel side.

Intel has been treading water much too long in the grand scheme so it's no wonder AMD's gained a sizable presence over time relative to them.

In terms of gaming there is reason to target between 8-16 C/T because that's what the current crop of consoles are offering. There isn't much practical upside to offer more for the same reason though. It can be nice to have a bit more beyond it for varying reasons and for reasons outside of gaming. If you're focus is gaming though you probably aren't seeing a lot of practical benefit aside from maybe if you also do a bit of game streaming and want to offload that with a few more available cores.

 

[chrcoluk]

Yeah this is why I think 32 e-cores is ridiculous, I think for a p-core only chip I would want 10-12 cores, so I have cores to utilise for background, and then 8 cores dedicated to gaming. e-cores I think 4 would be enough but maybe up to about 8 maximum if there is no HTT. The only thing that comes in to question of course is Intel are getting rid of HTT, I have the opinion HTT is outdated, but I do accept there may be games coded to expect to be able to utilise more than 8 threads, this might then need more than 8 p-cores or a better scheduler to handle such situations and move those extra threads to idle e-cores. If that is handled well then removing HTT should work well, if its handled badly it might be a mess.

 

[InVasMani]

I think as you close the gap difference between P cores and E cores in terms of frequency the scheduling concerns diminish pretty readily. As for HT removal that'll reduce MT a bit, but could be offset a bit with some supplemental E cores potentially. What I think would help Intel's flexibility a good bit if is the E core clusters were smaller instead of 4c per cluster why not 3c or 2c which provides them with a bit more leeway on design options in turn. Beyond that smaller clusters per E core die would result in more clusters of tied to multipliers rather than fewer clusters tied to multipliers which offers more granularity in terms of dynamically tuning multipliers to free up a bit of thermal headroom or scale relative efficiency and performance a bit one way or another.

 

[ir_cow]

Intel is returning to its roots. 1 CPU per socket none sense.

 

[persondb]

Yeah, but until Intel has a answer to AMD's X3D variant CPU's a pure P core design isn't going to be very competitive. The E cores for better or worse at least give them flexibility to offer higher MT than AMD's been willing or able to offer within consumer level market of CPU's. A traditional P core design isn't enough with that few cores. It had a lot of P cores it might be enough to offset the X3D advantage AMD is able to offer. The question is how many P cores does it take for that to make sense and what's the cooling, power, and cost association of that for Intel relative to performance difference of what AMD is offering!?

I believe you are overstating the advantage of the X3D. Those are focused for gaming and often have worse performance than their non-X3D counterpart in other things, if you want to have a computer that is good for multiple things and not just game, it's become a lot harder and not just X3D is better.

And that is if you even need that performance from the X3D in the first place, for most users their bottleneck will likely be in the graphics card and that would have a much bigger influence than the CPU, be it X3D vs non-X3D. CPU gaming benchmarks are done with the best available GPU at lower resolution, so that differences in CPU SKUs actually show themselves.

If you were to use a 3060 on something like that, there would probably be a flatline in quite a few CPU models as the CPU pushes as many frames as the GPU can handle and then gets stuck.

So I do not think that X3D is anything to be considered for Intel strategy. It's a great product that gives value to the people who need exactly that but it's pretty niche. The cost of designing a new die and getting it to production to compete with that would probably outweigh the sales generated by it.

AMD can manage to do it cheaply because their normal chips already include all the circuitry necessary, they just decide if they want to add the 3d SRAM die or not.

Beyond that the platform itself has been holding Intel back. They switch sockets too frequently and their MB designs in certain regards these days have become inferior AMD offers more PCIE lanes and that's pretty important for future expand ability of discrete add in components for NVME devices to GPU's to network adapters and so on.

For the LGA1700, they offer about the same number of PCIe lanes as AMD, the difference is that they put more of them into the chipset connection. This has it's benefits and drawbacks.

The benefits is that it's a more versatile allocation of resources as the chipset can offer a lot more lanes downstream. The drawback of course, is that those lanes end up sharing the chipset lane bandwidth upstream.
This is however mitigated by the fact that devices connected by the chipset are usually not using the full bandwidth 100% of the time. But there can be issues, like how people often complain about USB audio from the chipset and how when it gets busy, their audio device can stutter due not being allocated the bandwidth. This is far more common with x4 connectivity as a single PCIe SSD could fill the chipset connectivity. With x8 that Intel has for their Z-series chipset, this would be far less common as you need an increasing amount of devices active at the same time.

For network adapters and such pcie devices, those are better handled by chipset lanes in either case as it would be a waste to use a PCIe 4 or 5 lane on those in the majority of cases.

Intel however generally fails at middle and below chipsets as the B and H series ones generally don't have a lot of niceties. For the 600/700 series, they are probably inferior to AMD B650 and such due to those having an aditional 4 PCIe lanes.

The new 800 series chipset by the leaks is going to be superior to AMDs though. With 24 lanes from the CPU(20 being 5.0 and 4 being 4.0), x8 chipset connectivity and 2 TB4/USB 4 from the CPU(don`t need to use 4 PCIe lanes). Those are for the Z890 though, the B860 and H810 are going to be the same, except get a single TB4/USB4 port from the CPU.

Intel's Upcoming 800-series Chipsets Leak in Detail

What appears to be the full details of all of Intel's upcoming 800-series chipsets have leaked and although we've already seen what the Z890 chipset will offer, no details of the other SKUs have leaked to date. Maybe the biggest news is that there won't be an H870 chipset this time around...

www.techpowerup.com

AMD's pretty much cornered the sweet spot of the market that the general consumer cares most about. Not everyone needs a hyper multi-core workstation beast and for just gaming beyond 16 cores or threads isn't very relevant given everything is ported from consoles primarily because why target just PC gamer's!!?

I think you mean 8 cores or 16 threads(SMT). As it`s beyond that which isn`t relevant for gaming.
Gamers will likely stick with the 8 cores Ryzen 7s, they don't really get anything by getting 12 or 16 cores Ryzen 9s, aside from the higher clock that those parts tend to have.

And from that, I don't think that Intel strategy of P+E is that inferior to AMD's as it largely doesn't matter. Though ironically enough that strategy brought them similar scheduling problems that AMD has.

 

[TumbleGeorge]

It looks like they're cutting down the release cycles

Extend.

 

[kapone32]

Yeah, but until Intel has a answer to AMD's X3D variant CPU's a pure P core design isn't going to be very competitive. The E cores for better or worse at least give them flexibility to offer higher MT than AMD's been willing or able to offer within consumer level market of CPU's. A traditional P core design isn't enough with that few cores. It had a lot of P cores it might be enough to offset the X3D advantage AMD is able to offer. The question is how many P cores does it take for that to make sense and what's the cooling, power, and cost association of that for Intel relative to performance difference of what AMD is offering!?

Beyond that the platform itself has been holding Intel back. They switch sockets too frequently and their MB designs in certain regards these days have become inferior AMD offers more PCIE lanes and that's pretty important for future expand ability of discrete add in components for NVME devices to GPU's to network adapters and so on.

If they were to offer a 8P+32E chip on LGA1700 later on after node die shrink later on they'd probably have people on the platform interested by it quite a bit. In fact even a 12P+ 24E or a 8P + 24E would garner a lot of interest. That is of course is if Intel is winning to do some of the things AMD has done with AM4 to increase socket longevity.

Intel's in hole it's dug itself in the end. This mess is their own doing. If their unable to learn from their mistakes and change they can bury themselves inside of it. Intel wasted so much resources to tread water at 14nm trying desperate to look like it was competing more earnestly against AMD's Ryzen architecture and what it offered consumers in terms of value and productivity relative to the dumpster fire that Intel was offering at the time. Intel scrambled to add more cores and did so, but created hot power hungry chips in the process. Their still very much stuck in that scenario to some extent.

AMD's pretty much cornered the sweet spot of the market that the general consumer cares most about. Not everyone needs a hyper multi-core workstation beast and for just gaming beyond 16 cores or threads isn't very relevant given everything is ported from consoles primarily because why target just PC gamer's!!?

That may be why they gave up on Hyperthreading.

I believe you are overstating the advantage of the X3D. Those are focused for gaming and often have worse performance than their non-X3D counterpart in other things, if you want to have a computer that is good for multiple things and not just game, it's become a lot harder and not just X3D is better.

And that is if you even need that performance from the X3D in the first place, for most users their bottleneck will likely be in the graphics card and that would have a much bigger influence than the CPU, be it X3D vs non-X3D. CPU gaming benchmarks are done with the best available GPU at lower resolution, so that differences in CPU SKUs actually show themselves.

If you were to use a 3060 on something like that, there would probably be a flatline in quite a few CPU models as the CPU pushes as many frames as the GPU can handle and then gets stuck.

So I do not think that X3D is anything to be considered for Intel strategy. It's a great product that gives value to the people who need exactly that but it's pretty niche. The cost of designing a new die and getting it to production to compete with that would probably outweigh the sales generated by it.

AMD can manage to do it cheaply because their normal chips already include all the circuitry necessary, they just decide if they want to add the 3d SRAM die or not.

For the LGA1700, they offer about the same number of PCIe lanes as AMD, the difference is that they put more of them into the chipset connection. This has it's benefits and drawbacks.

The benefits is that it's a more versatile allocation of resources as the chipset can offer a lot more lanes downstream. The drawback of course, is that those lanes end up sharing the chipset lane bandwidth upstream.
This is however mitigated by the fact that devices connected by the chipset are usually not using the full bandwidth 100% of the time. But there can be issues, like how people often complain about USB audio from the chipset and how when it gets busy, their audio device can stutter due not being allocated the bandwidth. This is far more common with x4 connectivity as a single PCIe SSD could fill the chipset connectivity. With x8 that Intel has for their Z-series chipset, this would be far less common as you need an increasing amount of devices active at the same time.

For network adapters and such pcie devices, those are better handled by chipset lanes in either case as it would be a waste to use a PCIe 4 or 5 lane on those in the majority of cases.

Intel however generally fails at middle and below chipsets as the B and H series ones generally don't have a lot of niceties. For the 600/700 series, they are probably inferior to AMD B650 and such due to those having an aditional 4 PCIe lanes.

The new 800 series chipset by the leaks is going to be superior to AMDs though. With 24 lanes from the CPU(20 being 5.0 and 4 being 4.0), x8 chipset connectivity and 2 TB4/USB 4 from the CPU(don`t need to use 4 PCIe lanes). Those are for the Z890 though, the B860 and H810 are going to be the same, except get a single TB4/USB4 port from the CPU.

Intel's Upcoming 800-series Chipsets Leak in Detail

What appears to be the full details of all of Intel's upcoming 800-series chipsets have leaked and although we've already seen what the Z890 chipset will offer, no details of the other SKUs have leaked to date. Maybe the biggest news is that there won't be an H870 chipset this time around...

www.techpowerup.com

I think you mean 8 cores or 16 threads(SMT). As it`s beyond that which isn`t relevant for gaming.
Gamers will likely stick with the 8 cores Ryzen 7s, they don't really get anything by getting 12 or 16 cores Ryzen 9s, aside from the higher clock that those parts tend to have.

And from that, I don't think that Intel strategy of P+E is that inferior to AMD's as it largely doesn't matter. Though ironically enough that strategy brought them similar scheduling problems that AMD has.

Games like City Skylines 2 or the new Space Marine 2 show that CPU performance matters.

 

[persondb]

Games like City Skylines 2 or the new Space Marine 2 show that CPU performance matters.

Yes there are games like that, most of them are not though.

And also the specific topic was about the core counts.

I do not know about City Skylines 2 but Space Marines 2 do not seem to care a lot about core counts, but rather about the core performance and cache. Noting that a single core from the 3700X has access to 16MB of L3 as Zen 2 is divided into clusters of CCX with 4 cores each.

8 cores seem to be enough really, which was kinda of the point. Maybe in City Skylines 2 or BeamNG your 16 cores CPU has an advantage but in general, it's really not that, so those big core count CPUs aren't for people who are only going to game.

 

[InVasMani]

I'm betting on developers not all wanting to knee cap themselves indefinitely into the future purely based around current standards. Even so you don't have a lot to worry about and concern yourself with if you're at least meeting what consoles are currently pushing and targeting. It's easy enough to look at from both angles the rest is just a matter of cost and other usage considerations.