Intel Core Ultra 200 "Arrow Lake-S" Desktop Processor Core Configurations Surface

[P4-630]

That's wrong.

IPC.

 

[Tek-Check]

So... will i5 be the only one to host 20A process and all higher SKUs on TSMC N3 node?

 

[ARF]

So... will i5 be the o ly one to host 20A process and all higher SKUs on TSMC N3 node?

Err, intel 7 is 10nm, intel 4 is 7nm, intel 3 is 6nm, intel 2 (20) is what... 5nm, 4nm (in TSMC equivalent)?

 

[P4-630]

Err, intel 7 is 10nm, intel 4 is 7nm, intel 3 is 6nm, intel 2 (20) is what... 5nm, 4nm (in TSMC equivalent)?

Doesn't matter right? We should just all buy AMD!...

 

[Lycanwolfen]

I wish intel would just make a full on desktop performance CPU no E cores no Igpu's, A pure 8 or 12 or 16 core Desktop CPU. That's one thing Intel has missed over the years. Also I miss the pentium name all together. These core i and core ultra seriously come up with a better name. The naming convention is horrible these days even for AMD as well. My next upgrade might be a threadripper or xeon. I was a pure CPU no E cores and XE cores I want a pure hard core compute CPU that can really crunch them numbers.

 

[ARF]

Doesn't matter right?

It matters. We should know the truth.

We should just all buy AMD!...

Unless intel offers something considerably faster, energy efficient and relatively cheap.
I had in the past Core 2 Quad Q9450, it isn't if I wasn't forced somehow to buy the faster option at that time. Not that Phenom X4 was that bad... but still...
Was younger, and the to-help-AMD myself wasn't mature...

 

[Wirko]

I wish intel would just make a full on desktop performance CPU no E cores no Igpu's, A pure 8 or 12 or 16 core Desktop CPU. That's one thing Intel has missed over the years. Also I miss the pentium name all together. These core i and core ultra seriously come up with a better name. The naming convention is horrible these days even for AMD as well. My next upgrade might be a threadripper or xeon. I was a pure CPU no E cores and XE cores I want a pure hard core compute CPU that can really crunch them numbers.

Which Xeons are you thinking of, LGA1700 or LGA4677? Only the former achieve near-Core clocks but Intel also succeeded in making LGA1700 two-way incompatible with LGA1700: Core CPUs don't work workstation chipsets and Xeons don't work with desktop chipsets, not even the W680 and R680E.

 

[Noyand]

Err, intel 7 is 10nm, intel 4 is 7nm, intel 3 is 6nm, intel 2 (20) is what... 5nm, 4nm (in TSMC equivalent)?

Your hatred made you got the equivalency backwards: Intel 10nm was renamed Intel 7 because that’s a tsmc 7nm “equivalent” when it comes to density. You probably forgot about it already, but even when Intel was the leader you couldn’t make a comparison at equal numbers: their 32nm was closer to the competition 28nm

 

[AnarchoPrimitiv]

So will arrowlake desktop have a node advantage over Zen5 desktop? Isn't arrowlake on TSMC 3nm and Zen5 is on TSMC 4nm?

 

[Daven]

Your hatred made you got the equivalency backwards: Intel 10nm was renamed Intel 7 because that’s a tsmc 7nm “equivalent” when it comes to density. You probably forgot about it already, but even when Intel was the leader you couldn’t make a comparison at equal numbers: their 32nm was closer to the competition 28nm

I think he got the numbers right. Just the wording is a little off. Maybe something like this:

New Intel 7 = Intel 10 nm
New Intel 4 = Intel 7 nm
New Intel 3 = Intel 6 nm

But after the new Intel 3, the node names are a little more suspect. How could Intel go from new Intel 7 to four new nodes (4, 3, 20A, 18A) almost simultaneously (or at least in a one-year time frame)? The answer: not much is different between these four nodes and they are more equivalent to higher TSMC node names. For example, new Intel 20A is probably closer to a TSMC 4 nm, maybe almost TSMC 3 nm but not quite.

 

[N/A]

20A is the former 5nm. 4x density that of 7nm (former 10nm). Tsmc equivalent 2.5 or 25 angstroms. But iam afraid 20 A may have been cancelled.

 

[rv8000]

So we have between 10- 15% IPC increase over Raptor Lake but about a 5-10% decrease in clocks.

Intel Core Ultra 9 285K rumored to boost up to 5.7 GHz - VideoCardz.com

Flagship Core Ultra 9 SKU with 5.7 GHz boost A rumor shared by OneRaichu suggests clocks will be lower on next-gen desktop CPU series. The Intel Core Ultra 9 285K is currently known as the fastest upcoming Arrow Lake-S processor. The series is said to make use of Lion Cove Performance Cores and...

videocardz.com

So single thread improvement is in the 5-10% range.

Doing some really rough napkin math, strictly related to cb23, it’s looking really bad for MT performance on the i9 (~36,000-37,000). With losing HT, not even the 50% IPC improvement to the ecores is going to make enough difference. I hope there’s some extra sauce due to negated mitigations from HT removal.

 

[kondamin]

Starting to get tiered of the constant drip of information and wish it went like it did 20 years ago.
Nothing, product presentation, for sale.

 

[InVasMani]

Doing some really rough napkin math, strictly related to cb23, it’s looking really bad for MT performance on the i9 (~36,000-37,000). With losing HT, not even the 50% IPC improvement to the ecores is going to make enough difference. I hope there’s some extra sauce due to negated mitigations from HT removal.

Yeah I think it'll take a hit to peak MT, but overall it seems like a fair trade off for a pretty substantial ST uplift. I think consistency should be a bit better too between P cores and E cores which really makes a lot of complaints around core types a more muted talking point. Like I really think we'll just start seeing more core types in clusters a lot like a food pyramid in layers hitting different power target envelopes with trade-offs shifting between ST/MT.

As power target drops MT uplift increases, but at lower peak ST is how I suspect they will arrange it. I think the goal is ultimately more linear and granularity with less compromises in between. That said they've got to contend with consumers needs and desires and so Intel's trying to bump up relative E core performance closer and closer to P core performance and both are being pushed to their knees a bit at the same time.

We've got 3 core types effectively now, but I reckon the next big shift will be inserting another 1-2 in between layers with their own intrinsic sets of relative strengths and weaknesses.

 

[rv8000]

Yeah I think it'll take a hit to peak MT, but overall it seems like a fair trade off for a pretty substantial ST uplift. I think consistency should be a bit better too between P cores and E cores which really makes a lot of complaints around core types a more muted talking point. Like I really think we'll just start seeing more core types in clusters a lot like a food pyramid in layers hitting different power target envelopes with trade-offs shifting between ST/MT.

As power target drops MT uplift increases, but at lower peak ST is how I suspect they will arrange it. I think the goal is ultimately more linear and granularity with less compromises in between. That said they've got to contend with consumers needs and desires and so Intel's trying to bump up relative E core performance closer and closer to P core performance and both are being pushed to their knees a bit at the same time.

We've got 3 core types effectively now, but I reckon the next big shift will be inserting another 1-2 in between layers with their own intrinsic sets of relative strengths and weaknesses.

Too many more core types on a single cpu seems like a scheduling nightmare. More typical users also don’t need ridiculous core counts if I’m being honest, the very strong ST should be great.

Overall I’m surprised they didn’t bump p/e-core counts for the i9. The waffling release date is also worrying to me, December would be problematic if true.

 

[InVasMani]

Too many more core types on a single cpu seems like a scheduling nightmare. More typical users also don’t need ridiculous core counts if I’m being honest, the very strong ST should be great.

Overall I’m surprised they didn’t bump p/e-core counts for the i9. The waffling release date is also worrying to me, December would be problematic if true.

Well I think in general a ideal multipurpose CPU should be flexible in terms of ST/MT and power usage across many power levels. That isn't to say it's the best CPU for a specific need by contrast like X3D that serves more of a niche purpose well. I do agree though not everyone needs a heap of MT and have even argued that point. Intel's got plenty of flexibility room to cater to more than one target audience though. There are inherent differences to P cores and E cores in design that make them relatively better and worse subjectively speaking though for intended usage. That's where more layers of core types and Intel's approach has some merit to it though. Having a layered tiered approach of different core types fitting different niche target demographic needs is kind of practical.

Is the scheduling a little more complex sure, but that can be sorted out. Games were once ST. We overcame that. Give developers some credit they can figure it out. I mean worrying about scheduling in AI era is especially strange. Realistically the AI will be able to end up figuring it out probably better than actual developers have done so when fed the right input data to crunch the math.

 

[Lycanwolfen]

Which Xeons are you thinking of, LGA1700 or LGA4677? Only the former achieve near-Core clocks but Intel also succeeded in making LGA1700 two-way incompatible with LGA1700: Core CPUs don't work workstation chipsets and Xeons don't work with desktop chipsets, not even the W680 and R680E.

Asus w790 ace seem to be a good board for the LGA 4677. My last upgrade I went from a Core i7 3770K to a Core i9 9900 KFC, Was a huge improvement overall and still is my current gaming box works quite well. But The latest Intel stuff does not impress me at all. I have heard stable issues with the 13th gen and 14 gen. They just do not get it. Look if intel wants to make desktop CPU's go ahead make 8P+16E things for Ai and whatever but in terms of Gaming We need pure Performance cores period. Intel should just make Gaming CPU's and Desktop CPU's split them up. Ryzen is looking like my next big upgrade change. Seems AMD is finally spot on fixed all their problems from the Past of stable issues.

 

[lexluthermiester]

That's wrong. They remove the multi-threading support simply because their architecture and manufacturing processes are power hogs, in order to free some TDP headroom.
The SMT will always be an important feature.

We will see how the top Ultra 9 28-thread will compete with the 32-thread Ryzen 9.

LOL! You keep thinking that.

 

[Minus Infinity]

I wish intel would just make a full on desktop performance CPU no E cores no Igpu's, A pure 8 or 12 or 16 core Desktop CPU. That's one thing Intel has missed over the years. Also I miss the pentium name all together. These core i and core ultra seriously come up with a better name. The naming convention is horrible these days even for AMD as well. My next upgrade might be a threadripper or xeon. I was a pure CPU no E cores and XE cores I want a pure hard core compute CPU that can really crunch them numbers.

You missed the story about Batrlett Lake next year with 12P + 0E config for i9?

 

[mahirzukic2]

So we have between 10- 15% IPC increase over Raptor Lake but about a 5-10% decrease in clocks.

Intel Core Ultra 9 285K rumored to boost up to 5.7 GHz - VideoCardz.com

Flagship Core Ultra 9 SKU with 5.7 GHz boost A rumor shared by OneRaichu suggests clocks will be lower on next-gen desktop CPU series. The Intel Core Ultra 9 285K is currently known as the fastest upcoming Arrow Lake-S processor. The series is said to make use of Lion Cove Performance Cores and...

videocardz.com

So single thread improvement is in the 5-10% range.

You mean 0-10%?

 

[phints]

cant wait to see what these cpus can churn out

Same, finally a new lithography and updated architecture. Hope things improve for Intel. For now I'm expecting to do a Ryzen 9800X3D but certainly open to one of these new Intel CPUs if they deliver.

Err, intel 7 is 10nm, intel 4 is 7nm, intel 3 is 6nm, intel 2 (20) is what... 5nm, 4nm (in TSMC equivalent)?

This is wrong, you would have to look at transistor density. The nm ratings mean nothing these days since the average feature sizes are still like 20nm for both Intel and TSMC. The renaming Intel did is completely reasonable just to have competitive naming convention.

 

[TheinsanegamerN]

Greed and stupidity. Intel thinks they can make more money changing sockets every few years. AMD has proven them wrong with AM4 being VERY profitable and lasting, what, 8 years now, almost 9?

I don't know if you've checked lately, but Intel still makes WAY more money then AMD, despite being in a serious rut.

Speaking of stupidity, remember all the issues AMD had supporting their CPUs across AM4 motherboards? How the 300 series still doesn't fully support the 5000s?

Guess who doesn't have that issue.

 

[lexluthermiester]

I don't know if you've checked lately, but Intel still makes WAY more money then AMD, despite being in a serious rut.

Never said they didn't. Just said it was greedy and stupid.

Speaking of stupidity, remember all the issues AMD had supporting their CPUs across AM4 motherboards? How the 300 series still doesn't fully support the 5000s?

And? Most boards made for the first gen Ryzen can and do run Ryzen 5000 series CPUs. So there were some glitch and hickups, which were usually swiftly solve. So who cares..

Guess who doesn't have that issue.

Who give a flying rats bum? At least AMD is trying/succeeding. I'm still an Intel guy, but I own AMD Ryzen and admire them.

 

[Wirko]

Guess who doesn't have that issue.

I (kind of) don't. Intel never promised me full compatibility of all LGA1151 boards with all LGA1151 processors, and it kept its non-promise.

 

[RandallFlagg]

Err, intel 7 is 10nm, intel 4 is 7nm, intel 3 is 6nm, intel 2 (20) is what... 5nm, 4nm (in TSMC equivalent)?

Yep, and TSMC N7 is 10nm, N5 is 7nm, and N3 is 5nm.

TSMC’s 7nm, 5nm, and 3nm “are just numbers… it doesn’t matter what the number is”

"These numbers are just numbers. They're like models in a car - it’s like BMW 5-series or Mazda 6."

www.pcgamesn.com