Intel Rebadges 10nm Enhanced SuperFin Node as "Intel 7," Invents Other Creative Node Names

[persondb]

Why all the bashing? It's really just aligning the names to be closer to the competitors.

Nodes naming have been BS for years and if we are going to be mad about it then be mad at all foundries. I guess people are just too much in the Intel hate bandwagon.

 

[renz496]

waste of time really, still might confuse avg consumers

most often average consumer did not care about such details. so there is nothing to confuse about.

 

[windwhirl]

The only thing I have to say about this is that all these node names are just bullshit marketing. These days, there's no chip feature that you can pick and say "this is why the node is called X nm, because this feature measures X nm". Not talking just about Intel but all the foundries with relatively advanced nodes (TSMC, Samsung, even GF regardless of them discarding their 7 nm node, etc).

At best, we can compare transistor density. That's about it since rarely if ever any design is ported to multiple foundries' nodes

 

[Richards]

So, about that.
These numbers are probably from SRAM cells and the high density/low power variation of the node. Basically maximum possible number which does not necessarily mean any real usage in that range.
In High performance variations density suffers. A lot. GPUs are probably the easiest way to get real-ish numbers, for example RDNA2 GPUs are around 50-51 MTr/mm2. RDNA1 was 40-41 MTr/mm2 (both an older version of N7 and likely RDNA2 gains some density from huge cache).

Edit:
Also, your quote has Samsung's 7nm number.
Samsung 8nm is an older 10nm half-node evolution with claimed density of 61.18 MTr/mm2.
The Samsung 8nm variation used for Nvidia Ampere GPUs has in that application density of 44-45MTr/mm2.

Yeah all these companies get 78 or 80 % of the peak advertising density no company gets 100% yields

 

[watzupken]

They're not wrong in doing this. The "Xnm" moniker means nothing anymore, it isn't in any way representative of the transistors' geometry, it's just a purely commercial/marketing term. What matters is the density you can achieve on a process node, and in terms of density, Intel's 10nm is superior to both TSMC's 7nm and Samsung's 7nm.

Intel "10nm" - 100.76 MTr/mm2
Samsung "7nm" - 95.08 MTr/mm2
TSMC "7nm" - 91.2 MTr/mm2

TSMC has the smaller "nm" number and yet the worst density of all. Seems disingenuous if you ask me.

Just to make my point even stronger:

TSMC "10nm" - 52.51 MTr/mm2
Samsung "10nm" - 51.82 MTr/mm2

If anyone is trying to fool us is TSMC and Samsung. Intel should've aligned their naming scheme ages ago.

EDIT: Samsung 7nm

I don't disagree that the XXnm naming is misleading and not reflective of the actual product/node itself. However if the TSMC is inferior to Intel's node, there is no reason why Intel will want to use TSMC, particularly with their more cutting edge node.

 

[tabascosauz]

Why all the bashing? It's really just aligning the names to be closer to the competitors.

Nodes naming have been BS for years and if we are going to be mad about it then be mad at all foundries. I guess people are just too much in the Intel hate bandwagon.

Because while the rationale is reasonable, it's an ill-timed and absolutely knuckleheaded move by Intel and I'd be surprised if it doesn't gain hilarious amounts of bad press. 10++ becoming 10nm Enhanced Superfin was the only sensible renaming scheme that actually was a fresh start (though you could argue that at that point, 10+ and 10++ should already have simply become something like 9nm to distance them from the name). This is what, the 4th time Intel has renamed 10nm? Not all investors are necessarily tech-savvy, but after 4 rebranding efforts with nothing to show for it other than Tiger Lake and ICL Xeon (kinda not really), PR stunts like this do nothing to improve confidence.

No one's doubting that TSMC and Samsung love to give old shit a new coat of paint and call it by a new name, but when you can't even deliver compelling products, it doesn't matter if your node is denser or more efficient than the name suggests or that the competitors are misleading consumers. 10SF is much improved and dense as hell - so what? Tiger Lake is still a power hog and ICL Xeon isn't impressive.

At least begin the new naming scheme on the (former) 7nm node. A fresh start to forget about the 10nm debacle, if you will - I thought that was exactly what Pat aimed for when he became CEO? I get that they want to pull out all the stops for Alder Lake and Sapphire Rapids, but shit like this just reeks of a lack of confidence in Alder Lake.

That said, the news post in certain parts reads like a highly suspect opinion piece rather than news.

 

[Redwoodz]

Yes, my numbers are "stolen" from wikichip and represent raw cell-level density, so a best case scenario. We don't really have a common architecture implemented on all process nodes, so, we can't really compare anything else than these theoretical numbers. But, as you say, density suffers on complex chip designs, but it is the case for all process nodes, it's not like TSMC's 7nm is magic... The extent of the suffering, that's difficult to guestimate. Even if Intel's 10nm would be the most affected, it still has a 10%-ish initial advantage. Good catch on the Samsung 7nm.

Well all that means is that Intel's design really sucks if they have higher density and lower performance. They just eliminated their only viable excuse.

 

[TheLostSwede]

Dear Intel, the unit is Å, not A, as his name was Ångström. I guess your marketing department couldn't figure out how to type it, so now your new nodes are in Ampere...

 

[Tartaros]

I was hoping they would come up with and x and then an X and ended like up xXxXxXx like in early 00s IRC and msn.

 

[goodeedidid]

There is a problem with naming nodes this way. The marketing guys will try pushing deceptive titles. If the engineers protest the title, they will be ignored and/or fired.

Yeah sure they will fire their engineers because of some marketing guys who are disposable like diapers.

 

[R-T-B]

This is intel, they dont give a damn about people

Ah yes, I forgot, they sell only to robots.

 

[MDDB]

What puzzles me is why they would drop the "nm" in the next few nodes, only to reintroduce an "A" for angstroms a couple of years later. Good planning!

 

[ZoneDymo]

What puzzles me is why they would drop the "nm" in the next few nodes, only to reintroduce an "A" for angstroms a couple of years later. Good planning!

Eh naming schemes are always a mess where sooner or later they will become a problem so you have to start over.
Im suprised they actually went past 10 with the processor names, personally I dont have a problem with 12900k but I guess a lot of people do because usually that is the moment to start over.

 

[GURU7OF9]

I don't disagree that the XXnm naming is misleading and not reflective of the actual product/node itself. However if the TSMC is inferior to Intel's node, there is no reason why Intel will want to use TSMC, particularly with their more cutting edge node.

Now that is a very good point ! Couldnt have put it better myself! Times have changed !
I guess Intel are really feeling the heat these days ! They have been talking it up big time how good their future products will be! What about their current products ! Then having to rename their products to show how they compare to their competitors, must be a first for Intel ! Normally they just make the product and everyone else has to follow! Intel having to drop the price on their processors, must be a first !

I have to say i get sick of hearing about Intel this and that but at the end of the day whether their 10nm is better than TSMC 7nm etc or not show us some products! TSMC make products!
All good to sprout about it with nothing to show for it!

 

[Xuper]

allright , 5 Ghz CPU confirms on 10nm+ ?

 

[mtcn77]

You'll have to excuse Intel since it is still a proprietary foundry that can market each step of the way whereas TSMC is likely not going to issue a statement on which licensing partner develops what... TSMC's job is keeping secrets which is quite on the contrary with Intel.

 

[ZoneDymo]

emotions running high again I see, personally I think this is fine, nm is like megapixels for cameras, it means nothing, its just marketing.
and heck, even if the numbers were a representation of reality, who cares?

If Intel would make a chip on 4nm whereas AMD would be 7nm but the AMD chip consumes less power and is faster, then who cares that the Intel one would be build on a smaller node? nobody.
We associate a lower nm number with performance but ultimately we only care about that performance soooo....yeah who cares, just wait for reviews on the actual product.

 

[GreiverBlade]

actually, even tho i disliked Intel behavior these last years ("childish bashing on glue" "improvements giving them an edge which turned out to be vulnerabilities") ... i am not really surprised for that and i kinda agree with it ... the AMD PRating system was also legit since the rating effectively meant performances on par or above while running at a lower speed, i am not one to care about that "proprietary" naming for a node, if a 10/14nm (++++++++++++++++++++++++++++) compete with a smaller node...

although i am sure it will be misleading in some ways ... some will think "Intel is definitely better in all aspect since their bigger node are same as AMDs smaller node" and not look at the real result ... just like the 12900K beating a 5950X (kinda a pointless PR move ... next gen beating previous gen ... even tho it shows the good point of big core small core nonetheless)

the pricing and availability will always dictate for me ... not the PR

 

[Deleted member 24505]

actually, even tho i disliked Intel behavior these last years ("childish bashing on glue" "improvements giving them an edge which turned out to be vulnerabilities") ... i am not really surprised for that and i kinda agree with it ... the AMD PRating system was also legit since the rating effectively meant performances on par or above while running at a lower speed, i am not one to care about that "proprietary" naming for a node, if a 10/14nm (++++++++++++++++++++++++++++) compete with a smaller node...

although i am sure it will be misleading in some ways ... some will think "Intel is definitely better in all aspect since their bigger node are same as AMDs smaller node" and not look at the real result ... just like the 12900K beating a 5950X (kinda a pointless PR move ... next gen beating previous gen ... even tho it shows the good point of big core small core nonetheless)

the pricing and availability will always dictate for me ... not the PR

If a bigger node beats a smaller one, surely the bigger node is the better

 

[GreiverBlade]

If a bigger node beats a smaller one, surely the bigger node is the better

nonono no ... i did not mean it like that...

if it was like that i would have taken a 11600K to replace my 6600K and not a R5 3600 which albeit being previous gen still hold up to the 11600K (no... i do not consider a few fps more to be an abyssal gape or a "Intel destroy AMD in gaming!" validation )

i mean it like "IN ALL ASPECT" ... which in reality it is not the case... beating in PR slides is one thing... optimized and standardized "variable" do wonder... sure the 12900k beat a 5950X now ... but what would it be actually with the correct counterpart in a benchmark from a trusted reviewer and not just some PR without context and background. (it's Intel we are talking ... and "i disliked Intel behavior these last years" is a reflection on how low they dived once they got their challenge met with Ryzen)

edit: also yes the bigger node packing more transistor is bound to be better than a smaller node with less transistor ...

 

[Deleted member 24505]

nonono no ... i did not mean it like that...

if it was like that i would have taken a 11600K to replace my 6600K and not a R5 3600 which albeit being previous gen still hold up to the 11600K (no... i do not consider a few fps more to be an abyssal gape or a "Intel destroy AMD in gaming!" validation )

i mean it like "IN ALL ASPECT" ... which in reality it is not the case... beating in PR slides is one thing... optimized and standardized "variable" do wonder... sure the 12900k beat a 5950X now ... but what would it be actually with the correct counterpart in a benchmark from a trusted reviewer and not just some PR without context and background. (it's Intel we are talking ... and "i disliked Intel behavior these last years" is a reflection on how low they dived once they got their challenge met with Ryzen)

edit: also yes the bigger node packing more transistor is bound to be better than a smaller node with less transistor ...

in theory though a smaller should have more transistors than the bigger in a smaller package, smaller should always be better.

There's no doubt Intel was sore at losing their crown. Love em or hate em though, i have no doubt they will be back.

 

[londiste]

You'll have to excuse Intel since it is still a proprietary foundry that can market each step of the way whereas TSMC is likely not going to issue a statement on which licensing partner develops what... TSMC's job is keeping secrets which is quite on the contrary with Intel.

TSMC is quite actively marketing their nodes. For 7nm they did announce N7 (all the various stages), N7P, N7+ and now N6. Usually along with big partners who are the major users of each node.
Whether Intel wants to or not (while the obviously do want) they need to use the process improvements for marketing as long as AMD piggybacks on TSMCs marketing for the same.

 

[Aquinus]

in theory though a smaller should have more transistors than the bigger in a smaller package, smaller should always be better.

The issue with cramming more transistors into a small area is heat flux. You do need to remove the heat from these circuits and putting them closer together basically keeps power consumption similar (or lower,) but over a much smaller area. The end result is heat being more concentrated in a smaller area. This is actually an advantage of AMD's MCM design, because it spreads that thermal load across a larger area. The same issue exists with stacked circuits. Either way, I think the juice is typically worth the squeeze for the smaller node, but there are definitely tradeoffs.

 

[medi01]

Still a 10nm node

No, not really.

Intel 14nm transistor size (as measured by crazy German dude):
24 by 24 nm

TSMC 7nm
22 by 22 nm

 

[DeathtoGnomes]

As with all of Intel's marketing team, ideas are laid out like a hopscotch template and the dice are rolled. Its very common for team Intel marketing to not talk ignore the rest of Intel about what they are planning on doing.