Intel renames nodes: 10nm+ and 7nm become Intel 7 and 4

[Kenrou]

https://www.guru3d.com/news-story/intel-renames-nodes-10nm-and-7nm-become-intel-7-and-4.html

"Processes at Intel will no longer be measured in nanometers. To differentiate itself from rivals, the company is adopting new names that are intended to provide a "clearer picture." As a result, Intel's upcoming 10nm+ and 7nm nodes will be named Intel 7 and 4." - "Intel renames its upcoming 7nm process Intel 4 to avoid confusion with its predecessor, Intel 3."

I think the comments below said it best :

"What process do you use?"
"4"
"4 nm?"
"No just 4"
"Which is what?"
"7 nm actually"
"But you call it "4"?"
"Yes, it gives a clearer picture"

"Trying to get 10nm right gave intel ptsd didn't it..."

 

[dejo]

I think maybe the media has spoken to them and made it clear that this works with the public

 

[EarthDog]

I mean 14nm++++++++++++++++++++++++++++++++++ was a joke, but this...the numbers don't even match?

 

[mackerel]

Do I really want to go through this all over again? I posted this on another (much bigger) forum and it only just settled down. People's problems with the name change boil down to either fanboyism or a fundamental failure to understand what they're talking about.

 

[Voodoo Rufus]

The proof will be in their power efficiency and clocks. People have played numbers games with "equivalent clock speeds" in the past, also.

 

[EarthDog]

Do I really want to go through this all over again? I posted this on another (much bigger) forum and it only just settled down. People's problems with the name change boil down to either fanboyism or a fundamental failure to understand what they're talking about.

Or, the naming convention isn't telling........regardless of fanboyism or failure to understand?

Can you shed some light on how it makes sense to call a 10nm(regardless of +++++) 7 anything and 7nm 4? I mean, I get it.... it's simply putting a name in front that doesn't represent the process size. It marks improvements in the arch versus people poking at ++++++++ nomenclature. Clever marketing to not name the process node but simply use a number to mark different iterations/improvements.

In the end, I have to agree, because, 99% of people couldn't care less about the process node and only care about performance in the end. Personally, IDGAH(oot) about the node so long as the performance and temps are in order.

The proof will be in their power efficiency and clocks. People have played numbers games with "equivalent clock speeds" in the past, also.

Wasn't it AMD who listed their processor names according to Intel equivalent clocks back in the day? Athlon 64 2800+ for example was ~= P4 2.8 GHz or something along those lines.

 

[Janus67]

Absolutely amazing lol

 

[mackerel]

Can you shed some light on how it makes sense to call a 10nm(regardless of +++++) 7 anything and 7nm 4? I mean, I get it.... it's simply putting a name in front that doesn't represent the process size. It marks improvements in the arch versus people poking at ++++++++ nomenclature. Clever marketing to not name the process node but simply use a number to mark different iterations/improvements.

This is Intel copying the competition, and they are now playing the same numbers game that Samsung and TSMC have been doing for a while. The "nm" rating has diverged from any relation to physical aspects in the fabrication process a long time ago and today is nothing more than a marketing number. Intel might have been consistent within themselves, but the rest of the industry took a different route leading to the mismatch.

We can roughly put people into 3 groups:
1, If you don't know or care about process technology, none of this matters.
2, If you have heard of process technology but never looked into it deeply, this is the group that will benefit most from this naming change because they were not sufficiently informed to understand similar looking numbers were not directly comparable.
3, If you have a good understanding of it, this wont matter.

This is mainly to help with the 2nd group, maybe 3rd indirectly as a quality of life thing, to save them re-explaining it to group 2 every time it pops up.


I suppose the more fundamental problem is that people tend to want a single number to represent how "good" something is, and that most measures can't be distilled to a single number like that. Intel was losing out because their number which was not directly comparable to someone else's number, was being compared. This goes away looking forward, unless others escalate the numbers game.

 

[Kenrou]

Mack, while I agree with the whole explanation, it still boils down to buyers getting even more confused than normal because the names at this point seem arbitrary, you might as well do away with the whole nomenclature and just stick a barcode to it. I mean, again comments nailed it, when you have laptops that might be called something like "Intel 4 i7 1385G7 on 7nm" or worse you will have store clerks getting even more chances of getting something wrong or using it on purpose to cheat customers of their $. While this is hilarious from our point of view that know something (or all) about it, it's also bad from the point of view of the "average Joe" that knows nothing about what they are purchasing, no ?

 

[mackerel]

Average Joe will not care at all or be impacted by this, as they wont care at all about process. They are not the main target, who are those somewhat interested in technical details but don't know as much as they need to in order to not get mislead by differences between Intel's competitors and Intel's old system. The new system better aligns it with competitors so numerical comparisons are not as bad as they are now.

 

[Kenrou]

[video=youtube;kSLaE_jTqJY]

 

[Voodoo Rufus]

On a positive note, in their roadmap they reference the "Angstrom era". If all these semiconductor geniuses can make it work, it's another major leap.

Just like when we went from microns to nanometers.

 

[EarthDog]

This was a good vid from Ian and Marco (about the new naming convention...as well as the Anand article). There's jumps to the spot in the video.

[video=youtube;6cNvkPigYic]

 

[David]

On a positive note, in their roadmap they reference the "Angstrom era". If all these semiconductor geniuses can make it work, it's another major leap.

Just like when we went from microns to nanometers.

The fun bit of this is that it's a fundamental limit ... distances between atoms are typically ~1.2 - 3.5 Angstroms ...

 

[mackerel]

I was looking up that the other day. The effective radius of hydrogen is 120pm and for silicon 210pm, so if we're talking spacing it'll be double that. In terms of practical feature size we're still in the ball park of two magnitudes away from that.

 

[Voodoo Rufus]

The fun bit of this is that it's a fundamental limit ... distances between atoms are typically ~1.2 - 3.5 Angstroms ...

Yeah, I realize we're quick approaching known limits. I wonder what they're working on next.

 

[David]

Yeah, I realize we're quick approaching known limits. I wonder what they're working on next.

I suspect after that we need to replace bits with qubits ... hopefully at temperatures higher than 4 K. Helium compressors are heavy.

 

[Kenrou]

[video=youtube;wxKGFxmwcDo]