Semiwiki on LinkedIn Semiwiki on Facebook Semiwiki on Twitter Register / Log in 
Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/threads/intel-18a-dimensions-leaked.22352/
        )

    [addOns] => Array
        (
            [DL6/MLTP] => 13
            [Hampel/TimeZoneDebug] => 1000070
            [SV/ChangePostDate] => 2010200
            [SemiWiki/Newsletter] => 1000010
            [SemiWiki/WPMenu] => 1000010
            [SemiWiki/XPressExtend] => 1000010
            [ThemeHouse/XLink] => 1000970
            [ThemeHouse/XPress] => 1010570
            [XF] => 2021770
            [XFI] => 1050270
        )

    [wordpress] => /var/www/html
)

Intel 18A Dimensions Leaked?

I get this from WiKi, seems Intel 18nm process is much larger than TSMC 2nm?

1742807922104.png
 
Last edited:
Stan Chen said:
I get this from WiKi, seems Intel 18nm process is much larger than TSMC 2nm?

View attachment 2911
Click to expand...
Those numbers are all from Techinsights, and they have N3 as ~20% more dense than 18A (283MTr/mm2), and then N2 as ~10% more dense than N3.
The 18A figure Techinsights has looks like it is way higher than what one might get if they used Mark Bohr's formula, and the same follows for N3. Perhaps they are using an UHD 18A lib, and single fin N3, because even the 2+1 N3 lib doesn't achieve those numbers the techinsight report claims (or perhaps they aren't using that formula at all).
From this set of leaked numbers though, it would appear 18A HD is roughly on par with N3 2-2 density, and N3 is still marginally ahead (15-20%) with their finflex 2-1 density.
 
Geddagod said:
Those numbers are all from Techinsights, and they have N3 as ~20% more dense than 18A (283MTr/mm2), and then N2 as ~10% more dense than N3.
The 18A figure Techinsights has looks like it is way higher than what one might get if they used Mark Bohr's formula, and the same follows for N3. Perhaps they are using an UHD 18A lib, and single fin N3, because even the 2+1 N3 lib doesn't achieve those numbers the techinsight report claims (or perhaps they aren't using that formula at all).
From this set of leaked numbers though, it would appear 18A HD is roughly on par with N3 2-2 density, and N3 is still marginally ahead (15-20%) with their finflex 2-1 density.
Click to expand...
Mark's Bohr figure is based on actual pitches I am pretty sure Techinsights also uses this but their density calculation in this one is based on the public density gains which was not true as iirc N5 was shown to be closer to 171Mxtor/mm2 but it was actually 135Mxtor/mm2.

I guess this stems from that as for what lib are those pitches is not clear.

www.angstronomics.com

The TRUTH of TSMC 5nm

Claimed Density vs Reality
www.angstronomics.com www.angstronomics.com
 
siliconbruh999 said:
Mark's Bohr figure is based on actual pitches I am pretty sure Techinsights also uses this but their density calculation in this one is based on the public density gains which was not true as iirc N5 was shown to be closer to 171Mxtor/mm2 but it was actually 135Mxtor/mm2.

I guess this stems from that as for what lib are those pitches is not clear.

www.angstronomics.com

The TRUTH of TSMC 5nm

Claimed Density vs Reality
www.angstronomics.com www.angstronomics.com
Click to expand...
Using Mark Bohr's equation wouldn't get you 238MTr/mm2 for 18A though, afaik, it gets you ~185MTr/mm2, at least using these rumored specs.
Which casts doubts on the accuracy of these rumored pitches....
However considering techinsights appears to have gotten wrong numbers for TSMC N3 as well, it may just be that they are the ones wrong. Or they are using single fin stuff.
For Techinsight potentially just using the public density gain figure, idk, I would think that after they realized that the one for N5 was wrong, they wouldn't repeat the same thing for N3. Especially since those figures came from 2024. (Actually, their numbers for N3E is 273, not 283, my bad).
The funny thing is though, using either the rumored specs or tech insights, you would get the same story- TSMC N3 is marginally ahead of 18A in HD (or in TSMC's case, finflex HD) density.
 
Last edited:
Scotten Jones said:
To the best of my knowledge Intel has only disclosed 18A pitches under NDA, I have not seen a public disclosure and I talk to Intel all the time.
Click to expand...
Is there any ETA when the NDA will be lifted? i.e. the data made public

I doubt the leak even more now and the author deleted his Post as well.
 
Last edited:
You must log in or register to reply here.
Back
Top