Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/index.php?whats-new/profile-posts/2942167/
        )

    [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] => 2021370
            [XFI] => 1050270
        )

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

New profile posts

In that wolf story the boy was eventually, tragically, correct. Will that be the story of High NA? Well, we have a couple of years to see it play out.

Of course, I'm more interested in EBL, so I'm just an interested bystander on HNA. If it goes sour, better for me
Dylan's latest article analyzing hybrid bonding is the best SemiAnalysis yet.

I like Dylan, and I like Fred. Both are educational.

Fred is often "the boy who cried wolf". If you believed everything he has written, EUV would never have worked - but I have been using a phone built with EUV for over 2 years. You need to give credit to the engineers who use those things and who really make them effective.
N
nghanayem
I don't wholly disagree about Fred. As a memory guy he obviously has a different perspective that is partially validated by micron leading w/SAPQ. As for Dylan, I've seen far to much pulling things out of his rear and "trust me bro" to highly regard him. If it is correct it was publicly available, but the second he is making inferences or leaks from his "sources" I don't think I have seen a single accurate statement.
AI might recover a logical design from geometry. Lower layers should identify cells without ambiguity. Macro blocks should pop out due to repetition. Add human - AI interaction like "what drives inputs A on block 53" to cut through the jumble of masks.

Probably less successful with inventive analog solutions, still likely to need a real human in that loop.
T
Tanj
I have no data on use of mask services. But I think that is a clever strategy, and for DUV masks the products should be equivalent. The main downside I see is getting an matching good schedule from two shops.
T
Tanj
Would separating vias from metals make the masks inscrutable? Especially if you add some extra features on metal which go nowhere.
C
cliff
We make the layout at the same time, just send met4 and below to one mask shop, and above met4 to another.
  • Like
Reactions: Tanj
Hi Kevin

You mentioned on one post that you have a couple of smart sons doing programming in CA. Do they want to work with my company making compute-in-memory? I have a lot of USC and NCSU guys rejoining us in January to work on the OpenTitan project.
C
cliff
We have Calibre DRC perfectly integrated in case the user insists on saving a nanometer. We have no interest in being foundry qualified. Creation is easier. Imec, etc will run the DRCs and give our customers the Calibre error file if there is a problem. It is a good idea to run a 3rd party checker anyway.
C
cliff
K
KevinK
Like most startups, the VC owned most of Jasper - My wife was just a well-compensated CEO that has a reasonable amount of equity. Agree that 402 characters is a pain. I'm kevin@kranen.com.
Thanks for the response to my last post.

Is a request for an NDA for detailed design documents as outrageous as Cliff mentions? Or is Cliff being Cliff?
I am still finding my sea legs here. Thx
N
nghanayem
I think Cliff is being Cliff. In my opinion there is nothing wrong with you asking for an NDA.
Saw your post - yes a simulation power estimate is best.
I continue to struggle to keep up with things - I have a working method that requires the use of semiconductor designs - and yet zero experience in semiconductor / engineering.
I am very lucky the design is super simple.
C
cliff
I am putting together the WARP-V this summer anyway. It really costs me nothing to do this.
C
Chris9594
Excellent - I would like to do an NDA and then I can share the Architecture Specifications - Make sure you are sitting down with you read.

It is professionally written.
Can you send a link to one or more of your IEEE publications?
C
Chris9594
There are four - and focused on consumer implementations - and each is relatively similar. Applications move to Data Center and Semiconductor.
Interested in reviewing your network paper. What kind of NDA would you require? I am not building a network myself, but I occasionally consult for people who do.
C
Chris9594
Just a general NDA so that there are no issues in the patent filing process. The paper doesnt get too much in depth and once the details are filed into USPTO then the it will establish prior art. Looking to register non-public status for the next month or so.
C
Chris9594
If you can send me an email at chris@crossport.ai - I can send you a standard one.
I am working on raising some money later in the year - it is all bootstrapped - so disregard the mess in websites lol.
Intel’s prolonged underperformance at a time of economic upheaval will push it further behind rivals like AMG, TSMC, and Nvidia. Stalled momentum on its various pivots and chip factory expansion plans eventually might require Intel to sell off assets or consider risky partnerships to withstand continued decline.
VLSI Design | Maven-Silicon.com

VLSI Design or VLSI Chip Design is a process of integrating hundreds of thousands of transistors onto a singular silicon semiconductor microchip. Get VLSI Design training From Industry Experts of Maven Silicon.
Back
Top