Quote from: Warren on May 06, 2022, 00:54:58
Why is this article using the word "proprietary" to describe Intel process nodes? Only reason I can see is to subtle introduce bias, which is not necessary in a news report. After all, TSMC's N4 is proprietary too...

I suspect the use of the word "proprietary" refers to the fact that the design details of intel's 7nm process are not publicly available and thus only intel can make designs for its process.  As a foundy, TSMC publishes Process Design Kits (PDKs) which are widely available, so that many companies (eg. Apple, Qualcomm, AMD, intel, etc...) can design products to use their process technology.

Quote from: Warren on May 06, 2022, 00:54:58
Why is this article using the word "proprietary" to describe Intel process nodes? Only reason I can see is to subtle introduce bias, which is not necessary in a news report. After all, TSMC's N4 is proprietary too...

Awkward word choice, but not necessarily inaccurate.

On a related note, when will we see some AMD chips made at Samsung? They need moar capacity.

Why deadname Intel 4 by putting "7nm" in parenthesis. These's nothing inherently "7nm" to base the clarification off of

Most likely this is a sign that they simply aren't hitting their power efficiency targets - plenty of signs on the current mixed Intel/TSMC chips - and that they think shouldering out NVIDIA/AMD/Quallcomm/Samsung will be a better move in the short term than trying to improve their fabs.

I can't agree. I know they're in a tight spot, but this strategy is a hybrid strategy, and that never succeeds. They aren't committed to one direction (vertical integration) or the other (AMD/NVIDIA/Apple style). And how generously will TSMC look on a rival trying to extinguish TSMC's customer base? They'll be burning cash and still not have enough volume, I'd wager.

AMD focused on investing in share buy backs critical time when they should have invested heavily in TSMC supply and ensured they locked Intel out of TSMC. Its like they didnt believe in thier own success, unless they never got an option.

I honestly don't think this is a good idea. Intel sells a lot more chips than AMD and even AMD cant get enough chips out of TSMC to meet their demands.

Even if they are a little bit back compared to their competitors, Intel node production has been solid once they get over their initial problems. 10nm when from being useless to good enough to really good in just 1 year and a half.

I wasn't aware Intel had a 7nm process to abandon.

Why is this article using the word "proprietary" to describe Intel process nodes? Only reason I can see is to subtle introduce bias, which is not necessary in a news report. After all, TSMC's N4 is proprietary too...

Fake news.. already powered on .. only option is to dual source if they need more supply