Quote from: Bau on January 31, 2023, 20:19:44Intel is in late with implementing new processes that make use of Extreme UV lithography, but in reality they are not so behind in terms of density respect to TSMC and Samsung.

Density is one thing, gate leakage is another. It is obvious while Samsung and Intel can fit as many transistors into the same silicon footprint, their gates are leaking much more currents causing a lot of waste and heat. This is most evident on ARM chips because Samsung and MediaTek use the same ARM reference design, so any difference in efficiency is the result of the fab process. While here we can only guess how much of the deficiency is due to poor architectural design and how much is due to fabrication.

AMD instantly wins in terms of noise and battery life as soon as the load drops below average.

Intel are just cheaters - they overclocked their SoC at the cost of the wildest TDP. If you force TDP to be lowered to the AMD level, they immediately lose more than 1.5 times in speed and at least 30% in battery life, which is proven in reviews and even officially from Lenovo in psref on the same Intel / AMD models.

At some point, the authorities will force both to stop the TDP race, and they have already reached the physical limits of growth, so they will have to compete further on the same and much lower TDP - for example, only 65W for the entire laptop consumption, including the processor and discrete chip. And who will win here?

Bau nailed it. Way back in the "35nm" era, the "nm" rating stopped meaning "average distance between transistors". It's just a label.

Franky, I'm surprised Intel didn't jump to a "2nm" label to sound better than TSMC, milking public ignornance like they did back in the "GHz myth" days. Those were the days when flagship AMD CPUs briefly passed up Intel in performance, while running at significantly lower clocks.

Please, stop with the comparison between "10nm" and "7-5-4-3 nm". It is almost 10 years that such "nm" does not indicate an actual dimension. Intel 10nm (now Intel 7) is just the name of a proccess such as TSMC "7 nm" or Samsung "8 nm". They could call such process "Jack", "Amelia", "Horse" or whatever, is just a name. They choose to use "x nm" just because it is easier for a consumer understand that is new. Unfurtunatly, it is evident that nobody is really able to understand that.

Intel is in late with implementing new processes that make use of Extreme UV lithography, but in reality they are not so behind in terms of density respect to TSMC and Samsung.

I don't see any difference with 1260p at all. Everything again rests on PL1 / PL2, and the noise of the laptop is directly related to them, and the more compact the model, the worse it gets in terms of noise.

Intel continues to sell us obsolete 10nm+++++++, instead of AMD's already "4nm" in the 7x40 series with Phoenix.

One problem - there are NO 15.6"+ laptops on the market even with 6680/6850U in mass quantities, because AMD is not able to mass-produce cheap processors - it has become a completely paper company. Yes, and Intel, if you look - you can find tasty models at 1240p (1250 ) and on 1260(1270) in 15.6+ is practically unrealistic on the market in any country, including the USA.

So both companies can continue to tell us tales about their new series, which will actually hit the market en masse at best by mid-2024...

Amen.