Intel 10nm, TSMC 7nm / 5nm are marketing terms that cannot be compared. You can say Intel 7nm will have X% higher density than Intel 10nm. Or TSMC 5nm will be Y% dense than TSMC 7nm. Right now, TSMC 7nm is less dense than Intel 10nm. Secondly transistor density itself cannot be considered a valid parameter, there are differences between different types of FinFETs and the sizes differ based on whether the transistor is part of the CPU core or cache or some other logic and also depends on the operating frequencies. Moreover transistors are not distributed at the same density across the whole piece of silicon, even though its possible to do that if you are just creating something as simple as LLC. Most high performance logic circuits will have blank regions and barriers to make the CPU handle higher clock frequencies. Please dont mislead people with vague misinformation.

Quote from: rs on June 08, 2020, 00:42:01

Quote from: S.Yu on June 07, 2020, 23:45:18
TSMC 5nm is twice the density as Intel 10nm+? Do you have sources for that?

According to the public numbers it seems more like 70+%. Intel's 10nm density is ~100 MTr per mm², while TSMC's 5nm density is reported to be >170 MTr per mm².

Very interesting...

Quote from: S.Yu on June 07, 2020, 23:45:18
TSMC 5nm is twice the density as Intel 10nm+? Do you have sources for that?

According to the public numbers it seems more like 70+%. Intel's 10nm density is ~100 MTr per mm², while TSMC's 5nm density is reported to be >170 MTr per mm².

TSMC 5nm is twice the density as Intel 10nm+? Do you have sources for that?

After the ryzen in xps article and this article , i feel like notebookcheck losing the professionalism until the point i would totally stop taking reference from this site already .

This is a quote from an Anandtech article by Dr Cutress from AMD's FAD 2020.  Article is March 5, 2020.


"Zen 4 based Genoa has already been announced as the CPU to power the El Capitan supercomputer, and in this roadmap AMD has put it as coming out by 2022. We asked AMD for clarification, and they stated that in this sort of graph, we should interpret it as the full stack of Genoa should be formally launched by the end of 2022."

what is the source of the claim that AMD will have a 2021 server with pcie5 and ddr5?  Neither of these were shown on their roadmaps presented at FAD 2020.

"AMD's 2021 server offering will be built on the new SP5 platform and also offer support for PCIe 5.0 and DDR5 memory. "

A leaked Intel Russia roadmap presentation offered a glimpse at Intel's enterprise server plans for 2020.....
You must mean 2021, right?

Notebookcheck, please don't turn into fudzilla with these kind of articles. You are not professional at all.
As for sapphire rapids...yeah, it'll probably still be worse than amd offerings. Intel needs a bit of time to recover and I think that will happen by 2022-2023, when ocean cove based products launch on 7nm.

Many times the writers of notebookcheck compare directly the technology node and pass comments such as the one in this article - tsmc 5nm could be twice as fast as intel 10nm. This is not entirely true and ultimately it depends on the transistor density and other factors. Please check this before you pass the message to the readers.