Trickle down effect! This is what happens when Taiwan produces a massive brain trust from their Universities and we don't. The result is skilled groups of engineers, able to advance the company's and country's progress.

Just say how many transistors fit in 1 happy meal and there you have your industry standard!

schmidtbag:

That's funny, but, I get why Intel is doing this. Even here, there are people who think Intel's 10nm is worse than TSCM's 7nm. Sites that compare CPU stats aren't doing Intel any favors either. Intel has already been facing image problems for almost 4 years, and with the release of the 11900K, it hasn't got better. Using their competitor's way of measuring transistors is one way to seem like they're actually making competitive progress, especially when you consider how much people make such a fuss about it.

Completely agree. Ever since they all moved from 2D transistors to 2.5D with FinFET the nm size has been purely marketing anyhow. So why not market your node the same way as Samsung and TSMC?

Intel: we want to remove the stigma of nanometer shaming because we're still using fat chips. Rest of the world: F*** off, Intel. Stop being lazy and lose the weight on your chips, you're lousy competition.

Neo Cyrus:

I wouldn't put it past them, but in this extremely rare case Intel aren't lying through their teeth, and it is true. Their 10nm is sort of comparable to TSMC's 7nm. Edit - Typo.

That's quite statement to make. Each process has wide range of available densities. And intel's process can overlap with TSMC's on density scale. But when both use exactly same density. (Same count of transistors per mm^2.) What's operational voltage of each process to reach 3GHz, 4GHz? What's power draw per FLOP? Because 14nm at highest available density will overlap with 10nm density range. Those processes are sold as improvements of not only density, but achievable clock (at reasonable voltage), and energy efficiency.

And what happens in 2023 when they fully release their 7nm chips (yeah.. sure...) while other fabs name their own nodes at 4nm? Rename again? Make a standard or release products that stand for themselves regardless of process names. Anything else is just pitiful...