Hacker News Reader

Comment by wahern | Hacker News Reader

Anybody expect 7nm processes to result in longevity issues? As far as I understand (and IME) the first components to fail are capacitors. Might that begin to change?

Apropos the article, I'm trying to convince myself to build an EPYC 3201 server now rather than waiting for the Zen 2 version, for which I presume I'd have to wait until October or November at the earliest.

View on HN

I think, electromigration will still kill the chip earlier than individual device failures.

It was Intel is said to switched to cobalt wiring in latest node, and seems to be paying dearly for that. TSMC and others seem to go the conventional road and continued to perfect the salicide for smaller nodes without any issues.

Officially, Intel says 10nm has lithography problems. They did try a more aggressive node than TSMC's first "7nm", entirely using 193nm UV, and were the only company to attempt Self-Aligned Quadruple Patterning (SAQP) for the top metal layers.

First generation anything is not as good as it's going to be.

Up until recently every CPU generation was on a brand new node - there was no second gen on the same process. We never had reliability issues with CPUs before.

The Intel Atom C2000 product family had a terrible problem which would brick systems, see for example https://www.tomshardware.com/news/intel-cpu-failure-atom-pro... or https://www.theregister.co.uk/2017/02/06/cisco_intel_decline...

There was also a weak transistor on some Sandy Bridge chips in the SATA controller. But these failures are really quite rare.

Reply on news.ycombinator.com