4 ms·
I'm just going to link the comment I wrote above :) https://news.ycombinator.com/item?id=36867709 https://news.ycombinator.com/item?id=36867709 Needless to sa
by MayeulC 3y ago
I'm just going to link the comment I wrote above :)
https://news.ycombinator.com/item?id=36867709 https://news.ycombinator.com/item?id=36867709
Needless to say, this would be game-changing. But extraordinary claims require extraordinary evidence etc, let's be cautiously optimistic here.
- pcdoodle 3y agoThank bro! are we talking like no heat loss in the traces on the motherboard type a deal or also on a wafer level internal to the CPU?
- MayeulC 3y agoProbably both? At least the first. For integrated circuits, I think lead is a hard pass (for now), and the deposition process needs to be worked out. Ideally, this could be useful for the hottest paths: clock tree, high-speed buses, as well as the power supplies. There are a few hurdles though: high-speed voltage changes create changing currents, which creates variable magnetic fields, which IIRC may be a problem depending on the superconductor's characteristics. Processors also work at low voltages, which means that they need huge currents. Both magnetic fields and large current (as well as high temperatures) can break down superconductivity. So it's challenging, but probably doable. There are also superconducting structures that could replace transistors, see applications of https://en.wikipedia.org/wiki/Josephson_effect https://en.wikipedia.org/wiki/Josephson_effect
- pcdoodle 3y agoAppreciate the follow up. This could be really cool!