3 ms·

consider that a big part of why we cant make better processors is because they get hot, and they get hot because of resistance. I'd say dramatically faster chip

consider that a big part of why we cant make better processors is because they get hot, and they get hot because of resistance. I'd say dramatically faster chips in smaller form factors (think enormous graphics card in your phone, most of the bulk in those is cooling) would be quite a revolution !

ben_w 3y ago
Depends on the details, as always. I don't know the superconductivity mechanism here (does anyone?), but IIRC Cooper pairs in BCS theory can be separated by much larger gaps than the transistors in a modern CPU — hundreds, rather than single-digits, of nanometers. Might make fully-3D processors much more viable though, from lack of heat dissipation; and if it does, that in turn might be able to make up for a coarser resolution.
Kirby64 3y ago
Heat is a big factor, but part of limitations on speed really come from physical dimensions. At some point, speed of light travel becomes a limitation. Also, the transistors themselves generate a lot of the heat... So you still have large amounts of heat to handle even if the interconnect traces are heatless.
- jacquesm 3y ago
  The transistors create a lot of heat because the current carrying capacity of the interconnect is severely limited which increases the time to switch the transistor from off to on or vv. This puts the transistor into a resistive domain where it consumes power and that leads to heat. If you use a superconducting interconnect that transition time will drop sharply and that in turn will lead to less power used by the transistor.
- RugnirViking 3y ago
  > Heat is a big factor, but part of limitations on speed really come from physical dimensions. undoubtably. We can't make things infinately fast. But being able to put the most powerful desktop processors and graphics cards we have today into a phone would definately be useful, not to mention laptops. It would also dramatically change how server farms work (their existance right now revolves around cooling, they could be so much more dense)

ben_w 3y ago

Depends on the details, as always. I don't know the superconductivity mechanism here (does anyone?), but IIRC Cooper pairs in BCS theory can be separated by much larger gaps than the transistors in a modern CPU — hundreds, rather than single-digits, of nanometers. Might make fully-3D processors much more viable though, from lack of heat dissipation; and if it does, that in turn might be able to make up for a coarser resolution.

Kirby64 3y ago

Heat is a big factor, but part of limitations on speed really come from physical dimensions. At some point, speed of light travel becomes a limitation. Also, the transistors themselves generate a lot of the heat... So you still have large amounts of heat to handle even if the interconnect traces are heatless.

jacquesm 3y ago
The transistors create a lot of heat because the current carrying capacity of the interconnect is severely limited which increases the time to switch the transistor from off to on or vv. This puts the transistor into a resistive domain where it consumes power and that leads to heat. If you use a superconducting interconnect that transition time will drop sharply and that in turn will lead to less power used by the transistor.
RugnirViking 3y ago
> Heat is a big factor, but part of limitations on speed really come from physical dimensions. undoubtably. We can't make things infinately fast. But being able to put the most powerful desktop processors and graphics cards we have today into a phone would definately be useful, not to mention laptops. It would also dramatically change how server farms work (their existance right now revolves around cooling, they could be so much more dense)