Ask HN: Room temperature superconductor: what to expect?

Re:Ask HN: Room temperature superconductor: what to expect? by JohnBooty on Sunday July 30, 2023 @11:24AM (#36932155)

    CPUs etc - highly unlikely, we are talking dozens-atoms-level 
    stuff, going to be really hard to replicate superconductor 
    properties AND the usual semiconductor stuff

Possibly moronic question:

Maybe a RTSC based chip doesnt need to compete with modern chips in terms of process size or complexity

What if you could just make an RTSC based chip with complexity and process size comparable to e.g. a 1993 era 80386 but it runs at 500ghz or something

(You'd need memory, or at least a cache, that could keep up I guess)

[ Parent | on HN ]

Re:Ask HN: Room temperature superconductor: what to expect? by PaulHoule on Sunday July 30, 2023 @11:41AM (#36932376)

Already it ought to be possible to clock an indium phosphide cpu to 50 GHz or beyond and there was an SBIR awarded to make one back in the 1990s. Memory is a problem though, particular the whole system has to be small since it already takes a nanosecond for light to travel one meter.

Note silicon (probably any semiconductor) is about to hit the wall in terms of feature size. An obvious answer is to stack logic elements in top of each other but it is already difficult to extract heat from 2-d silicon chips. If superconducting gates really use 1000 times less power that stacking might be possible.

[ Parent | on HN ]

Re:Ask HN: Room temperature superconductor: what to expect? by out_of_protocol on Sunday July 30, 2023 @01:24PM (#36933589)

"superconducting" obviously different from "semiconductor". Fat chance combining these. Maybe in the future, that's going to be separate instant Nobel prize, on top of the one we're talking about now

[ Parent | on HN ]

Re:Ask HN: Room temperature superconductor: what to expect? by mikewarot on Sunday July 30, 2023 @03:10PM (#36934818)

They combine those in quantum computers all the time. You'd be able to do it without liquid helium if this works out.

Personally, I want to own a SQUID[1] or two for some experiments, which is otherwise FAR outside my budget. They've been made with HTS, I see no reason why they couldn't be made with this.

[1] https://en.wikipedia.org/wiki/SQUID

[ Parent | on HN ]

Re:Ask HN: Room temperature superconductor: what to expect? by out_of_protocol on Sunday July 30, 2023 @06:32PM (#36936677)

In quantum computers we are talking double digits cubits. In regular computing, we do store and process billions snd billions bits. Regular 1TB microSD is 10^13 bits or so, regular CPU contains a dozen or two of billions of transistors. So, clearly, it doesn't work same way

[ Parent | on HN ]

Re:Ask HN: Room temperature superconductor: what to expect? by mikewarot on Sunday July 30, 2023 @07:16PM (#36937015)

Take a look at the DWave computers, they use the same processes for silicon and superconductors in one big wafer.

The chemistry and processes can be made to be compatible.

So, in theory you could assemble a device with multiple SQUIDs on it, sampling the A field[1] in 3 dimensions, instead of the B field (the Curl of the A field) that is usually measured, just for one instance.

[1] https://en.wikipedia.org/wiki/Magnetic_vector_potential

[ Parent | on HN ]

Re:Ask HN: Room temperature superconductor: what to expect? by out_of_protocol on Monday July 31, 2023 @03:37AM (#36939739)

measurement parts and computation parts can be on the same wafer but they are separate beings. in regular CPU you don't need measurement part, mostly. Transistors work precisely because they are made of semiconductors - due to some quantum mechanisms like https://en.wikipedia.org/wiki/Band_gap And transistors is where you loose most of energy. You can make capacitors out of superconducting material, and it could work out somehow for memory elements (at this moment these are made from transistors as well)

[ Parent | on HN ]