Saturday, February 9, 2019

Building Quantum Computers Is Hard

ExtremeTech reports:
You may have read that quantum computers one day could break most current cryptography systems. They will be able to do that because there are some very clever algorithms designed to run on quantum computers that can solve a hard math problem, which in turn can be used to factor very large numbers. One of the most famous is Shor’s Factoring Algorithm. The difficulty of factoring large numbers is essential to the security of all public-private key systems — which are the most commonly used today. Current quantum computers don’t have nearly enough qubits to attempt the task, but various experts predict they will within the next 3-8 years. That leads to some potentially dangerous situations, such as if only governments and the super-rich had access to the ultra-secure encryption provided by quantum computers.

Why Building Quantum Computers Is Hard

There are plenty of reasons quantum computers are taking a long time to develop. For starters, you need to find a way to isolate and control a physical object that implements a qubit. That also requires cooling it down to essentially zero (as in .015 degrees Kelvin, in the case of IBM‘s Quantum One). Even at such a low temperature, qubits are only stable (retaining coherence) for a very short time. That greatly limits the flexibility of programmers in how many operations they can perform before needing to read out a result.

Not only do programs need to be constrained, but they need to be run many times, as current qubit implementations have a high error rate. Additionally, entanglement isn’t easy to implement in hardware either. In many designs, only some of the qubits are entangled, so the compiler needs to be smart enough to swap bits around as needed to help simulate a system where all the bits can potentially be entangled.
So quantum computing is extraordinarily difficult, but they will break most of our computer security systems, and experts predict it within the next 3-8 years.

2 comments:

  1. Oh ye gods. Another 'all we need to do is... just send lots more money and treat computer programming like magic!' diatribe. Swap bits around? seriously? That's like making a guess about a guess about a calculation you can't actually do based on theory.

    Save some money. Go buy a magic eight ball, or a Ouija board instead, it will save a lot of money and give just as accurate answers.

    ReplyDelete
  2. It's like, here (http://www.imechanica.org/node/23078) I got a great post, and you don't even notice it, Roger!

    That's bad.

    Hmmm... Happens with a lot of Americans...

    Also with Indians...

    But where can I live my life, you know?

    Saturday 16 February 2019 02:15:26 AM IST

    --Ajit

    ReplyDelete