Micro-architectures
You have probably already seen that although the actual speedup for many quantum algorithms lies in the “quantum” nature of the qubits, many of these algorithms still require a fair amount of classical information processing for it to work. It is, therefore, crucial that the channels containing classical and quantum information can work together seamlessly.
This is where the micro-architecture comes into play. In this first video, QuTech researcher Nader Khammasi will introduce the arbiter, a component that determines whether to execute a given instruction on a classical or quantum processor. Doing this spares the quantum computer from performing tasks that can be carried out more reliably and efficiently on classical hardware.
Prerequisite knowledge
- The compiler of a quantum computer
- Quantum programming languages
- Quantum speedup algorithms (specifically Shor’s algorithm)
- Quantum measurement
- Coherence time of qubits
Okay, so the arbiter decides which piece of information goes to the classical channel, and which to the quantum channel. Let’s focus on the quantum channel for now. After the quantum processor has performed its quantum algorithm, we should also think about how to acquire the information from the qubits. In other words, we need to think about measuring qubits. Professor at QuTech Koen Bertels will talk about this process in the context of the micro-architecture in the next video. At the end of the video, Prof. Bertels explains why the field of Quantum Error Correction (QEC) is such an essential part of the subject we are discussing here.
Further thinking
In the second video, Professor Bertels emphasizes that we should repeat our quantum algorithm many times since we rely on quantum measurements. Why is that? Suppose we increase our number of measurements by a factor of 2, can we “trust” our outcome twice as good as well? Why (not)?
Further reading
Check this article about microarchitecture for superconducting quantum processors.