.svg)
Quantum Simulator Sickness
Simulator sickness is a known phenomena in virtual reality and first-person games. It is the sensation of discomfort caused by sensory conflict between what the brain thinks is happening and what the body feels is going on.
While quantum computing does not create sensory conflict, there is often a sense of discomfort when moving from a quantum simulator to actual quantum hardware.
Why is that the case?
Some might take the 'glass half-full" view and say that simulators are too perfect, while others might see the 'glass half-empty' and say that quantum computers are too imperfect. Simulators all too often fail to appropriately model imperfections in gates and qubits. They don't capture the variance between qubits on the same machine, or the variance from one machine to another.
As a result, quantum scientists that think they perfected an algorithm on a simulator are sometimes surprised to find that it does not give the desired results on actual hardware.
What to do?
One solution would be to make improvements. Improved simulators can do a better job in modeling noise and other imperfections. Improved quantum computers can reduce the noise and become more consistent.
Another solution is to allow deeper analysis of quantum circuits. This can be something such as counting the multi-qubit gates in a particular design, creating a 'heat map' to pinpoint busy qubits. Such tools are available today in some quantum development environments, such as Classiq's.
Yet another solution would be to work in such a way to allow testing the algorithm on machines from different vendors. Given that many quantum cloud providers carry more than one type of quantum computer, this might be possible. However, make sure that the development environment does not require you to painstakingly modify the circuit to fit the architecture, gate set and qubit count of another vendor. Environments that can do this automatically are very useful in this context.
Designing quantum algorithms is hard. There is saying at Classiq: "if it were easy, we'd have marketing do it". But thinking through the problem and choosing the right development environment can reduce or eliminate the discomfort when moving from simulators to real hardware.
Simulator sickness is a known phenomena in virtual reality and first-person games. It is the sensation of discomfort caused by sensory conflict between what the brain thinks is happening and what the body feels is going on.
While quantum computing does not create sensory conflict, there is often a sense of discomfort when moving from a quantum simulator to actual quantum hardware.
Why is that the case?
Some might take the 'glass half-full" view and say that simulators are too perfect, while others might see the 'glass half-empty' and say that quantum computers are too imperfect. Simulators all too often fail to appropriately model imperfections in gates and qubits. They don't capture the variance between qubits on the same machine, or the variance from one machine to another.
As a result, quantum scientists that think they perfected an algorithm on a simulator are sometimes surprised to find that it does not give the desired results on actual hardware.
What to do?
One solution would be to make improvements. Improved simulators can do a better job in modeling noise and other imperfections. Improved quantum computers can reduce the noise and become more consistent.
Another solution is to allow deeper analysis of quantum circuits. This can be something such as counting the multi-qubit gates in a particular design, creating a 'heat map' to pinpoint busy qubits. Such tools are available today in some quantum development environments, such as Classiq's.
Yet another solution would be to work in such a way to allow testing the algorithm on machines from different vendors. Given that many quantum cloud providers carry more than one type of quantum computer, this might be possible. However, make sure that the development environment does not require you to painstakingly modify the circuit to fit the architecture, gate set and qubit count of another vendor. Environments that can do this automatically are very useful in this context.
Designing quantum algorithms is hard. There is saying at Classiq: "if it were easy, we'd have marketing do it". But thinking through the problem and choosing the right development environment can reduce or eliminate the discomfort when moving from simulators to real hardware.
