What is the significance of different transpiling time, validating time, in queue time, running time for the same circuit run on different quantum systems for same no. of shots in IBM Q experience?
What does longer transpiling time mean?
Quantum Computing Stack Exchange is a question and answer site for engineers, scientists, programmers, and computing professionals interested in quantum computing. It only takes a minute to sign up.Sign up to join this community
Transpiling time is the time it takes for your circuit to be translated into a circuit that can be run on the backend of your choosing. This process includes converting gates into the standard basis gates ['cx', 'u1', 'u2', 'u3', 'id'], optimizing the circuit so it is shorter, mapping virtual qubits in the circuit to physical qubits, etc. Converting gates into basis gates is mostly independent of the backend you choose (unless it has a different basis set), so the time it takes depends on the number and complexity of the gates in the circuit.
The optimization is dependent on the backend, as mapping virtual qubits to physical qubits, condensing gates, etc. depends on the geometry of the qubits for that backend, specifically which qubits are entangled and which aren't. The default optimization level is 1, which includes mapping qubits and collapsing adjacent gates, but if you specify a higher level (the max is 3) like so
job = (circuit, backend, optimization_level=3)
then transpilation will take longer. More information on transpilation can be found here: https://qiskit.org/documentation/tutorials/circuits_advanced/4_transpiler_passes_and_passmanager.html
Validating time is the time it takes to verify that the circuit is able to be run, so I think the time between different backends is arbitrary.
Queue time for different backends is more dependent on how busy the queue is, and how much other people use the queue. A more detailed explanation of the queue algorithm can be found here: https://quantum-computing.ibm.com/docs/manage/backends/queue/
Running time is the time it takes the backend to run the circuit itself. With more complex circuits, the run time will increase. For backends with higher quantum volume, the same circuit will run faster than on other backends with lower quantum volume.
Transpilation (in this context) is backend-specific. Longer transpiling time just means the software took longer to make your circuits compatible and more optimized for the backend you selected (e.g. because your circuits require a lot of remapping).
In general, transpiling time is important to those who do research on that topic. Much like classical compilers, the speed of the transpilation is also important in addition to the quality. But researchers typically do their own transpilation instead of letting IBM Q Experience do it (in which case tranpiling time is displayed as N/A).
Validating time is mostly just tourist information. While it's also backend-specific, the time difference is usually very small.
Queuing time depends on how busy the backend is as well as your priority. It gives some users satisfaction that their universities are in the Q network, which allows them higher priorities.
Running time can differ based on backend configuration. Newer backends may have new technologies that allow circuits to run faster. Also the transpiled circuits are likely of difference sizes for different backends.