search

qiskit-community / qiskit-aqua

Quantum Algorithms & Applications (**DEPRECATED** since April 2021 - see readme for more info)
https://qiskit.org/aqua
Apache License 2.0
574 stars 376 forks source link

Additional timing information beyond `eval_time` when running variational algorithms available in aqua #309

Closed ajavadia closed 5 years ago

ajavadia commented 5 years ago

Original issue by @JuanIgnacioAdame in https://github.com/Qiskit/qiskit-terra/issues/1706

What is the expected enhancement?

My understanding is that when running the QAOA implementation in aqua, the only timing information you can retrieve is the total wall-clock time, which is returned as the value of the key eval_time in the results dictionary. My understanding is that eval_time consists of the sum of the varying time in the queue, the classical optimization process involving updating the gammas and betas, the actual time it took to run the quantum circuits, etc. It would be great if a breakdown of eval_time was available. This would allow one to look into what parts of the overall application/algorithm it would be most worthwhile to look into to try to speed up.

woodsp-ibm commented 5 years ago

Looking into parts of the application/algorithm to alter/optimize to speed things up overall is more the job of a profiler. A search for Python Profilers will turn up helpful information on what they can do, how to use etc if you are not familiar.

woodsp-ibm commented 5 years ago

Closing as this level of timing is the purview of a Python Profiler