One common thing we have in some benchmarks is sweeping on a parameter, like number of qubits. It would be great if there was a way a user could specify as a runtime flag values (or a sweep either linear or log) to use as a value for these parameters. Something like:
would run that benchmark with a linear sweep from 10 to 100 with 50 elements on the num_qubits parameter (obviously the syntax is just an example, and doesn't necessarily have to be like that). This would make it a lot easier to quickly benchmark things over a wide range and also do arbitrary scale testing with red queen.
One common thing we have in some benchmarks is sweeping on a parameter, like number of qubits. It would be great if there was a way a user could specify as a runtime flag values (or a sweep either linear or log) to use as a value for these parameters. Something like:
would run that benchmark with a linear sweep from 10 to 100 with 50 elements on the num_qubits parameter (obviously the syntax is just an example, and doesn't necessarily have to be like that). This would make it a lot easier to quickly benchmark things over a wide range and also do arbitrary scale testing with red queen.