Add gate error and balanced objective to BIP mapper #6778
Conversation
| @@ -84,9 +84,9 @@ def __init__( | |||
| If None, all qubits in the coupling_map will be considered. | |||
| objective (str): Type of objective function: | |||
|
|
|||
| * ``'error_rate'``: [NotImplemented] Predicted error rate of the circuit | |||
| * ``'error_rate'``: Predicted error rate of the circuit | |||
There was a problem hiding this comment.
can you add some description of how this "prediction" works? I think it multiplies gate errors, and thus it is only a function of circuit size but not depth. But just needs to be explicit.
There was a problem hiding this comment.
Add more description on 'error_rate' objective at 0c24c75
There was a problem hiding this comment.
I think error_rate is not a good name because a true error rate is a combination of gate errors and depth of circuit. This completely ignores depth. We should just call it size. So we have "size"/"depth"/"balanced"
There was a problem hiding this comment.
I agree "error_rate" is not a precise name and "size" is better, but "size" still sounds like a bit misleading name to me since the objective considers 2q-gate errors (not counts). How about "2qgate_error"? I think it's the most appropriate name. Is it too long?
There was a problem hiding this comment.
good point. If we want to be completely accurate in naming these objective function then even depth might not be too accurate. I'll merge this for now and if we find a better name we can replace.
There was a problem hiding this comment.
changed to "gate_error" for now.
mtreinish
added
the
Changelog: New Feature
itoko
changed the title
Summary
Add
gate_errorandbalancedobjectives to the BIP mapper.Details and comments
Previously, only the
depthobjective is implemented inBIPMappingpass. This commit adds the implement ofgate_errorandbalancedobjective functions, which are described in the original paper (https://arxiv.org/abs/2106.06446). Note thatgate_erroris referred as "error_rate" in the paper.