QStabilizer has recently been refactored to introspect and throw on cases of gates that it cannot apply. As a result, some QStabilizerHybrid implementation is redundant.
I notice that this PR itself adds a redundant check for separability on one control qubit in cases of non-Clifford MCMtrx(), and I notice that test_qft_cosmology benchmarks have recently slightly degraded, but any regression in performance in my tests doesn't seem to be a matter of this PR.
QStabilizerhas recently been refactored to introspect and throw on cases of gates that it cannot apply. As a result, someQStabilizerHybridimplementation is redundant.I notice that this PR itself adds a redundant check for separability on one control qubit in cases of non-Clifford
MCMtrx(), and I notice thattest_qft_cosmologybenchmarks have recently slightly degraded, but any regression in performance in my tests doesn't seem to be a matter of this PR.