Evaluating the Performance of a D-Wave Quantum Annealing System for Feature Subset Selection in Software Defect Prediction