In this thesis we look into the time complexity and performance of the model-log conformance calculation implemented in ProM, which is a means of evaluating model accuracy with regards to a log, through finding the most model-log synchronizing (optimal) alignment. We specifically focus on a particular part of the algorithm, which calculates integer linear pro- grams repeatedly in three distinct phases, and we identify per phase what optimizations can be made. Overall a significant increase of about 50% in performance is obtained, and we present large scale tests on random (real-life case sized) data to validate the improvement. All evaluated techniques, even those unsuccessful, are explained and evaluated in detail. The successful techniques can serve as a basis for future improvements.