This PR adds another benchmark that stresses the transaction mechanism. The benchmark is based on a naïve leftist heap data structure. The problem with that is that every operation modifies the root. That is a fundamental sequential bottleneck and it means that the data structure simply cannot scale to multiple cores such that thruput would improve by adding workers. Nevertheless this could be a useful benchmark of the transaction mechanism.
This PR adds another benchmark that stresses the transaction mechanism. The benchmark is based on a naïve leftist heap data structure. The problem with that is that every operation modifies the root. That is a fundamental sequential bottleneck and it means that the data structure simply cannot scale to multiple cores such that thruput would improve by adding workers. Nevertheless this could be a useful benchmark of the transaction mechanism.