Create a benchmark to verify the correctness of Monte Carlo by comparing the End-To-End actual runtime (and cost or carbon) of applications to the Runtime Estimate of Monte Carlo.
For the MapReduce benchmark (Our only conditional benchmark), use a mix of small and large input sizes (50-50 for simplicity or some reasonable ratio), then compare the actual end-to-end runtime distribution to the Monte Carlo Distribution for runtime using our Solver.
Create a benchmark to verify the correctness of Monte Carlo by comparing the End-To-End actual runtime (and cost or carbon) of applications to the Runtime Estimate of Monte Carlo.
For the MapReduce benchmark (Our only conditional benchmark), use a mix of small and large input sizes (50-50 for simplicity or some reasonable ratio), then compare the actual end-to-end runtime distribution to the Monte Carlo Distribution for runtime using our Solver.