Widespread inconsistencies and flaws in benchmarking practices for graph processing systems lead to misleading and non-reproducible results.
