Benchmarks measure overhead of acquiring a connection or setting up EntityManager

[lukaseder]

I came here from a reddit discussion:
https://www.reddit.com/r/java/comments/o2z0yu/hibernate_community_performance_comparison_with/h2aqdjx?utm_source=share&utm_medium=web2x&context=3

I think these resource management lines should be moved out of the benchmark into the setup section:

querydsl-entityql/src/test/java/pl/exsio/querydsl/entityql/java/benchmark/execution/jdbc/JDBCQueryExecutionBenchmark.java

Line 39 in fb61553

Connection c = settings.dataSource.getConnection();

querydsl-entityql/src/test/java/pl/exsio/querydsl/entityql/java/benchmark/execution/jpa/JpaJPQLQueryExecutionBenchmark.java

Line 34 in fb61553

EntityManager em = settings.emf.createEntityManager();

Or, at least, create a separate benchmark that allows for seeing how much overhead is produced by this resource management.

I don't know your product, but I'm assuming this overhead is already being moved to the setup section when you measure your own product:
https://github.com/eXsio/querydsl-entityql/blob/master/src/test/java/pl/exsio/querydsl/entityql/java/benchmark/execution/entityql/EntityQLDynamicQueryExecutionBenchmark.java

[eXsio]

Hi Lukas!

Thanks for noticing this, you are absolutely correct. I will update the benchmarks and post the new results.