Renaissance 0.16 Released

We are pleased to announce a new release of the Renaissance benchmark suite. This release adds result validation to several benchmarks, improves compatibility with current Java releases, and updates numerous underlying frameworks and libraries.

Validation was improved or added to the following 7 workloads: als, fj-kmeans, future-genetic, gauss-mix, log-regression, movie-lens, and philosophers, bringing Renaissance closer to having fully validated workloads (there are still 4 missing).

Issues related to the Hadoop Client API library (which plagued the Apache Spark benchmarks) and the Chronicle Map library (which plagued the db-shootout benchmark) were resolved. As a result, all benchmarks in this release (with the exception of neo4j-analytics which requires at least version 17 to run), run on LTS versions from 11 to 21 as well as on early access version of Java 24.

The philosophers and scala-stm-bench7 benchmarks were migrated to Scala 3, continuing the migration towards more recent Scala versions. The reactors benchmark is the last to require Scala 2.12.

Even though the source code of most benchmarks remains unchanged, the actual code executed at runtime may be affected by the dependency updates, resulting in different performance. Apart from the indirect impact of library updates, the performance of the future-genetic, gauss-mix, movie-lens, and philosophers benchmarks is known to differ from the previous release due to benchmark changes.

See the GitHub release for more details.

Any comments and contributions are welcome.

Happy benchmarking!