A couple of benchmarks in tree traversal scenarios comparing different stream libraries (such as Java Stream, Jool, Vavr, StreamEx, and others) in the JVM.

Run with, for example:

./gradlew clean jmhJar
java -jar ./treebench/build/libs/treebench-jmh.jar -i 4 -wi 4 -f 1 -r 2 -w 2

According to our experiments, the above arguments provide consistent observations. You may vary these parameters accordingly.

• -i 4 → --iterations - Number of measurement iterations JMH will perform after warmup.
• -wi 4 → --warmupIterations - Number of warmup iterations before actual measurement.
• -f 1 → --forks - Number of forked JVMs in which the benchmark will run. Each fork is a fresh JVM process.
• -r 2 → --time - Time per iteration. Default unit is seconds unless a suffix (ms, s, m) is provided. Here: 2 seconds per iteration.
• -w 2 → --warmupTime Controls the duration of warmup time per iteration.