Benchmarks are based on Oracle Java Microbenchmark Harness.

Please be cautious with conclusions based on microbenchmarking as there are plenty possible pitfalls for goals, test compositions, input data, an envionment and the analyze itself.

Execute all benchmark methods with 4 worker threads:

mvn clean install
java -jar target/benchmarks.jar  ".*" -t 4

or specify a filter for benchmark methods and the number of forks and warmup/measurements iterations, e.g.:

java -jar target/benchmarks.jar  -t 4  -f 3 -i 10 -wi 5  ".*CounterBenchmark.*"
java -jar target/benchmarks.jar  -t 4  -f 3 -i 10 -wi 5  ".*ReservoirBenchmark.*"
java -jar target/benchmarks.jar  -t 4  -f 3 -i 10 -wi 5  ".*MeterBenchmark.*"

The whole list of command line options is available by:

java -jar target/benchmarks.jar -h