Java Micro-Benchmarks With JMH

These benchmarks are build with OpenJDK JMH.

Just another collection of typical Java Code Benchmarks.

How To Run The Benchmarks

Build the project mvn clean install.

You can call java -jar target/benchmarks.jar -help to see available command line parameters.

For instance,

  java -Xmx1G -jar targget/benchmarks.jar -f 1 -r 5 -i -wi 5 EmptyBenchmark

Create Your Own Benchmark

mvn archetype:generate           
       -DinteractiveMode=false           
       -DarchetypeGroupId=org.openjdk.jmh          
       -DarchetypeArtifactId=jmh-java-benchmark-archetype
       -DarchetypeVersion=1.21           
       -DgroupId=org.sample           
       -DartifactId=test           
       -Dversion=1.0-SNAPSHOT

Benchmark Results

REMEMBER: The numbers below are just data. To gain reusable insights, you need to follow up on why the numbers are the way they are. Use profilers (see -prof, -lprof), design factorial experiments, perform baseline and negative tests that provide experimental control, make sure the benchmarking environment is safe on JVM/OS/HW level, ask for reviews from the domain experts. Do not assume the numbers tell you what you want them to tell.

Use Multiply Instead Of Divide

Check MultiplyVsDivisionBenchmark to see that it is faster to multiply than to divide doubles.

    double divisor = 10.0;
    double a = a / divisor // repeat this 1_000 times
    ...

better use

    double divisor = 10.0;
    double factor = 1 / divisor;
    double a = nextDouble() * factor; // repeat this 1_000 times

Results with JDK 13:

java -Xmx1G -jar target/benchmarks.jar -f 1 -r 1 -i 5 -w 1 -wi 5 MultiplyVsDivisionBenchmark

Benchmark	Mode	Cnt	Score	Error	Units
MultiplyVsDivisionBenchmark.divideDoubles	thrpt	5	207619,369	± 2662,392	ops/s
MultiplyVsDivisionBenchmark.divideDoublesWithMultiply	thrpt	5	508495,675	± 8033,898	ops/s

Random Subset Without Duplicates

RandomSubsetBenchmark

Results with JDK 13:

java -Xmx1G -jar target/benchmarks.jar -f 1 -r 1 -i 5 -w 1 -wi 5 RandomSubsetBenchmark

Benchmark	Mode	Cnt	Score	Error	Units
RandomSubsetBenchmark.subsetWithForLoop	thrpt	5	1720,292	± 330,771	ops/s
RandomSubsetBenchmark.subsetWithParallelStreams	thrpt	5	779,079	± 23,238	ops/s
RandomSubsetBenchmark.subsetWithStreams	thrpt	5	1200,939	± 140,346	ops/s

Date Parsing And Adding An Hour Benchmark

DateBenchmark

java -Xmx1G -jar target/benchmarks.jar -f 1 -r 1 -i 5 -w 1 -wi 5 DateBenchmark

Benchmark	Mode	Cnt	Score	Error	Units
DateBenchmark.dateLong	thrpt	5	386571214,471	± 10171784,541	ops/s
DateBenchmark.dateLongWithFormat	thrpt	5	3358368,673	± 31265,425	ops/s
DateBenchmark.dateParsingWithFormat	thrpt	5	728327,799	± 19937,432	ops/s
DateBenchmark.parsingLocalDateTime	thrpt	5	873308,951	± 97291,855	ops/s

Using Initial Capacity At ArrayList

Run with JDK 12 and java -jar target/benchmarks.jar -f 1 -r 2 -w 2 -wi 5 InitialCapacityBenchmark

Benchmark	Mode	Cnt	Score	Error	Units
InitialCapacityBenchmark.addWithInitialCapacity	thrpt	5	29,552	± 13,596	ops/s
InitialCapacityBenchmark.addWithInitialCapacityWithStreams	thrpt	5	25,123	± 10,103	ops/s
InitialCapacityBenchmark.addWithoutInitialCapacity	thrpt	5	19,304	± 3,389	ops/s
InitialCapacityBenchmark.addWithoutInitialCapacityWithStreams	thrpt	5	19,012	± 6,849	ops/s

Created By

CROWDCODE GmbH & Co. KG