(for different libs there should be different classes, for more details look in the code):
net.datafaker.benchmark.java_object_population.single_object
Project | Java Version | Mode | Cnt | Score | Units |
---|---|---|---|---|---|
Datafaker | JDK 17.0.5, OpenJDK 64-Bit Server VM, 17.0.5+8 | avgt | 10 | 925.698 ± 131.938 | ms/op |
Kotlin Faker | JDK 17.0.5, OpenJDK 64-Bit Server VM, 17.0.5+8 | avgt | 10 | 3481.494 ± 232.195 | ms/op |
Instancio | JDK 17.0.5, OpenJDK 64-Bit Server VM, 17.0.5+8 | avgt | 10 | 26881.156 ± 730.465 | ms/op |
ATTENTION! In this test we measure not throughput but the average time of one operation => lower is better.
(for different libs there should be different classes, for more details look in the code):
net.datafaker.benchmark.java_object_population.sequence
Project | Java Version | Mode | Cnt | Score | Units |
---|---|---|---|---|---|
Datafaker | JDK 17.0.5, OpenJDK 64-Bit Server VM, 17.0.5+8 | avgt | 10 | 1024.540 ± 121.902 | ms/op |
Instancio | JDK 17.0.5, OpenJDK 64-Bit Server VM, 17.0.5+8 | avgt | 10 | 1349.173 ± 75.254 | ms/op |
ATTENTION! In this test we measure not throughput but the average time of one operation => lower is better.