Improve jackson benchmark in java#303
Conversation
There was a problem hiding this comment.
The order of running programs also affects the performance results in JVM. Simply running System.gc() as an warmup is not sufficient.
There was a problem hiding this comment.
The benchmark has 4 times warmup runs for each one and it avoids the affect of running order, I think. BTW, how does your benchmark avoid the order of running programs?
There was a problem hiding this comment.
For example, if you have two programs P1, P2, it runs the program P1, P2, P1, P2,... to reduce the order effect.
There was a problem hiding this comment.
Okay. The original benchmark code (https://github.com/msgpack/msgpack-java/blob/v07-develop/msgpack-jackson/src/test/java/org/msgpack/jackson/dataformat/benchmark/MessagePackDataformatPojoBenchmarkTest.java#L98-L130) does the same thing, right?
|
@xerial Can you review it again? Thanks |
|
ok. It looks good. |
…-java Improve jackson benchmark in java
2 participants
No description provided.