Hey @johnaohara, I was wondering whether keeping the creation of ObjectMapper et al. within the benchmark methods doesn't potentially distort the results. Typically, you'd set up the mapper object only once and re-use them then. I.e. extracting them to the JMH state class would be more representative potentially. WDYT?
And would be nice to add Qson too?
Hey @johnaohara, I was wondering whether keeping the creation of
ObjectMapperet al. within the benchmark methods doesn't potentially distort the results. Typically, you'd set up the mapper object only once and re-use them then. I.e. extracting them to the JMH state class would be more representative potentially. WDYT?And would be nice to add Qson too?