[WIP] Add JMH benchmarks for conjure map deserialization

[aldexis]

Before this PR

We don't have a great way to measure conjure serde performance.

After this PR

==COMMIT_MSG==
Add JMH benchmarks for conjure map deserialization
==COMMIT_MSG==

This measures

• The time it takes to deserialize
• The memory that gets allocated for deserialization
• The memory that is retained after by the deserialized object

Running this locally (and removing pieces that aren't that useful), I get the below results. Summarized:

• In the one-item map case, the custom singleton map handling allocates a bit more memory (~1% more), but retains a lot less (~47%)
• For empty maps, the custom handling allocates a bit less and retains a ton less (just the object wrapper)
• For all cases, the guava option allocates ~20% less and retains ~10% less memory

Benchmark                                                             (json)        (mapImpl)  Mode  Cnt        Score         Error   Units
ConjureBenchmarks.testAllocatingBenchmark                             NO_MAP           NORMAL  avgt    5      184.651 ±      29.672   ns/op
ConjureBenchmarks.testAllocatingBenchmark:gc.alloc.rate.norm          NO_MAP           NORMAL  avgt    5      968.004 ±       0.001    B/op
ConjureBenchmarks.testAllocatingBenchmark:mem.retained.total.norm     NO_MAP           NORMAL  avgt    5      112.092 ±       0.068    B/op

ConjureBenchmarks.testAllocatingBenchmark                             NO_MAP        SINGLETON  avgt    5      159.875 ±       7.069   ns/op
ConjureBenchmarks.testAllocatingBenchmark:gc.alloc.rate.norm          NO_MAP        SINGLETON  avgt    5      936.003 ±       0.001    B/op
ConjureBenchmarks.testAllocatingBenchmark:mem.retained.total.norm     NO_MAP        SINGLETON  avgt    5       16.027 ±       0.101    B/op

ConjureBenchmarks.testAllocatingBenchmark                             NO_MAP  GUAVA_IMMUTABLE  avgt    5      176.506 ±     183.164   ns/op
ConjureBenchmarks.testAllocatingBenchmark:gc.alloc.rate.norm          NO_MAP  GUAVA_IMMUTABLE  avgt    5      872.003 ±       0.004    B/op
ConjureBenchmarks.testAllocatingBenchmark:mem.retained.total.norm     NO_MAP  GUAVA_IMMUTABLE  avgt    5       16.032 ±       0.045    B/op

ConjureBenchmarks.testAllocatingBenchmark                              EMPTY           NORMAL  avgt    5      257.319 ±      56.192   ns/op
ConjureBenchmarks.testAllocatingBenchmark:gc.alloc.rate.norm           EMPTY           NORMAL  avgt    5     1208.005 ±       0.001    B/op
ConjureBenchmarks.testAllocatingBenchmark:mem.retained.total.norm      EMPTY           NORMAL  avgt    5      112.128 ±       0.310    B/op

ConjureBenchmarks.testAllocatingBenchmark                              EMPTY        SINGLETON  avgt    5      220.516 ±       5.144   ns/op
ConjureBenchmarks.testAllocatingBenchmark:gc.alloc.rate.norm           EMPTY        SINGLETON  avgt    5     1176.004 ±       0.001    B/op
ConjureBenchmarks.testAllocatingBenchmark:mem.retained.total.norm      EMPTY        SINGLETON  avgt    5       16.016 ±       0.106    B/op

ConjureBenchmarks.testAllocatingBenchmark                              EMPTY  GUAVA_IMMUTABLE  avgt    5      211.399 ±       8.693   ns/op
ConjureBenchmarks.testAllocatingBenchmark:gc.alloc.rate.norm           EMPTY  GUAVA_IMMUTABLE  avgt    5      984.004 ±       0.001    B/op
ConjureBenchmarks.testAllocatingBenchmark:mem.retained.total.norm      EMPTY  GUAVA_IMMUTABLE  avgt    5       16.047 ±       0.079    B/op

ConjureBenchmarks.testAllocatingBenchmark                          SINGLETON           NORMAL  avgt    5      405.794 ±      26.292   ns/op
ConjureBenchmarks.testAllocatingBenchmark:gc.alloc.rate.norm       SINGLETON           NORMAL  avgt    5     1528.008 ±       0.001    B/op
ConjureBenchmarks.testAllocatingBenchmark:mem.retained.total.norm  SINGLETON           NORMAL  avgt    5      272.118 ±       0.205    B/op

ConjureBenchmarks.testAllocatingBenchmark                          SINGLETON        SINGLETON  avgt    5      357.805 ±      36.661   ns/op
ConjureBenchmarks.testAllocatingBenchmark:gc.alloc.rate.norm       SINGLETON        SINGLETON  avgt    5     1552.007 ±       0.001    B/op
ConjureBenchmarks.testAllocatingBenchmark:mem.retained.total.norm  SINGLETON        SINGLETON  avgt    5      144.092 ±       0.159    B/op

ConjureBenchmarks.testAllocatingBenchmark                          SINGLETON  GUAVA_IMMUTABLE  avgt    5      320.432 ±      33.864   ns/op
ConjureBenchmarks.testAllocatingBenchmark:gc.alloc.rate.norm       SINGLETON  GUAVA_IMMUTABLE  avgt    5     1200.006 ±       0.001    B/op
ConjureBenchmarks.testAllocatingBenchmark:mem.retained.total.norm  SINGLETON  GUAVA_IMMUTABLE  avgt    5      160.142 ±       0.289    B/op

ConjureBenchmarks.testAllocatingBenchmark                              THREE           NORMAL  avgt    5      574.595 ±      18.514   ns/op
ConjureBenchmarks.testAllocatingBenchmark:gc.alloc.rate.norm           THREE           NORMAL  avgt    5     1920.011 ±       0.001    B/op
ConjureBenchmarks.testAllocatingBenchmark:mem.retained.total.norm      THREE           NORMAL  avgt    5      552.286 ±       0.216    B/op

ConjureBenchmarks.testAllocatingBenchmark                              THREE        SINGLETON  avgt    5      580.841 ±      51.899   ns/op
ConjureBenchmarks.testAllocatingBenchmark:gc.alloc.rate.norm           THREE        SINGLETON  avgt    5     1920.011 ±       0.001    B/op
ConjureBenchmarks.testAllocatingBenchmark:mem.retained.total.norm      THREE        SINGLETON  avgt    5      552.261 ±       0.402    B/op

ConjureBenchmarks.testAllocatingBenchmark                              THREE  GUAVA_IMMUTABLE  avgt    5      506.948 ±      58.892   ns/op
ConjureBenchmarks.testAllocatingBenchmark:gc.alloc.rate.norm           THREE  GUAVA_IMMUTABLE  avgt    5     1528.010 ±       0.001    B/op
ConjureBenchmarks.testAllocatingBenchmark:mem.retained.total.norm      THREE  GUAVA_IMMUTABLE  avgt    5      480.155 ±       0.354    B/op

ConjureBenchmarks.testAllocatingBenchmark                                SIX           NORMAL  avgt    5      863.696 ±      62.455   ns/op
ConjureBenchmarks.testAllocatingBenchmark:gc.alloc.rate.norm             SIX           NORMAL  avgt    5     2512.017 ±       0.001    B/op
ConjureBenchmarks.testAllocatingBenchmark:mem.retained.total.norm        SIX           NORMAL  avgt    5      976.401 ±       1.085    B/op

ConjureBenchmarks.testAllocatingBenchmark                                SIX        SINGLETON  avgt    5      889.678 ±     158.380   ns/op
ConjureBenchmarks.testAllocatingBenchmark:gc.alloc.rate.norm             SIX        SINGLETON  avgt    5     2512.017 ±       0.003    B/op
ConjureBenchmarks.testAllocatingBenchmark:mem.retained.total.norm        SIX        SINGLETON  avgt    5      976.380 ±       1.032    B/op

ConjureBenchmarks.testAllocatingBenchmark                                SIX  GUAVA_IMMUTABLE  avgt    5      758.321 ±      39.006   ns/op
ConjureBenchmarks.testAllocatingBenchmark:gc.alloc.rate.norm             SIX  GUAVA_IMMUTABLE  avgt    5     2008.015 ±       0.001    B/op
ConjureBenchmarks.testAllocatingBenchmark:mem.retained.total.norm        SIX  GUAVA_IMMUTABLE  avgt    5      864.340 ±       0.478    B/op

ConjureBenchmarks.testAllocatingBenchmark                                TEN           NORMAL  avgt    5     1278.512 ±      61.835   ns/op
ConjureBenchmarks.testAllocatingBenchmark:gc.alloc.rate.norm             TEN           NORMAL  avgt    5     3312.025 ±       0.001    B/op
ConjureBenchmarks.testAllocatingBenchmark:mem.retained.total.norm        TEN           NORMAL  avgt    5     1553.018 ±       1.519    B/op

ConjureBenchmarks.testAllocatingBenchmark                                TEN        SINGLETON  avgt    5     1280.454 ±     147.395   ns/op
ConjureBenchmarks.testAllocatingBenchmark:gc.alloc.rate.norm             TEN        SINGLETON  avgt    5     3312.025 ±       0.001    B/op
ConjureBenchmarks.testAllocatingBenchmark:mem.retained.total.norm        TEN        SINGLETON  avgt    5     1552.555 ±       0.988    B/op

ConjureBenchmarks.testAllocatingBenchmark                                TEN  GUAVA_IMMUTABLE  avgt    5     1128.815 ±     149.054   ns/op
ConjureBenchmarks.testAllocatingBenchmark:gc.alloc.rate.norm             TEN  GUAVA_IMMUTABLE  avgt    5     2712.022 ±       0.003    B/op
ConjureBenchmarks.testAllocatingBenchmark:mem.retained.total.norm        TEN  GUAVA_IMMUTABLE  avgt    5     1392.462 ±       1.501    B/op

ConjureBenchmarks.testAllocatingBenchmark                            HUNDRED           NORMAL  avgt    5    11248.318 ±     640.032   ns/op
ConjureBenchmarks.testAllocatingBenchmark:gc.alloc.rate.norm         HUNDRED           NORMAL  avgt    5    23720.220 ±       0.011    B/op
ConjureBenchmarks.testAllocatingBenchmark:mem.retained.total.norm    HUNDRED           NORMAL  avgt    5    14757.744 ±       5.804    B/op

ConjureBenchmarks.testAllocatingBenchmark                            HUNDRED        SINGLETON  avgt    5    11391.878 ±     553.685   ns/op
ConjureBenchmarks.testAllocatingBenchmark:gc.alloc.rate.norm         HUNDRED        SINGLETON  avgt    5    23720.222 ±       0.012    B/op
ConjureBenchmarks.testAllocatingBenchmark:mem.retained.total.norm    HUNDRED        SINGLETON  avgt    5    14758.101 ±       6.503    B/op

ConjureBenchmarks.testAllocatingBenchmark                            HUNDRED  GUAVA_IMMUTABLE  avgt    5     9613.995 ±     650.940   ns/op
ConjureBenchmarks.testAllocatingBenchmark:gc.alloc.rate.norm         HUNDRED  GUAVA_IMMUTABLE  avgt    5    18616.185 ±       0.016    B/op
ConjureBenchmarks.testAllocatingBenchmark:mem.retained.total.norm    HUNDRED  GUAVA_IMMUTABLE  avgt    5    13003.866 ±       9.984    B/op

ConjureBenchmarks.testAllocatingBenchmark                           THOUSAND           NORMAL  avgt    5   112590.970 ±   13575.673   ns/op
ConjureBenchmarks.testAllocatingBenchmark:gc.alloc.rate.norm        THOUSAND           NORMAL  avgt    5   210554.177 ±       0.343    B/op
ConjureBenchmarks.testAllocatingBenchmark:mem.retained.total.norm   THOUSAND           NORMAL  avgt    5   144405.510 ±      84.187    B/op

ConjureBenchmarks.testAllocatingBenchmark                           THOUSAND        SINGLETON  avgt    5   111950.084 ±    7945.716   ns/op
ConjureBenchmarks.testAllocatingBenchmark:gc.alloc.rate.norm        THOUSAND        SINGLETON  avgt    5   210554.703 ±       4.775    B/op
ConjureBenchmarks.testAllocatingBenchmark:mem.retained.total.norm   THOUSAND        SINGLETON  avgt    5   144400.054 ±      63.475    B/op

ConjureBenchmarks.testAllocatingBenchmark                           THOUSAND  GUAVA_IMMUTABLE  avgt    5    91829.734 ±    8998.321   ns/op
ConjureBenchmarks.testAllocatingBenchmark:gc.alloc.rate.norm        THOUSAND  GUAVA_IMMUTABLE  avgt    5   163961.788 ±       0.186    B/op
ConjureBenchmarks.testAllocatingBenchmark:mem.retained.total.norm   THOUSAND  GUAVA_IMMUTABLE  avgt    5   128243.556 ±      64.769    B/op

[changelog-app]

Generate changelog in changelog/@unreleased

Type (Select exactly one)

• Feature (Adding new functionality)
• Improvement (Improving existing functionality)
• Fix (Fixing an issue with existing functionality)
• Break (Creating a new major version by breaking public APIs)
• Deprecation (Removing functionality in a non-breaking way)
• Migration (Automatically moving data/functionality to a new system)

Description

Add JMH benchmarks for conjure map deserialization

Check the box to generate changelog(s)

• Generate changelog entry

[aldexis]

Not sure why, but I can't seem to run in IntelliJ with runtime = 25