2.x: Benchmark X-Map-Z operators

[akarnokd]

This PR adds JMH benchmarks to measure operators mapping onto other types, such as flatMapCompletable, concatMapSingle, etc. In addition, two new benchmarks were added to measure the flatMapIterable performance in additional situations.

Baseline results

i7 4770K, Windows 7 x64, Java 8u162

JMH Compare GUI workspace file: xmapz_ws.xml

The first diagram compares the dedicated, plain (same inner type as the outer type) and conversion-based flows:

The switchMap-based Observable operators look like they could use some optimizations. Overall, count == 1 is not optimized with the dedicated versions.

From the Observable's perspective:

Looks like the concatMapX operators could use some optimizations.

The flatMapIterable measures are as follows:

The Flowable version seem to be considerably slower, probably not explainable due to backpressure overhead.

[codecov]

Codecov Report

❗ No coverage uploaded for pull request base (2.x@90bca55). Click here to learn what that means.
The diff coverage is n/a.

@@          Coverage Diff           @@
##             2.x    #5914   +/-   ##
======================================
  Coverage       ?   98.19%           
  Complexity     ?     5996           
======================================
  Files          ?      655           
  Lines          ?    43927           
  Branches       ?     6084           
======================================
  Hits           ?    43134           
  Misses         ?      242           
  Partials       ?      551

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 90bca55...2cc8e7c. Read the comment docs.