Open vlsi opened 2 years ago
I have not researched/seen existing footprint benchmarks, however, the following tests might make sense:
a) "footprint overhead for storing N entries in a Map" (N=1..100_000) b) "footprint overhead in case N insertions and N/2 removals are performed in a random order" c) "footprint overhead in case the implementation allows to store Key within Value itself" (see https://github.com/vlsi/compactmap/blob/f641491ca13705a0821f0af51b64cbc3914a13f9/jol/src/test/java/NoDefaultsTest.java#L94-L108 )
You can find some information on https://collection-libs-comparison.develar.org/object-to-object
I have not researched/seen existing footprint benchmarks, however, the following tests might make sense:
a) "footprint overhead for storing N entries in a Map" (N=1..100_000) b) "footprint overhead in case N insertions and N/2 removals are performed in a random order" c) "footprint overhead in case the implementation allows to store Key within Value itself" (see https://github.com/vlsi/compactmap/blob/f641491ca13705a0821f0af51b64cbc3914a13f9/jol/src/test/java/NoDefaultsTest.java#L94-L108 )