Lists serialization benchmarks

[rutkowskij]

Observed Performance Degradation During Deserialization of Objects with Large Collections

I have observed a significant performance degradation when deserializing objects that contain large collections (lists). In my case, the deserialization time increased by nearly 3x 🧐
Interestingly, when deserializing previously created generic objects, the performance drop is not as noticeable. The most significant difference compared to the previous version is for objects serialized with version 2.1.1 and deserialized with version 2.1.1.
I added benchmarks on my local branch to measure the performance in version 1.10.1 and compared it to the current implementation (entries with name Avro4kV1*). I created a very simple model that contained just an Int list, and the performance drop was already noticeable.
I have added the results to the README for comparison. The remaining results (for v2) are comparable to those in the README on my machine, so I think I can include the results for reference.
Interestingly, there was a strange and intriguing result for JacksonAvro serialization (the fastests one). I’m not sure how to interpret it at this moment, but this effect does not seem to occur with more complex models than just an Int list.

[Chuckame]

Here are the results on my computer using the new kotlinx-io decoder, doubling the performances!! Still not better than jackson, but it's close!

Suggested change

	c.g.a.b.lists.Avro4kListsBenchmark.read thrpt 5 412.503 ± 6.127 ops/s -13.16%
	c.g.a.b.lists.JacksonAvroListsBenchmark.read thrpt 5 1002.527 ± 23.039 ops/s +111.09%
	c.g.a.b.lists.Avro4kListsBenchmark.read thrpt 5 813.448 ± 19.265 ops/s
	c.g.a.b.lists.JacksonAvroListsBenchmark.read thrpt 5 972.243 ± 59.303 ops/s