Should we increase `BufferedChecksum`'s buffer from 1 KB -> 2 KB?

Spinoff from [this comment](https://github.com/apache/lucene/issues/15552#issuecomment-3720969176) from @rmuir. Modern SIMD registers have gotten wider, and maybe that means it's worth using a larger `byte[]` when computing Lucene's CRC32 checkums?

Start by locating BufferedChecksum and reviewing how its byte[] buffer is used when computing Lucene CRC32 checksums. Benchmark the 1 KB and 2 KB alternatives under representative workloads, then document the measured trade-off and update the buffer only if the results support the change.

Written by the indexing model from the issue text.