[REVIEW][Java] Bindings, tests and benchmarks for RMM pooled memory

[ldematte]

Exposes cuvsRMMPoolMemoryResourceEnable/cuvsRMMMemoryResourceReset via the Java API, with tests and benchmarks

[copy-pr-bot[bot]]

This pull request requires additional validation before any workflows can run on NVIDIA's runners.

Pull request vetters can view their responsibilities here.

Contributors can view more details about this message here.

[mythrocks]

I've merged #1510, which seems to have introduced conflicts in the CagraBuildAndSearchIT.java file.

[ldematte]

Thanks @mythrocks, I'll rebase

[ldematte]

@mythrocks I have rebased this onto 25.12 too

[mythrocks]

I've taken @rapidsai/cuvs-python-codeowners, @rapidsai/ci-codeowners, etc. off of this review. The changes are all in Java, and straightforward.

[mythrocks]

/ok to test d74f38e

[mythrocks]

/ok to test bd8670a

[mythrocks]

Hmm. Odd C++ compile error:

/home/coder/cuvs/cpp/build/pip/cuda-13.0/release/_deps/raft-src/cpp/include/raft/cluster/kmeans.cuh(97): error: SamplingOp is not a template
                        SamplingOp<DataT, IndexT>& select_op,

Included from kmeans_common.cuh, and it's certainly a class-template.

[mythrocks]

/ok to test bd8670a

[mythrocks]

/ok to test bd8670a

[mythrocks]

/ok to test fc7b55c

[mythrocks]

/ok to test ed92aa9

[mythrocks]

/ok to test 6900dfe

[mythrocks]

/merge