Migrate from RAFT to CUVS

[tarang-jain]

Remove the dependency on raft::compiled and modify GPU implementations to use cuVS backend in place of RAFT.

A deeper insight into the dependency:
FAISS gets the ANN algorithm implementations such as IVF-Flat and IVF-PQ from cuVS. RAFT is meant to be a lightweight C++ header-only template library that cuVS relies on for the more fundamental / low-level utilities. Some examples of these are RAFT's device mdarray and mdspan objects; the RAFT resource object (raft::resource) that takes care of the stream ordering of device functions; linear algebra functions such as mapping, reduction, BLAS routines etc. A lot of the cuVS functions take the RAFT mdspan objects as arguments (for example raft::device_matrix_view). Therefore FAISS relies on both cuVS and RAFT. FAISS gets RAFT headers through cuVS and uses them to create the function arguments that can be consumed by cuVS. Note that we are not explicitly linking FAISS against raft::raft or raft::compiled. Only the required headers are included and compiled rather than compiling the whole RAFT shared library. This is the reason we still see mentions of raft in FAISS.

[tarang-jain]

@asadoughi can the system glibc version be updated for running the cuvs github action? The cmake seg fault went away when I downgrade to cmake 3.26.4, but I see this error https://github.com/facebookresearch/faiss/actions/runs/11239339471/job/31246908895?pr=3549 which might mean a libc version mismatch.

[asadoughi]

AFAICT, this CI is running on ubuntu-latest so you can try amending the shell commands in build_cmake/action.yaml to run ldd --version and the relevant sudo apt-get -y install command to install the specific glibc version.

[facebook-github-bot]

@asadoughi has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator.

[facebook-github-bot]

@asadoughi has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator.

[asadoughi]

Thanks! Looks like unpinned CMake with 24.08 works.

In addition to the torch CPU tensor issue, let's track

• migration to libcuvs 24.10, pending your discussion with @cjnolet
  • including cuVS build time 1+ hr more than other CI jobs
• cuVS in FAISS documentation - already filed Add CuVS (aka RAFT) documentation #3823