CHANGELOG.md

raft 26.04.00 (8 Apr 2026)

🚨 Breaking Changes

• Rename lanczos by @aamijar in rapidsai#2936
• Remove deprecated headers by @aamijar in rapidsai#2939

🐛 Bug Fixes

• Disable lanczos SM tests by @aamijar in rapidsai#2941
• Add binary_mult and const fixes for ratio and sqrt apis by @aamijar in rapidsai#2940
• Temporarily increase max_days_without_success by @aamijar in rapidsai#2950
• Remove "requires" keyword from raft::linalg::map by @tarang-jain in rapidsai#2949
• Add nvjitlink to cuda-toolkit pip extras for cusparse compatibility by @bdice in rapidsai#2948
• Fix cub missing header by @aamijar in rapidsai#2957
• Fix cub header by @aamijar in rapidsai#2959
• [REVIEW] Fix l_offset overflow in select_k kernels for 32-bit IdxT by @yan-zaretskiy in rapidsai#2974
• Use the new reference-based RMM API to set RMM pools in device_resources_snmg by @viclafargue in rapidsai#2972

📖 Documentation

• Raft docs cleanup by @aamijar in rapidsai#2935

🚀 New Features

• [FEA] Support raft::KeyValuePair Inputs in linalg::map by @tarang-jain in rapidsai#2913
• Backport memory resources PRs by @achirkin in rapidsai#2983
• Move PCA and TSVD from cuml to raft by @aamijar in rapidsai#2952

🛠️ Improvements

• Replace thrust::tuple with cuda::std::tuple by @miscco in rapidsai#2928
• Use verify-hardcoded-version pre-commit hook by @KyleFromNVIDIA in rapidsai#2931
• expand CI-skipping logic, other small build changes by @jameslamb in rapidsai#2933
• remove pip.conf migration code in CI scripts, update CI-skipping rules by @jameslamb in rapidsai#2938
• Drop Python 3.10 support by @gforsyth in rapidsai#2932
• Don't hard-code branch names and UCXX version by @KyleFromNVIDIA in rapidsai#2934
• refactor: build wheels and conda packages using Python limited API by @gforsyth in rapidsai#2943
• remove gitutils by @jameslamb in rapidsai#2946
• Use GHA id-token for sccache-dist auth token by @trxcllnt in rapidsai#2947
• Use Specific CCCL Includes by @divyegala in rapidsai#2953
• Drop uses of thrust/functional.h by @miscco in rapidsai#2954
• check-nightly-ci: update to new version by @jameslamb in rapidsai#2956
• Drop thrust::constant_iterator in favor of cuda::constant_iterator by @miscco in rapidsai#2955
• Drop thrust::counting_iterator in favor of cuda::counting_iterator by @miscco in rapidsai#2958
• check-nightly-ci: remove testing config by @jameslamb in rapidsai#2960
• Revert max_days_without_success by @aamijar in rapidsai#2962
• refactor(limited api): add explicit wheel.py-api to pyproject.toml by @gforsyth in rapidsai#2967
• Add support for Python 3.14 by @gforsyth in rapidsai#2966
• Drop uses of thrust::discard_iterator in favor of cuda::discard_iterator by @miscco in rapidsai#2963
• Update Cython lower bound pin to 3.2.2 by @vyasr in rapidsai#2969
• Remove pytest upper bound pin by @vyasr in rapidsai#2970
• add no_libucx matrix option in dependencies.yaml by @trxcllnt in rapidsai#2975
• fix verify-hardcoded-versions issues by @jameslamb in rapidsai#2980
• build wheels with CUDA 13.0.x, test wheels against mix of CTK versions by @jameslamb in rapidsai#2971

New Contributors

• @yan-zaretskiy made their first contribution in rapidsai#2974

Full Changelog: https://github.com/rapidsai/raft/compare/v26.04.00a...release/26.04

raft 26.02.00 (4 Feb 2026)

🚨 Breaking Changes

• Use CCCL's mdspan implementation by @bdice in rapidsai#2836
• Default to static linking of libcudart by @bdice in rapidsai#2890
• Remove neighbors/, cluster/, distance/, spatial/, sparse/neighbors/ apis by @aamijar in rapidsai#2885
• Remove cutlass and cuco dependencies by @divyegala in rapidsai#2916

🐛 Bug Fixes

• Include <thrust/for_each.h> where it is used by @bdice in rapidsai#2883
• Include CTest module in CMakeLists.txt by @bdice in rapidsai#2895
• Fix Lanczos Determinism by @aamijar in rapidsai#2894
• Change compile-time assertion to runtime assertion on is_strided by @bdice in rapidsai#2909
• Set memory pool through RMM by @viclafargue in rapidsai#2866

📖 Documentation

• New readme image by @aamijar in rapidsai#2907
• Readme improvements by @aamijar in rapidsai#2906

🚀 New Features

• Tile Policy for Uint8 Input (Pairwise) by @tarang-jain in rapidsai#2770
• Add copy_vectorized to RAFT by @lowener in rapidsai#2900

🛠️ Improvements

• Use strict priority in CI conda tests by @bdice in rapidsai#2879
• Use strict priority in CI conda tests by @bdice in rapidsai#2884
• Remove alpha specs from non-RAPIDS dependencies by @bdice in rapidsai#2886
• Enable merge barriers by @KyleFromNVIDIA in rapidsai#2889
• Fix is_exhaustive, no longer constexpr by @bdice in rapidsai#2888
• Add devcontainer fallback for C++ test location by @bdice in rapidsai#2893
• eigsh optional seed by @aamijar in rapidsai#2899
• Empty commit to trigger a build by @bdice in rapidsai#2904
• Update to C++20 by @divyegala in rapidsai#2908
• Use SPDX license identifiers in pyproject.toml, bump build dependency floors by @jameslamb in rapidsai#2910
• Remove neighbors/detail/faiss_select by @aamijar in rapidsai#2902
• Remove sparse/distance by @aamijar in rapidsai#2905
• Add CUDA 13.1 support by @bdice in rapidsai#2896
• Fix CCCL 3.2 mdspan constexpr issues by @bdice in rapidsai#2911
• build and test against CUDA 13.1.0 by @jameslamb in rapidsai#2912
• Laplacian Kernel for COO inputs by @aamijar in rapidsai#2891
• Empty commit to trigger a build by @jameslamb in rapidsai#2919
• Use main shared-workflows branch by @jameslamb in rapidsai#2921
• Fix update-version.sh incorrectly replacing main() function names by @AyodeAwe in rapidsai#2923
• Lanczos remove dead code by @aamijar in rapidsai#2918
• wheel builds: react to changes in pip's handling of build constraints by @mmccarty in rapidsai#2927
• fix(build): build package on merge to release/* branch by @gforsyth in rapidsai#2929

New Contributors

• @mmccarty made their first contribution in rapidsai#2927

Full Changelog: https://github.com/rapidsai/raft/compare/v26.02.00a...release/26.02

raft 25.12.00 (10 Dec 2025)

🚨 Breaking Changes

• More consistent container policies & host memory resource by @achirkin in rapidsai#2835
• Require CUDA 12.2+ by @jakirkham in rapidsai#2850

🐛 Bug Fixes

• Correct tagging in the irecv function of the STD communicator by @viclafargue in rapidsai#2829
• Fix copyright hook file exclusion by @KyleFromNVIDIA in rapidsai#2840
• Properly guard usage of openmp function calls by @robertmaynard in rapidsai#2839
• Fix reduce mdspan API by @lowener in rapidsai#2853
• Fix for STD comm waitall function by @viclafargue in rapidsai#2852
• Pin Cython pre-3.2.0 and PyTest pre-9 by @jakirkham in rapidsai#2864
• refactored update-version.sh to handle new branching strategy by @rockhowse in rapidsai#2863
• Fix laplacian scaling coefficients by @aamijar in rapidsai#2871
• Revert "Remove Deprecated API (#2813)" by @csadorf in rapidsai#2881

📖 Documentation

• Use current system architecture in conda environment creation command by @bdice in rapidsai#2862

🚀 New Features

• BENCH_PRIMS: convenience reporting of benchmark parameters and read throughput by @achirkin in rapidsai#2824

🛠️ Improvements

• Update to rapids-logger 0.2 by @bdice in rapidsai#2828
• Enable sccache-dist connection pool by @trxcllnt in rapidsai#2837
• Use main in RAPIDS_BRANCH by @bdice in rapidsai#2842
• Use main shared-workflows branch by @bdice in rapidsai#2844
• Use SPDX for all copyright headers by @KyleFromNVIDIA in rapidsai#2845
• Use ruff-check, ruff-format instead of black, flake8, isort by @KyleFromNVIDIA in rapidsai#2855
• Remove shims for CCCL < 3.1 compatibility by @bdice in rapidsai#2858
• Always convert warnings to errors by @jakirkham in rapidsai#2857
• Lanczos Solver with COO input and cusparse wrappers by @aamijar in rapidsai#2851
• COO support in sparse matrix utilities by @aamijar in rapidsai#2861
• Update RMM includes from <rmm/mr/device/*> to <rmm/mr/*> by @bdice in rapidsai#2867
• Use sccache-dist build cluster for conda and wheel builds by @trxcllnt in rapidsai#2859
• Remove Deprecated API by @jnke2016 in rapidsai#2813

New Contributors

• @rockhowse made their first contribution in rapidsai#2863

Full Changelog: https://github.com/rapidsai/raft/compare/v25.12.00a...release/25.12

raft 25.10.00 (8 Oct 2025)

🐛 Bug Fixes

• Workaround for an illegal memory access on SM 120 devices (#2821) @achirkin
• Fix sparse select_k: don't write beyond min(input_len, k) (#2814) @achirkin
• [BUG] Fix compilation error in matrix/detail/gather.cuh (#2811) @enp1s0
• Fix select_k for negative bfloat16 (#2799) @apivovarov
• Fix index types for coo kernels (#2793) @aamijar
• Fix the GEMM pointer mode setting (#2777) @achirkin
• Fix host_vector_policy issue (#2739) @viclafargue

📖 Documentation

• Fix UCX-Py mention to UCXX in docstring (#2804) @pentschev

🚀 New Features

• Update cutlass to a version that supports CUDA 13 (#2774) @robertmaynard

🛠️ Improvements

• Fix missed deps in update-version.sh (#2826) @AyodeAwe
• Empty commit to trigger a build (#2816) @msarahan
• Make warpsort kernels use the IEEE 754 bit representation for ordering (#2807) @achirkin
• Configure repo for automatic release notes generation (#2806) @AyodeAwe
• Support < 2 element arrays in rand_index/adjusted_rand_index (#2805) @jcrist
• update dependencies: use cuda-toolkit wheels (#2802) @jameslamb
• Use branch-25.10 again (#2800) @jameslamb
• Remove CMake find UCX package (#2798) @pentschev
• use dask-cuda[cu12, cu13] extras for wheel dependencies (#2797) @jameslamb
• Remove UCX-Py (#2791) @pentschev
• Update rapids-dependency-file-generator (#2790) @KyleFromNVIDIA
• Build and test with CUDA 13.0.0 (#2787) @jameslamb
• Fix template arg passing in adjusted_rand_index (#2785) @jinsolp
• Use build cluster in devcontainers (#2781) @trxcllnt
• Use rapids_cuda_enable_fatbin_compression (#2780) @robertmaynard
• Increase Dask tests verbosity in CI (#2779) @pentschev
• Update rapids_config to handle user defined branch name (#2778) @robertmaynard
• [REVIEW] Fix: skip default_allocation_limit() if unnecessary (#2775) @i-Pear
• Update rapids-build-backend to 0.4.1 (#2773) @KyleFromNVIDIA
• ci(labeler): update labeler action to [@v5 (#2772) @gforsyth](https://github.com/v5 (#2772) @gforsyth)
• Register bfloat16/bfloat162 in util/vectorized.cuh (#2769) @apivovarov
• Use mdspan::index_type to Only Instantiate Specific Kernels (#2767) @tarang-jain
• Allow latest OS in devcontainers (#2759) @bdice
• Update build infra to support new branching strategy (#2751) @robertmaynard
• Use GCC 14 in conda builds. (#2708) @vyasr

raft 25.08.00 (6 Aug 2025)

🚨 Breaking Changes

• MatrixLinewiseOp compile-time-invocation (#2701) @aamijar
• Remove CUDA 11 from dependencies.yaml (#2695) @KyleFromNVIDIA
• stop uploading packages to downloads.rapids.ai (#2688) @jameslamb
• Reduce instantiations of Reduction kernels (#2679) @divyegala

🐛 Bug Fixes

• Fix stream sync for Copy2DAsync test (#2744) @lowener
• Several small fixes to make Raft compile with LLVM. (#2735) @vitor1001
• Add missing header (#2734) @vitor1001
• [REVIEW] Fix static initialization order fiasco in lanczos.cu (#2733) @legrosbuffle
• [REVIEW] Fix assertion in fill_indices_by_rows_kernel. (#2732) @legrosbuffle
• libucx: consider post-releases in wheel builds (#2729) @jameslamb
• Fix laplacian cast (#2725) @aamijar
• Fix excess_subsample (#2723) @mfoerste4
• Fix the constructor for coordinate_structure for non-zero nnz. (#2717) @legrosbuffle
• [REVIEW] Fix compile error when using mdbuffer with all-static extents. (#2716) @legrosbuffle
• Fix unsafe cast coo_remove_scalar (#2713) @aamijar
• Fix laplacian self-loops (#2712) @aamijar
• [REVIEW] Fix a few memory leaks. (#2710) @legrosbuffle
• Fix MST bug for graph with identical edge weights (#2707) @jnke2016
• Missed update accounting for reduction related APIs (#2704) @divyegala
• Adding GH_TOKEN pass-through to summarize job (#2702) @msarahan
• Work around Cython ctypedef bug (#2686) @vyasr

📖 Documentation

• add docs on CI workflow inputs (#2728) @jameslamb

🛠️ Improvements

• An additional small change to remove cuda 11 stuff (#2763) @cjnolet
• Removing CUDA 11 from docs and code (#2757) @cjnolet
• fix(docker): use versioned -latest tag for all rapidsai images (#2745) @gforsyth
• Update protocol name for UCX-Py tests (#2743) @pentschev
• Remove sphinx upper bound (#2742) @bdice
• remove cuspatial references, avoid triggering tests on clang-format config changes (#2740) @jameslamb
• MST Edge Case (#2736) @tarang-jain
• Add missing #include <cassert> in cpp/include/raft/core/math.hpp (#2730) @trxcllnt
• Update leftover CUDA 12.8 to 12.9 in docs (#2724) @jakirkham
• Fix docs lanczos solver (#2722) @aamijar
• Use CUDA 12.9 in Conda, Devcontainers, Spark, GHA, etc. (#2721) @jakirkham
• Remove nvidia and dask channels (#2720) @vyasr
• [REVIEW] Fix compile error of abs_op when compiling with clang (#2718) @legrosbuffle
• Avoid using internal method std::experimental::details::alignTo(). (#2714) @vitor1001
• refactor(shellcheck): fix all remaining warnings/errors (#2703) @gforsyth
• MatrixLinewiseOp compile-time-invocation (#2701) @aamijar
• Remove pytest pin (#2699) @vyasr
• Fix several issues that breaks LLVM (#2698) @vitor1001
• Remove CUDA 11 from dependencies.yaml (#2695) @KyleFromNVIDIA
• Remove CUDA 11 devcontainers and update CI scripts (#2690) @bdice
• refactor(rattler): remove cuda11 options and general cleanup (#2689) @gforsyth
• stop uploading packages to downloads.rapids.ai (#2688) @jameslamb
• fix(devcontainers): typo in container name (#2687) @gforsyth
• Reduce instantiations of Reduction kernels (#2679) @divyegala
• Forward-merge branch-25.06 into branch-25.08 (#2675) @divyegala
• Add support for F16 in linalg::transpose (#2672) @enp1s0
• Forward-merge branch-25.06 into branch-25.08 (#2664) @gforsyth
• Support coo_matrix in coo_symmetrize and coo_remove_scalar (#2662) @aamijar
• Lanczos Solver which=SA,SM,LA,LM argument (#2628) @aamijar

raft 25.06.00 (5 Jun 2025)

🚨 Breaking Changes

• Decoupling multi gpu resources from nccl usage (#2647) @jinsolp

🐛 Bug Fixes

• NCCL comm resource fix (#2692) @viclafargue
• Fix the launch bounds for nn-descent kernel for 1210 and remove nn-descent tests (#2691) @viclafargue
• Prefer host gather when dataset is available both on host and device (#2671) @tfeher
• Fix warnings treated as errors downstream in cuVS (#2644) @achirkin
• Fix nccl_comm.hpp warning: #83-D: type qualifier specified more than once (#2643) @achirkin
• NVTX: null destination pointer warning-treated-as-error (#2639) @achirkin
• Add UCXX and NCCL to libraft conda recipe (#2636) @divyegala
• Fix building cutlass (#2619) @miscco
• Fix COO symmetrization (#2582) @viclafargue

🚀 New Features

• [Feat] add cudaMemcpy2DAsync wrapper (#2674) @rhdong
• Python wrapper for device_resources_snmg (#2666) @jinsolp
• Laplacian normalization primitives (#2648) @aamijar
• [FEA] Matrix shift rows and columns (#2634) @jinsolp
• Use NCCL wheels from PyPI for CUDA 12 builds (#2629) @divyegala
• Support strided matrix view as an input to matrix::samples_rows (#2626) @enp1s0
• [Feat] add support for bm25 and tfidf (#2567) @jperez999

🛠️ Improvements

• use 'rapids-init-pip' in wheel CI, other CI changes (#2677) @jameslamb
• Dask 2025.4.1 compatibility (#2673) @TomAugspurger
• Finish CUDA 12.9 migration and use branch-25.06 workflows (#2669) @bdice
• Update to clang 20 (#2665) @bdice
• Quote head_rev in conda recipes (#2660) @bdice
• CUDA 12.9 use updated compression flags (#2657) @robertmaynard
• Build and test with CUDA 12.9.0 (#2655) @bdice
• Exclude librmm.so from auditwheel (#2654) @bdice
• Fix cub include in normalize.cuh (#2652) @lowener
• Add support for Python 3.13 (#2649) @gforsyth
• Decoupling multi gpu resources from nccl usage (#2647) @jinsolp
• [BUGFIX] Fixed quoting in wheel paths in pylibraft and raft_dask wheel tests (#2645) @VenkateshJaya
• Download build artifacts from Github for CI (#2640) @VenkateshJaya
• Limit allowed wheel sizes (#2638) @divyegala
• Remove CUDA whole compilation ODR violations (#2633) @divyegala
• refactor(rattler): enable strict channel priority for builds (#2632) @gforsyth
• Vendor RAPIDS.cmake (#2631) @bdice
• Replace Thrust iterator facilities and replace them with libcu++ ones (#2627) @miscco
• Port all conda recipes to rattler-build (#2623) @gforsyth
• Add missing thrust include (#2618) @miscco
• Moving wheel builds to specified location and uploading build artifacts to Github (#2617) @VenkateshJaya
• Fixed pytest marker warnings by removing unused pytest.ini (#2591) @TomAugspurger
• Introduction of the raft::device_resources_snmg type (#2549) @viclafargue
• Create a NCCL sub-communicator using ncclCommSplit (#2495) @seunghwak

raft 25.04.00 (9 Apr 2025)

🚨 Breaking Changes

• Account for cugraph API breakage (#2581) @divyegala
• Use new rapids-logger library (#2566) @vyasr

🐛 Bug Fixes

• Backport build patch fix (#2620) @KyleFromNVIDIA
• Revert "Temporarily increase max_days_without_success (#2602)" (#2613" (#2613)) @divyegala
• Relax max duplicates in batched NN Descent (#2610) @jinsolp
• [Fix] Lanczos solver gemv fix (#2607) @aamijar
• [Fix] select-k-csr failure on CUDA11.x + H100 (#2604) @rhdong
• Temporarily increase max_days_without_success (#2602) @divyegala
• Swap blocks and threads_per_block in compute_graph_laplacian (#2597) @jcrist
• [BUG] Fix illegal memory access in linalg::reduction (#2592) @enp1s0
• Require sphinx<8.2.0 (#2590) @KyleFromNVIDIA
• Account for cugraph API breakage (#2581) @divyegala
• #include <numeric> for std::iota (#2578) @benfred
• Fix Laplacian calculation in spectral partitioning (#2568) @wphicks
• Take argument by const& as the input range is const (#2558) @miscco
• Allow some of the sparse utility functions to handle larger matrices (#2541) @viclafargue

🛠️ Improvements

• ci: pre-filter 11.4 jobs before they are enabled in shared workflows (#2608) @gforsyth
• Use conda-build instead of conda-mambabuild (#2595) @bdice
• Replace cub::Sum and cub::Max with cuda::std::plus and cuda::maximum (#2594) @miscco
• Update all conda_build_config.yamls RAPIDS UCX version (#2589) @jakirkham
• Drop cub::TransformInputIterator in favor of thrust::transform_iterator (#2588) @miscco
• Consolidate more Conda solves in CI (#2587) @KyleFromNVIDIA
• Fix duplicate indices in batch NN Descent (#2586) @jinsolp
• Require CMake 3.30.4 (#2584) @robertmaynard
• Create Conda CI test env in one step (#2580) @KyleFromNVIDIA
• Use shared-workflows branch-25.04 (#2576) @bdice
• Add shellcheck to pre-commit and fix warnings (#2575) @gforsyth
• Add build_type input field for test.yaml (#2573) @gforsyth
• Use rapids-pip-retry in CI jobs that might need retries (#2571) @gforsyth
• Avoid limited memory adaptor issue in balanced KMeans (#2570) @csadorf
• update telemetry and retarget 25.04 (#2569) @msarahan
• Use new rapids-logger library (#2566) @vyasr
• disallow fallback to Make in Python builds (#2563) @jameslamb
• Forward-merge branch-25.02 into branch-25.04 (#2561) @bdice
• Migrate to NVKS for amd64 CI runners (#2559) @bdice
• Add verify-codeowners hook (#2557) @KyleFromNVIDIA

raft 25.02.00 (13 Feb 2025)

🚨 Breaking Changes

• Update pip devcontainers to UCX 1.18 (#2550) @jameslamb
• Switch over to rapids-logger (#2530) @vyasr
• Adapt to rmm logger changes (#2513) @vyasr

🐛 Bug Fixes

• Rename test to tests. (#2546) @bdice
• Fix bit order of RMAT Rectangular Generator to match expectation (#2542) @mfoerste4
• Fix broken link to python doc (#2537) @lowener
• Fix lanczos solver integer overflow (#2536) @viclafargue
• Fix rnd bit generation in rmat_rectangular_kernel (#2524) @tfeher

📖 Documentation

• Fix docs builds (#2562) @bdice
• [DOC] Fix sample codes (#2518) @enp1s0

🚀 New Features

• Add cuda 12.8 support (#2551) @robertmaynard
• Add support for different data type of bitset (#2535) @lowener
• [Feat] Support bitset_to_csr (#2523) @rhdong
• Remove upper bounds on cuda-python to allow 12.6.2 and 11.8.5 (#2517) @bdice

🛠️ Improvements

• Revert CUDA 12.8 shared workflow branch changes (#2560) @vyasr
• Build and test with CUDA 12.8.0 (#2555) @bdice
• Update pip devcontainers to UCX 1.18 (#2550) @jameslamb
• use dynamic CUDA wheels on CUDA 11 (#2548) @jameslamb
• Normalize whitespace (#2547) @bdice
• Use cuda.bindings layout. (#2545) @bdice
• Revert "Introduction of the raft::device_resources_snmg type (#2487)" (#2543" (#2543)) @cjnolet
• Add missing #include <cstdint> (#2540) @jakirkham
• Use GCC 13 in CUDA 12 conda builds. (#2539) @bdice
• Use rapids-cmake for the logger (#2534) @vyasr
• Check if nightlies have succeeded recently enough (#2533) @vyasr
• remove unused 'joblib' and 'numba' dependencies, other packaging cleanup (#2532) @jameslamb
• introduce libraft wheels (#2531) @jameslamb
• Switch over to rapids-logger (#2530) @vyasr
• reduce duplication, removed unused things in dependencies.yaml (#2529) @jameslamb
• Update cuda-python lower bounds to 12.6.2 / 11.8.5 (#2522) @bdice
• [Opt] Optimizing the performance of bitmap_to_csr (#2516) @rhdong
• prefer system install of UCX in devcontainers, update outdated RAPIDS references (#2514) @jameslamb
• Adapt to rmm logger changes (#2513) @vyasr
• Require approval to run CI on draft PRs (#2512) @bdice
• Shrink wheel size limit following removal of vector search APIs. (#2509) @bdice
• Forward-merge branch-24.12 to branch-25.02 (#2508) @bdice
• Introduction of the raft::device_resources_snmg type (#2487) @viclafargue
• Add breaking change workflow trigger (#2482) @AyodeAwe
• Remove 'sample' parameter from stats::mean API (#2389) @mfoerste4

raft 24.12.00 (11 Dec 2024)

🚨 Breaking Changes

• Do not initialize the pinned mdarray at construction time (#2478) @achirkin

🐛 Bug Fixes

• Skip gtests for new lanczos solver when CUDA version is 11.4 or below. (#2520) @cjnolet
• Switch assert to static_assert (#2510) @divyegala
• Revert use of new Lanczos solver in spectral clustering (#2507) @lowener
• Put a ceiling on cuda-python (#2486) @bdice
• Don't presume pointers location infers usability. (#2480) @robertmaynard
• Use Python for sccache hit rate computation. (#2474) @bdice
• Allow compilation with CUDA 12.6.1 (#2469) @robertmaynard

🚀 New Features

• [FEA] Lanczos solver v2 (#2481) @lowener

🛠️ Improvements

• Skip gtests for Rmat Lanczos tests with cuda <= 11.4 (#2525) @benfred
• Upgrade to latest cutlass version (#2503) @vyasr
• Removing some left over places where implicit instantiations were being ignored in headers (#2501) @cjnolet
• Remove leftover template project code. (#2500) @bdice
• 2412 remove libraft vss instantiations (#2498) @cjnolet
• Remove raft-ann-bench (#2497) @cjnolet
• Pin FAISS Version for raft-ann-bench (#2496) @tarang-jain
• enforce wheel size limits and README formatting in CI, put a ceiling on Cython dependency (#2490) @jameslamb
• Do not initialize the pinned mdarray at construction time (#2478) @achirkin
• Use environment variables in cache hit rate computation. (#2475) @bdice
• devcontainer: replace VAULT_HOST with AWS_ROLE_ARN (#2472) @jjacobelli
• print sccache stats in builds (#2470) @jameslamb
• make package installations in CI stricter (#2467) @jameslamb
• Prune workflows based on changed files (#2466) @KyleFromNVIDIA
• Merge branch-24.10 into branch-24.12 (#2461) @jameslamb
• Update all rmm imports to use pylibrmm/librmm (#2451) @Matt711

raft 24.10.00 (9 Oct 2024)

🚨 Breaking Changes

• [Feat] add repeat, sparsity, eval_n_elements APIs to bitset (#2439) @rhdong

🐛 Bug Fixes

• Disable NN Descent Batch tests temporarily (#2453) @divyegala
• Fix sed syntax in update-version.sh (#2441) @raydouglass
• Use runtime check of cudart version for eig (#2430) @lowener
• [BUG] Fix bitset function visibility (#2429) @lowener
• Exclude any kernel symbol that uses cutlass (#2425) @robertmaynard

🚀 New Features

• [Feat] add repeat, sparsity, eval_n_elements APIs to bitset (#2439) @rhdong
• [Opt] Enforce the UT Coverity and add benchmark for transpose (#2438) @rhdong
• [FEA] Support for half-float mixed precise in brute-force (#2382) @rhdong

🛠️ Improvements

• bump NCCL floor to 2.19 (#2458) @jameslamb
• Deprecating vector search APIs and updating README accordingly (#2448) @cjnolet
• Update update-version.sh to use packaging lib (#2447) @AyodeAwe
• Switch traceback to native (#2446) @galipremsagar
• bump NCCL floor to 2.18.1.1 (#2443) @jameslamb
• Add missing cuda_suffixed: true (#2440) @trxcllnt
• Use CI workflow branch 'branch-24.10' again (#2437) @jameslamb
• Update to flake8 7.1.1. (#2435) @bdice
• Update fmt (to 11.0.2) and spdlog (to 1.14.1). (#2433) @jameslamb
• Allow coo_sort to work on int64_t indices (#2432) @benfred
• Adding NCCL clique to the RAFT handle (#2431) @viclafargue
• Add support for Python 3.12 (#2428) @jameslamb
• Update rapidsai/pre-commit-hooks (#2420) @KyleFromNVIDIA
• Drop Python 3.9 support (#2417) @jameslamb
• Use CUDA math wheels (#2415) @KyleFromNVIDIA
• Remove NumPy <2 pin (#2414) @seberg
• Update pre-commit hooks (#2409) @KyleFromNVIDIA
• Improve update-version.sh (#2408) @bdice
• Use tool.scikit-build.cmake.version, set scikit-build-core minimum-version (#2406) @jameslamb
• [FEA] Batching NN Descent (#2403) @jinsolp
• Update pip devcontainers to UCX v1.17.0 (#2401) @jameslamb
• Merge branch-24.08 into branch-24.10 (#2397) @jameslamb

raft 24.08.00 (7 Aug 2024)

🚨 Breaking Changes

• [Refactor] move popc to under util (#2394) @rhdong
• [Opt] Expose the detail::popc as public API (#2346) @rhdong

🐛 Bug Fixes

• Add timeout to UCXX generic operations (#2398) @pentschev
• [Fix] bitmap set/test issue (#2371) @rhdong
• Fix 0 recall issue in raft_cagra_hnswlib ANN benchmark (#2369) @divyegala
• Fix ef setting in HNSW wrapper (#2367) @divyegala
• Fix cagra graph opt bug (#2365) @enp1s0
• Fix a bug where the wrong API is used to free the memory (#2361) @PointKernel
• Allow anonymous user in devcontainer name (#2355) @bdice
• Fix compilation error when _CLK_BREAKDOWN is defined in cagra. (#2350) @jiangyinzuo
• ensure raft-dask wheel tests install pylibraft wheel from the same CI run, fix wheel dependencies (#2349) @jameslamb
• Change --config-setting to --config-settings (#2342) @KyleFromNVIDIA
• Add workaround for syevd in CUDA 12.0 (#2332) @lowener

🚀 New Features

• [FEA] add the support of masked_matmul (#2362) @rhdong
• [FEA] Dice Distance for Dense Inputs (#2359) @aamijar
• [Opt] Expose the detail::popc as public API (#2346) @rhdong
• Enable distance return for NN Descent (#2345) @jinsolp

🛠️ Improvements

• [Refactor] move popc to under util (#2394) @rhdong
• split up CUDA-suffixed dependencies in dependencies.yaml (#2388) @jameslamb
• Use workflow branch 24.08 again (#2385) @KyleFromNVIDIA
• Add cusparseSpMV_preprocess to cusparse wrapper (#2384) @Kh4ster
• Consolidate SUM reductions (#2381) @mfoerste4
• Use slicing kernel to copy distances inside NN Descent (#2380) @jinsolp
• Build and test with CUDA 12.5.1 (#2378) @KyleFromNVIDIA
• Add CUDA_STATIC_MATH_LIBRARIES (#2376) @KyleFromNVIDIA
• skip CMake 3.30.0 (#2375) @jameslamb
• Use verify-alpha-spec hook (#2373) @KyleFromNVIDIA
• Binarize Dice Distance for Dense Inputs (#2370) @aamijar
• [FEA] Add distance epilogue for NN Descent (#2364) @jinsolp
• resolve dependency-file-generator warning, other rapids-build-backend followup (#2360) @jameslamb
• Remove text builds of documentation (#2354) @vyasr
• Use default init in reduction (#2351) @akifcorduk
• ensure update-version.sh preserves alpha spec, add tests on version constants (#2344) @jameslamb
• remove unnecessary 'setuptools' dependencies (#2343) @jameslamb
• Use rapids-build-backend (#2331) @KyleFromNVIDIA
• Add FAISS with RAFT enabled Benchmarking to raft-ann-bench (#2026) @tarang-jain

raft 24.06.00 (5 Jun 2024)

🚨 Breaking Changes

• Rename raft-ann-bench module to raft_ann_bench (#2333) @KyleFromNVIDIA
• Scaling workspace resources (#2322) @achirkin
• [REVIEW] Adjust UCX dependencies (#2304) @pentschev
• Convert device_memory_resource* to device_async_resource_ref (#2269) @harrism

🐛 Bug Fixes

• Fix import of VERSION file in raft-ann-bench (#2338) @KyleFromNVIDIA
• Rename raft-ann-bench module to raft_ann_bench (#2333) @KyleFromNVIDIA
• Support building faiss main statically (#2323) @robertmaynard
• Refactor spectral scale_obs to use existing normalization function (#2319) @ChuckHastings
• Correct initializer list order found by cuvs (#2317) @robertmaynard
• ANN_BENCH: enable move semantics for configured_raft_resources (#2311) @achirkin
• Revert "Build C++ wheel (#2264)" (#2305" (#2305)) @vyasr
• Revert "Add compile-library by default on pylibraft build" (#2300) @vyasr
• Add VERSION to raft-ann-bench package (#2299) @KyleFromNVIDIA
• Remove nonexistent job from workflow (#2298) @vyasr
• libucx should be run dependency of raft-dask (#2296) @divyegala
• Fix clang intrinsic warning (#2292) @aaronmondal
• Replace too long index file name with hash in ANN bench (#2280) @tfeher
• Fix build command for C++ compilation (#2270) @lowener
• Fix a compilation error in CAGRA when enabling log output (#2262) @enp1s0
• Correct member initialization order (#2254) @robertmaynard
• Fix time computation in CAGRA notebook (#2231) @lowener

📖 Documentation

• Fix citation info (#2318) @enp1s0

🚀 New Features

• Scaling workspace resources (#2322) @achirkin
• ANN_BENCH: AnnGPU::uses_stream() for optional algo GPU sync (#2314) @achirkin
• [FEA] Split Bitset code (#2295) @lowener
• [FEA] support of prefiltered brute force (#2294) @rhdong
• Always use a static gtest and gbench (#2265) @robertmaynard
• Build C++ wheel (#2264) @vyasr
• InnerProduct Distance Metric for CAGRA search (#2260) @tarang-jain
• [FEA] Add support for select_k on CSR matrix (#2140) @rhdong

🛠️ Improvements

• ANN_BENCH: common AnnBase::index_type (#2315) @achirkin
• ANN_BENCH: split instances of RaftCagra into multiple files (#2313) @achirkin
• ANN_BENCH: a global pool of result buffers across benchmark cases (#2312) @achirkin
• Remove the shared state and the mutex from NVTX internals (#2310) @achirkin
• docs: update README.md (#2308) @eltociear
• [REVIEW] Reenable raft-dask wheel tests requiring UCX-Py (#2307) @pentschev
• [REVIEW] Adjust UCX dependencies (#2304) @pentschev
• Overhaul ops-codeowners (#2303) @raydouglass
• Make thrust nosync execution policy the default thrust policy (#2302) @abc99lr
• InnerProduct testing for CAGRA+HNSW (#2297) @divyegala
• Enable warnings as errors for Python tests (#2288) @mroeschke
• Normalize dataset vectors in the CAGRA InnerProduct tests (#2287) @enp1s0
• Use dynamic version for raft-ann-bench (#2285) @KyleFromNVIDIA
• Make 'librmm' a 'host' dependency for conda packages (#2284) @jameslamb
• Fix comments in cpp/include/raft/neighbors/cagra_serialize.cuh (#2283) @jiangyinzuo
• Only use functions in the limited API (#2282) @vyasr
• define 'ucx' pytest marker (#2281) @jameslamb
• Migrate to {{ stdlib("c") }} (#2278) @hcho3
• add --rm and --name to devcontainer run args (#2275) @trxcllnt
• Update pip devcontainers to UCX v1.15.0 (#2274) @trxcllnt
• #ifdef out pragma deprecation warning messages (#2271) @trxcllnt
• Convert device_memory_resource* to device_async_resource_ref (#2269) @harrism
• Update the developer's guide with new copyright hook (#2266) @KyleFromNVIDIA
• Improve coalesced reduction performance for tall and thin matrices (up to 2.6x faster) (#2259) @Nyrio
• Adds missing files to update-version.sh (#2255) @AyodeAwe
• Enable all tests for arm64 jobs (#2248) @galipremsagar
• Update nvtx3 link in cmake (#2246) @lowener
• Add CAGRA-Q subspace dim = 4 support (#2244) @enp1s0
• Get rid of cuco::sentinel namespace (#2243) @PointKernel
• Replace usages of raw get_upstream with get_upstream_resource() (#2207) @miscco
• Set the import mode for dask tests (#2142) @vyasr
• Add UCXX support (#1983) @pentschev

raft 24.04.00 (10 Apr 2024)

🐛 Bug Fixes

• Update pre-commit-hooks to v0.0.3 (#2239) @KyleFromNVIDIA
• MAINT: Simplify NCCL worker rank identification (#2228) @VibhuJawa
• Fix bug in blockRankedReduce (#2226) @akifcorduk
• Fix illegal acces mean/stdev, sum add Kahan Summation (#2223) @mfoerste4
• Batch cutlass distance kernels along N matrix dim (#2215) @mdoijade
• Fix out of bounds access in sum kernel (#2183) @tfeher
• Fix ANN bench ground truth generation for k>1024 (#2180) @tfeher
• Fixing cusparse aligned address issue and adding note (#2179) @cjnolet
• Launch neighborhood_recall kernel on CUDA stream (#2156) @divyegala
• Add compile-library by default on pylibraft build (#2090) @lowener

📖 Documentation

• Adding cuVS notice to README and front page of docs. (#2224) @cjnolet

🚀 New Features

• Add CAGRA-Q to ANN benchmarks (#2233) @achirkin
• Add CAGRA-Q build (compression) (#2213) @achirkin
• CAGRA-Q search (#2206) @enp1s0
• Demangle backtrace symbols on raft error (#2188) @achirkin
• Reapply: Support for fp16 in CAGRA and IVF-PQ (#2172) @achirkin
• Remove supports_streams from custom RAFT memory resources (#2121) @harrism
• [FEA] Add support for bitmap_view & the API of bitmap_to_csr (#2109) @rhdong

🛠️ Improvements

• Use conda env create --yes instead of --force (#2247) @bdice
• Align ucx version pinning with ucx-py/ucxx. (#2227) @bdice
• Add upper bound to prevent usage of NumPy 2 (#2222) @bdice
• Performance optimization of IVF-flat / select_k (#2221) @mfoerste4
• Replace local copyright check with pre-commit-hooks verify-copyright (#2220) @KyleFromNVIDIA
• Remove hard-coding of RAPIDS version where possible (#2219) @KyleFromNVIDIA
• Fix style. (#2214) @bdice
• Add explicit instantiations for IVF-PQ search kernels used in tests (#2212) @tfeher
• Improve RBC eps-neighborhood query performance (#2211) @mfoerste4
• Add test for spmm (#2210) @mfoerste4
• Only install necessary components in conda packages. (#2209) @bdice
• Automate C++ include file grouping and ordering using clang-format (#2202) @harrism
• Add support for Python 3.11, require NumPy 1.23+ (#2200) @jameslamb
• Pass std::optional instead of thrust::optional to RMM (#2199) @trxcllnt
• Update devcontainers to CUDA Toolkit 12.2 (#2192) @trxcllnt
• target branch-24.04 for GitHub Actions workflows (#2189) @jameslamb
• Fixing workaround for cuSPARSE bug with correct copy dimensions (#2185) @mfoerste4
• Allow topk larger than 1024 in CAGRA (#2181) @benfred
• IVF-FLAT support k > 256 (#2169) @mfoerste4
• Add environment-agnostic scripts for running ctests and pytests (#2165) @trxcllnt
• Ensure that ctest is called with --no-tests=error. (#2163) @bdice
• Update ops-bot.yaml (#2158) @AyodeAwe
• random sampling of dataset rows with improved memory utilization (#2155) @tfeher
• [FIX] Ensure hnswlib can be found from RAFT's build dir (#2145) @trxcllnt
• Improve analysis experience for ANN benchmarks (#2139) @achirkin
• Enable CAGRA index building without adding dataset to the index (#2126) @tfeher
• Add fused cosine 1-NN cutlass based kernel (#2125) @mdoijade
• Update raft for compatibility with the latest cuco (#2118) @PointKernel
• Support CUDA 12.2 (#2092) @jameslamb
• Cache IVF-PQ and select-warpsort kernel launch parameters to reduce latency (#1786) @achirkin

raft 24.02.00 (12 Feb 2024)

🚨 Breaking Changes

• Switch to scikit-build-core (#2051) @vyasr
• Update to CCCL 2.2.0. (#2049) @bdice
• Update raft-ann-bench output filenames and add features to plotting (#2043) @divyegala
• Remove selection_faiss (#2027) @benfred

🐛 Bug Fixes

• fix is_row/col_order for strided layouts (#2173) @mfoerste4
• Fix failing C++ tests and revert #2097, #2085. (#2168) @cjnolet
• Exclude tests from builds (#2162) @vyasr
• [HOTFIX] 24.02 Revert Random Sampling (#2144) @cjnolet
• Pin to pytest 7. (#2137) @bdice
• Conditionally include hnsw wrapper source in CMake (#2135) @divyegala
• [BUG] Fix SPMM strided view (#2124) @lowener
• Fixing small bug in CUSPARSE spmm w/ CUDA 12.2 (#2117) @cjnolet
• [BUG] Fix num_cta_per_query div (#2107) @lowener
• Remove extraneous host pinnings from libraft-headers-only. (#2102) @bdice
• Remove unneeded CI symbol excludes (#2098) @robertmaynard
• Properly taking ownership of nccl subcomm (and destroying it) (#2094) @cjnolet
• Fix max_queries for CAGRA (#2081) @lowener
• Fix compile failure on RTX 4090 (#2076) @JieFengWang
• Fix a crash in FAISS benchmark wrapper introduced in #2021 (#2062) @achirkin
• Correct function that wasn't returning a value (#2045) @robertmaynard
• Fixing small bug in raft-ann-bench (#2041) @cjnolet
• Make device_resources accessed from device_resources_manager thread-safe (#2030) @wphicks
• Fix ann-bench multithreading (#2021) @achirkin
• Fix ci/checks/copyright.py to mirror RAPIDS reference (#2008) @divyegala
• Fix pyproject versions (#2002) @vyasr

📖 Documentation

• Adding license info for wiki-all dataset (#2129) @cjnolet
• [DOC] Documentation updates for release 24.02 (#2093) @cjnolet
• Fix errors with ingroup exposed by doxygen 1.10 (#2079) @wphicks
• Fix a typo (#2070) @narangvivek10
• Add usage example for brute_force::build (#2029) @benfred
• Add filtering to vector search tutorial (#1996) @lowener

🚀 New Features

• Update to use rapids-cmake for all deps (#2096) @robertmaynard
• Add IVF-PQ example into the template project (#2091) @achirkin
• Support for fp16 in CAGRA and IVF-PQ (#2085) @achirkin
• Add random subsampling for IVF methods (#2077) @tfeher
• Update raft-ann-bench output filenames and add features to plotting (#2043) @divyegala
• Add brute_force index serialization (#2036) @wphicks
• Add eps-neighbor search via RBC (#2028) @mfoerste4
• libraft and pylibraft API for CAGRA build and HNSW search (#2022) @divyegala
• Export Pareto frontier in raft-ann-bench.data_export (#2009) @divyegala
• Implement maybe-owning multi-dimensional container (mdbuffer) (#1999) @wphicks
• Add support for 1024+ dim vectors in CAGRA search (#1994) @enp1s0
• Replace GEMM backend: cublas.gemm -> cublaslt.matmul (#1736) @achirkin

🛠️ Improvements

• Remove get_mem_info functions from RAFT custom memory resources (#2108) @harrism
• Replace call to mr::get_mem_info() (#2099) @harrism
• Allow topk larger than 1024 in CAGRA (#2097) @benfred
• Remove usages of rapids-env-update (#2095) @KyleFromNVIDIA
• Provide explicit pool size for pool_memory_resources and clean up includes (#2088) @harrism
• refactor CUDA versions in dependencies.yaml (#2086) @jameslamb
• ANN bench fix latency measurement overhead (#2084) @tfeher
• Remove hardcoded limit in print_results function (#2080) @narangvivek10
• [FEA] Add support for SDDMM by wrapping the cusparseSDDMM (#2067) (#2067 (#2067)) @rhdong
• Benchmark brute force knn (#2063) @benfred
• [BUG] fix empty initialization of device_ndarray in pylibraft (#2061) @mfoerste4
• Improve parallelism of refine host (#2059) @anaruse
• Subsampling for IVF-PQ codebook generation (#2052) @abc99lr
• Switch to scikit-build-core (#2051) @vyasr
• Update to CCCL 2.2.0. (#2049) @bdice
• Use cuda::proclaim_return_type on device lambda. (#2048) @bdice
• Removing code that explicitly compares equality of rmm memory resources (#2047) @cjnolet
• Add public enum for select-k algorithm selection (#2046) @benfred
• Update dependencies.yaml to new pip index (#2042) @vyasr
• Remove RAFT_BUILD_WHEELS and standardize Python builds (#2040) @vyasr
• Fix ucx-py version pinning in dependencies.yaml. (#2035) @bdice
• [REVIEW] Fix typos in parameter tuning guide (#2034) @abc99lr
• Add AIR-Top-k reference (#2031) @tfeher
• Remove selection_faiss (#2027) @benfred
• Fixing json parse error in raft-ann-bench.data_export (#2025) @cjnolet
• Updating cagra build constraint (#2016) @cjnolet
• Update to fmt 10.1.1 and spdlog 1.12.0. (#1957) @bdice
• Enable host dataset for IVF-Flat (#1635) @tfeher
• add half/bfloat support to myInf and abs (#1592) @Kh4ster

raft 23.12.00 (6 Dec 2023)

🐛 Bug Fixes

• Update actions/labeler to v4 (#2037) @raydouglass
• pylibraft only depends on numpy at runtime, not build time. (#2013) @bdice
• Fixes to update-version.sh (#1991) @raydouglass
• Adjusting end-to-end start time so it doesn't include stream creation time (#1989) @cjnolet
• CAGRA graph optimizer: clamp rev_graph_count (#1987) @tfeher
• Catching conversion errors in data_export instead of fully failing (#1979) @cjnolet
• Fix syncing mechanism in raft-ann-bench C++ search (#1961) @divyegala
• Fixing hnswlib in latency mode (#1959) @cjnolet
• Fix ucx-py alpha version update for raft-dask (#1953) @divyegala
• Reduce NN Descent test threshold (#1946) @divyegala
• Fixes to new YAML config raft-bench-ann (#1945) @divyegala
• Set RNG seeds in NN Descent to diagnose flaky tests (#1931) @divyegala
• Fix FAISS CPU algorithm names in raft-ann-bench (#1916) @divyegala
• Increase iterations in NN Descent tests to avoid flakiness (#1915) @divyegala
• Fix filepath in raft-ann-bench/split_groundtruth module (#1911) @divyegala
• Remove dynamic entry-points from raft-ann-bench (#1910) @benfred
• Remove unnecessary dataset path check in ANN bench (#1908) @tfeher
• Fixing Googletests and re-enabling in CI (#1904) @cjnolet
• Fix NN Descent overflows (#1875) @divyegala
• Build fix for CUDA 12.2 (#1870) @benfred
• [BUG] Fix a bug in NN descent (#1869) @enp1s0

📖 Documentation

• Brute Force Index documentation fix (#1944) @lowener
• Add wiki_all dataset config and documentation. (#1918) @cjnolet
• Updates to raft-ann-bench docs (#1905) @cjnolet
• End-to-end vector search tutorial in docs (#1776) @cjnolet

🚀 New Features

• Adding dry-run option to raft-ann-bench (#1970) @cjnolet
• Add ANN bench scripts to generate ground truth (#1967) @tfeher
• CAGRA build + HNSW search (#1956) @divyegala
• Verify conda-cpp-post-build-checks (#1935) @robertmaynard
• Make all cuda kernels have hidden visibility (#1898) @robertmaynard
• Update rapids-cmake functions to non-deprecated signatures (#1884) @robertmaynard
• [FEA] Helpers for identifying contiguous layouts. (#1861) @trivialfis
• Add raft::stats::neighborhood_recall (#1860) @divyegala
• [FEA] Helpers and CodePacker for IVF-PQ (#1826) @tarang-jain

🛠️ Improvements

• Pinning fmt and spdlog for raft-ann-bench-cpu (#2018) @cjnolet
• Build concurrency for nightly and merge triggers (#2011) @bdice
• Using EXPORT_SET in rapids_find_package_root (#2006) @cjnolet
• Remove static checks for serialization size (#1997) @cjnolet
• Skipping bad json parse (#1990) @cjnolet
• Update select-k heuristic (#1985) @benfred
• ANN bench: use different offset for each thread (#1981) @tfeher
• Allow raft-ann-bench/run to continue after encountering bad YAML configs (#1980) @divyegala
• Add build and search params to raft-ann-bench.data_export CSVs (#1971) @divyegala
• Use new rapids-dask-dependency metapackage for managing dask versions (#1968) @galipremsagar
• Remove unused header (#1960) @wphicks
• Adding pool back in and fixing cagra benchmark params (#1951) @cjnolet
• Add constraints to hnswlib in raft-bench-ann (#1949) @divyegala
• Add support for iterating over batches in bfknn (#1947) @benfred
• Fix ANN bench latency (#1940) @tfeher
• Add YAML config files to run parameter sweeps for ANN benchmarks (#1929) @divyegala
• Relax ucx pinning (#1927) @vyasr
• Try using contiguous rank to fix cuda_visible_devices (#1926) @VibhuJawa
• Unpin dask and distributed for 23.12 development (#1925) @galipremsagar
• Adding throughput and latency modes to raft-ann-bench (#1920) @cjnolet
• Providing aarch64 yaml environment files (#1914) @cjnolet
• CAGRA ANN bench: parse build options for IVF-PQ build algo (#1912) @tfeher
• Fix python script location in ANN bench description (#1906) @tfeher
• Refactor install/build guide. (#1899) @cjnolet
• Check return values of raft-ann-bench subprocess calls (#1897) @benfred
• ANN bench options to specify CAGRA graph and dataset locations (#1896) @cjnolet
• Add check-json to pre-commit linters, and fix invalid ann-bench JSON config (#1894) @benfred
• Use branch-23.12 workflows. (#1886) @bdice
• Setup Consistent Nightly Versions for Pip and Conda (#1880) @divyegala
• Fix and improve one-block radix select (#1878) @yong-wang
• [FEA] Improvements on bitset class (#1877) @lowener
• Branch 23.12 merge 23.10 (#1873) @AyodeAwe
• Branch 23.12 merge 23.10 (#1868) @cjnolet
• Replace raft::random calls to not use deprecated API (#1867) @lowener
• raft: Build CUDA 12.0 ARM conda packages. (#1853) @bdice
• Documentation for raft ANN benchmark containers. (#1833) @dantegd
• [FEA] Support vector deletion in ANN IVF (#1831) @lowener
• Provide a raft::copy overload for mdspan-to-mdspan copies (#1818) @wphicks
• Adding FAISS cpu to raft-ann-bench (#1814) @cjnolet

raft 23.10.00 (11 Oct 2023)

🚨 Breaking Changes

• Change CAGRA auto mode selection (#1830) @enp1s0
• Update CAGRA serialization (#1755) @benfred
• Improvements to ANN Benchmark Python scripts and docs (#1734) @divyegala
• Update to Cython 3.0.0 (#1688) @vyasr
• ANN-benchmarks: switch to use gbench (#1661) @achirkin

🐛 Bug Fixes

• [BUG] Fix a bug in the filtering operation in CAGRA multi-kernel (#1862) @enp1s0
• Fix conf file for benchmarking glove datasets (#1846) @dantegd
• raft-ann-bench package fixes for plotting and conf files (#1844) @dantegd
• Fix update-version.sh for all pyproject.toml files (#1839) @raydouglass
• Make RMM a run dependency of the raft-ann-bench conda package (#1838) @dantegd
• Printing actual exception in require base set (#1816) @cjnolet
• Adding rmm to raft-ann-bench dependencies (#1815) @cjnolet
• Use conda mambabuild not mamba mambabuild (#1812) @bdice
• Fix raft-dask naming in wheel builds (#1805) @divyegala
• neighbors::refine_host: check the dataset bounds (#1793) @achirkin
• [BUG] Fix search parameter check in CAGRA (#1784) @enp1s0
• IVF-Flat: fix search batching (#1764) @achirkin
• Using expanded distance computations in pylibraft (#1759) @cjnolet
• Fix ann-bench Documentation (#1754) @divyegala
• Make get_cache_idx a weak symbol with dummy template (#1733) @ahendriksen
• Fix IVF-PQ fused kernel performance problems (#1726) @achirkin
• Fix build.sh to enable NEIGHBORS_ANN_CAGRA_TEST (#1724) @enp1s0
• Fix template types for create_descriptor function. (#1680) @csadorf

📖 Documentation

• Fix the CAGRA paper citation (#1788) @enp1s0
• Add citation info for the CAGRA paper preprint (#1787) @enp1s0
• [DOC] Fix grouping for ANN in C++ doxygen (#1782) @lowener
• Update RAFT documentation (#1717) @lowener
• Additional polishing of README and docs (#1713) @cjnolet

🚀 New Features

• [FEA] Add bitset_filter for CAGRA indices removal (#1837) @lowener
• ann-bench: miscellaneous improvements (#1808) @achirkin
• [FEA] Add bitset for ANN pre-filtering and deletion (#1803) @lowener
• Adding config files for remaining (relevant) ann-benchmarks million-scale datasets (#1761) @cjnolet
• Port NN-descent algorithm to use in cagra::build() (#1748) @divyegala
• Adding conda build for libraft static (#1746) @cjnolet
• [FEA] Provide device_resources_manager for easy generation of device_resources (#1716) @wphicks

🛠️ Improvements

• Add option to brute_force index to store non-owning reference to norms (#1865) @benfred
• Pin dask and distributed for 23.10 release (#1864) @galipremsagar
• Update image names (#1835) @AyodeAwe
• Fixes for OOM during CAGRA benchmarks (#1832) @benfred
• Change CAGRA auto mode selection (#1830) @enp1s0
• Update to clang 16.0.6. (#1829) @bdice
• Add IVF-Flat C++ example (#1828) @tfeher
• matrix::select_k: extra tests and benchmarks (#1821) @achirkin
• Add index class for brute_force knn (#1817) @benfred
• [FEA] Add pre-filtering to CAGRA (#1811) @enp1s0
• More updates to ann-bench docs (#1810) @cjnolet
• Add best deep-100M configs for IVF-PQ to ANN benchmarks (#1807) @tfeher
• A few fixes to raft-ann-bench recipe and docs (#1806) @cjnolet
• Simplify wheel build scripts and allow alphas of RAPIDS dependencies (#1804) @divyegala
• Various fixes to reproducible benchmarks (#1800) @cjnolet
• ANN-bench: more flexible cuda_stub.hpp (#1792) @achirkin
• Add RAFT devcontainers (#1791) @trxcllnt
• Cagra memory optimizations (#1790) @benfred
• Fixing a couple security concerns in raft-dask nccl unique id generation (#1785) @cjnolet
• Don't serialize dataset with CAGRA bench (#1781) @benfred
• Use copy-pr-bot (#1774) @ajschmidt8
• Add GPU and CPU packages for ANN benchmarks (#1773) @dantegd
• Improvements to raft-ann-bench scripts, docs, and benchmarking implementations. (#1769) @cjnolet
• [REVIEW] Introducing host API for PCG (#1767) @vinaydes
• Unpin dask and distributed for 23.10 development (#1760) @galipremsagar
• Add ivf-flat notebook (#1758) @tfeher
• Update CAGRA serialization (#1755) @benfred
• Remove block size template parameter from CAGRA search (#1740) @enp1s0
• Add NVTX ranges for cagra search/serialize functions (#1737) @benfred
• Improvements to ANN Benchmark Python scripts and docs (#1734) @divyegala
• Fixing forward merger for 23.08 -> 23.10 (#1731) @cjnolet
• [FEA] Use CAGRA in C++ template (#1730) @lowener
• fixed box around raft image (#1710) @nwstephens
• Enable CUTLASS-based distance kernels on CTK 12 (#1702) @ahendriksen
• Update bench-ann configuration (#1696) @lowener
• Update to Cython 3.0.0 (#1688) @vyasr
• Update CMake version (#1677) @vyasr
• Branch 23.10 merge 23.08 (#1672) @vyasr
• ANN-benchmarks: switch to use gbench (#1661) @achirkin

raft 23.08.00 (9 Aug 2023)

🚨 Breaking Changes

• Separate CAGRA index type from internal idx type (#1664) @tfeher
• Stop using setup.py in build.sh (#1645) @vyasr
• CAGRA max_queries auto configuration (#1613) @enp1s0
• Rename the CAGRA prune function to optimize (#1588) @enp1s0
• CAGRA pad dataset for 128bit vectorized load (#1505) @tfeher
• Sparse Pairwise Distances API Updates (#1502) @divyegala
• Cagra index construction without copying device mdarrays (#1494) @tfeher
• [FEA] Masked NN for connect_components (#1445) @tarang-jain
• Limiting workspace memory resource (#1356) @achirkin

🐛 Bug Fixes

• Remove push condition on docs-build (#1693) @raydouglass
• IVF-PQ: Fix illegal memory access with large max_samples (#1685) @achirkin
• Fix missing parameter for select_k (#1682) @ucassjy
• Separate CAGRA index type from internal idx type (#1664) @tfeher
• Add rmm to pylibraft run dependencies, since it is used by Cython. (#1656) @bdice
• Hotfix: wrong constant in IVF-PQ fp_8bit2half (#1654) @achirkin
• Fix sparse KNN for large batches (#1640) @viclafargue
• Fix uploading of RAFT nightly packages (#1638) @dantegd
• Fix cagra multi CTA bug (#1628) @enp1s0
• pass correct stream to cutlass kernel launch of L2/cosine pairwise distance kernels (#1597) @mdoijade
• Fix launchconfig y-gridsize too large in epilogue kernel (#1586) @mfoerste4
• Fix update version and pinnings for 23.08. (#1556) @bdice
• Fix for function exposing KNN merge (#1418) @viclafargue

📖 Documentation

• Critical doc fixes and updates for 23.08 (#1705) @cjnolet
• Fix the documentation about changing the logging level (#1596) @enp1s0
• Fix raft::bitonic_sort small usage example (#1580) @enp1s0

🚀 New Features

• Use rapids-cmake new parallel testing feature (#1623) @robertmaynard
• Add support for row-major slice (#1591) @lowener
• IVF-PQ tutorial notebook (#1544) @achirkin
• [FEA] Masked NN for connect_components (#1445) @tarang-jain
• raft: Build CUDA 12 packages (#1388) @vyasr
• Limiting workspace memory resource (#1356) @achirkin

🛠️ Improvements

• Pin dask and distributed for 23.08 release (#1711) @galipremsagar
• Add algo parameter for CAGRA ANN bench (#1687) @tfeher
• ANN benchmarks python wrapper for splitting billion-scale dataset groundtruth (#1679) @divyegala
• Rename CAGRA parameter num_parents to search_width (#1676) @tfeher
• Renaming namespaces to promote CAGRA from experimental (#1666) @cjnolet
• CAGRA Python wrappers (#1665) @dantegd
• Add notebook for Vector Search - Question Retrieval (#1662) @lowener
• Fix CMake CUDA support for pylibraft when raft is found. (#1659) @bdice
• Cagra ANN benchmark improvements (#1658) @tfeher
• ANN-benchmarks: avoid using the dataset during search when possible (#1657) @achirkin
• Revert CUDA 12.0 CI workflows to branch-23.08. (#1652) @bdice
• ANN: Optimize host-side refine (#1651) @achirkin
• Cagra template instantiations (#1650) @tfeher
• Modify comm_split to avoid ucp (#1649) @ChuckHastings
• Stop using setup.py in build.sh (#1645) @vyasr
• IVF-PQ: Add a (faster) direct conversion fp8->half (#1644) @achirkin
• Simplify bench/ann scripts to Python based module (#1642) @divyegala
• Further removal of uses-setup-env-vars (#1639) @dantegd
• Drop blank line in raft-dask/meta.yaml (#1637) @jakirkham
• Enable conservative memory allocations for RAFT IVF-Flat benchmarks. (#1634) @tfeher
• [FEA] Codepacking for IVF-flat (#1632) @tarang-jain
• Fixing ann bench cmake (and docs) (#1630) @cjnolet
• [WIP] Test CI issues (#1626) @VibhuJawa
• Set pool memory resource for raft IVF ANN benchmarks (#1625) @tfeher
• Adding sort option to matrix::select_k api (#1615) @cjnolet
• CAGRA max_queries auto configuration (#1613) @enp1s0
• Use exceptions instead of exit(-1) (#1594) @benfred
• [REVIEW] Add scheduler_file argument to support MNMG setup (#1593) @VibhuJawa
• Rename the CAGRA prune function to optimize (#1588) @enp1s0
• This PR adds support to __half and nb_bfloat16 to myAtomicReduce (#1585) @Kh4ster
• [IMP] move core CUDA RT macros to cuda_rt_essentials.hpp (#1584) @MatthiasKohl
• preprocessor syntax fix (#1582) @AyodeAwe
• use rapids-upload-docs script (#1578) @AyodeAwe
• Unpin dask and distributed for development and fix merge_labels test (#1574) @galipremsagar
• Remove documentation build scripts for Jenkins (#1570) @ajschmidt8
• Add support to __half and nv_bfloat16 to most math functions (#1554) @Kh4ster
• Add RAFT ANN benchmark for CAGRA (#1552) @enp1s0
• Update CAGRA knn_graph_sort to use Raft::bitonic_sort (#1550) @enp1s0
• Add identity matrix function (#1548) @lowener
• Unpin scikit-build upper bound (#1547) @vyasr
• Migrate wheel workflow scripts locally (#1546) @divyegala
• Add sample filtering for ivf_flat. Filtering code refactoring and cleanup (#1541) @alexanderguzhva
• CAGRA pad dataset for 128bit vectorized load (#1505) @tfeher
• Sparse Pairwise Distances API Updates (#1502) @divyegala
• Add CAGRA gbench (#1496) @tfeher
• Cagra index construction without copying device mdarrays (#1494) @tfeher

raft 23.06.00 (7 Jun 2023)

🚨 Breaking Changes

• ivf-pq::search: fix the indexing type of the query-related mdspan arguments (#1539) @achirkin
• Dropping Python 3.8 (#1454) @divyegala

🐛 Bug Fixes

• [HOTFIX] Fix distance metrics L2/cosine/correlation when X & Y are same buffer but with different shape and add unit test for such case. (#1571) @mdoijade
• Using raft::resources in rsvd (#1543) @cjnolet
• ivf-pq::search: fix the indexing type of the query-related mdspan arguments (#1539) @achirkin
• Check python brute-force knn inputs (#1537) @benfred
• Fix failing TiledKNNTest unittest (#1533) @benfred
• ivf-flat: fix incorrect recomputed size of the index (#1525) @achirkin
• ivf-flat: limit the workspace size of the search via batching (#1515) @achirkin
• Support uint64_t in CAGRA index data type (#1514) @enp1s0
• Workaround for cuda 12 issue in cusparse (#1508) @cjnolet
• Un-scale output distances (#1499) @achirkin
• Inline get_cache_idx (#1492) @ahendriksen
• Pin to scikit-build<17.2 (#1487) @vyasr
• Remove pool_size() calls from debug printouts (#1484) @tfeher
• Add missing ext declaration for log detail::format (#1482) @tfeher
• Remove include statements from inside namespace (#1467) @robertmaynard
• Use pin_compatible to ensure that lower CTKs can be used (#1462) @vyasr
• fix ivf_pq n_probes (#1456) @benfred
• The glog project root CMakeLists.txt is where we should build from (#1442) @robertmaynard
• Add missing resource factory virtual destructor (#1433) @cjnolet
• Removing cuda stream view include from mdarray (#1429) @cjnolet
• Fix dim param for IVF-PQ wrapper in ANN bench (#1427) @tfeher
• Remove MetricProcessor code from brute_force::knn (#1426) @benfred
• Fix is_min_close (#1419) @benfred
• Have consistent compile lines between BUILD_TESTS enabled or not (#1401) @robertmaynard
• Fix ucx-py pin in raft-dask recipe (#1396) @vyasr

📖 Documentation

• Various updates to the docs for 23.06 release (#1538) @cjnolet
• Rename kernel arch finding function for dispatch (#1536) @mdoijade
• Adding bfknn and ivf-pq python api to docs (#1507) @cjnolet
• Add RAPIDS cuDF as a library that supports cuda_array_interface (#1444) @miguelusque

🚀 New Features

• IVF-PQ: manipulating individual lists (#1298) @achirkin
• Gram matrix support for sparse input (#1296) @mfoerste4
• [FEA] Add randomized svd from cusolver (#1000) @lowener

🛠️ Improvements

• Require Numba 0.57.0+ (#1559) @jakirkham
• remove device_resources include from linalg::map (#1540) @benfred
• Learn heuristic to pick fastest select_k algorithm (#1523) @benfred
• [REVIEW] make raft::cache::Cache protected to allow overrides (#1522) @mfoerste4
• [REVIEW] Fix padding assertion in sparse Gram evaluation (#1521) @mfoerste4
• run docs nightly too (#1520) @AyodeAwe
• Switch back to using primary shared-action-workflows branch (#1519) @vyasr
• Python API for IVF-Flat serialization (#1516) @tfeher
• Introduce sample filtering to IVFPQ index search (#1513) @alexanderguzhva
• Migrate from raft::device_resources -> raft::resources (#1510) @benfred
• Use rmm allocator in CAGRA prune (#1503) @enp1s0
• Update recipes to GTest version >=1.13.0 (#1501) @bdice
• Remove raft/matrix/matrix.cuh includes (#1498) @benfred
• Generate dataset of select_k times (#1497) @benfred
• Re-use memory pool between benchmark runs (#1495) @benfred
• Support CUDA 12.0 for pip wheels (#1489) @divyegala
• Update cupy dependency (#1488) @vyasr
• Enable sccache hits from local builds (#1478) @AyodeAwe
• Build wheels using new single image workflow (#1477) @vyasr
• Revert shared-action-workflows pin (#1475) @divyegala
• CAGRA: Separate graph index sorting functionality from prune function (#1471) @enp1s0
• Add generic reduction functions and separate reductions/warp_primitives (#1470) @akifcorduk
• [ENH] [FINAL] Header structure: combine all PRs into one (#1469) @ahendriksen
• use matrix::select_k in brute_force::knn call (#1463) @benfred
• Dropping Python 3.8 (#1454) @divyegala
• Fix linalg::map to work with non-power-of-2-sized types again (#1453) @ahendriksen
• [ENH] Enable building with clang (limit strict error checking to GCC) (#1452) @ahendriksen
• Remove usage of rapids-get-rapids-version-from-git (#1436) @jjacobelli
• Minor Updates to Sparse Structures (#1432) @divyegala
• Use nvtx3 includes. (#1431) @bdice
• Remove wheel pytest verbosity (#1424) @sevagh
• Add python bindings for matrix::select_k (#1422) @benfred
• Using raft::resources across raft::random (#1420) @cjnolet
• Generate build metrics report for test and benchmarks (#1414) @divyegala
• Update clang-format to 16.0.1. (#1412) @bdice
• Use ARC V2 self-hosted runners for GPU jobs (#1410) @jjacobelli
• Remove uses-setup-env-vars (#1406) @vyasr
• Resolve conflicts in auto-merger of branch-23.06 and branch-23.04 (#1403) @galipremsagar
• Adding base header-only conda package without cuda math libs (#1386) @cjnolet
• Fix IVF-PQ API to use device_vector_view (#1384) @lowener
• Branch 23.06 merge 23.04 (#1379) @vyasr
• Forward merge branch 23.04 into 23.06 (#1350) @cjnolet
• Fused L2 1-NN based on cutlass 3xTF32 / DMMA (#1118) @mdoijade

raft 23.04.00 (6 Apr 2023)

🚨 Breaking Changes

• Pin dask and distributed for release (#1399) @galipremsagar
• Remove faiss_mr.hpp (#1351) @benfred
• Removing FAISS from build (#1340) @cjnolet
• Generic linalg::map (#1337) @achirkin
• Consolidate pre-compiled specializations into single libraft binary (#1333) @cjnolet
• Generic linalg::map (#1329) @achirkin
• Update and standardize IVF indexes API (#1328) @viclafargue
• IVF-Flat index splitting (#1271) @lowener
• IVF-PQ: store cluster data in individual lists and reduce templates (#1249) @achirkin
• Fix for svd API (#1190) @lowener
• Remove deprecated headers (#1145) @lowener

🐛 Bug Fixes

• Fix primitives benchmarks (#1389) @ahendriksen
• Fixing index-url link on pip install docs (#1378) @cjnolet
• Adding some functions back in that seem to be a copy/paste error (#1373) @cjnolet
• Remove usage of Dask's get_worker (#1365) @pentschev
• Remove MANIFEST.in use auto-generated one for sdists and package_data for wheels (#1348) @vyasr
• Revert "Generic linalg::map (#1329)" (#1336" (#1336)) @cjnolet
• Small follow-up to specializations cleanup (#1332) @cjnolet
• Fixing select_k specializations (#1330) @cjnolet
• Fixing remaining bug in ann_quantized (#1327) @cjnolet
• Fixign a couple small kmeans bugs (#1274) @cjnolet
• Remove no longer instantiated templates from list of extern template declarations (#1272) @vyasr
• Bump pinned deps to 23.4 (#1266) @vyasr
• Fix the destruction of interruptible token registry (#1229) @achirkin
• Expose raft::handle_t in the public header (#1192) @vyasr
• Fix for svd API (#1190) @lowener

📖 Documentation

• Adding architecture diagram to README.md (#1370) @cjnolet
• Adding small readme image (#1354) @cjnolet
• Fix serialize documentation of ivf_flat (#1347) @lowener
• Small updates to docs (#1339) @cjnolet

🚀 New Features

• Add Options to Generate Build Metrics Report (#1369) @divyegala
• Generic linalg::map (#1337) @achirkin
• Generic linalg::map (#1329) @achirkin
• matrix::select_k specializations (#1268) @achirkin
• Use rapids-cmake new COMPONENT exporting feature (#1154) @robertmaynard

🛠️ Improvements

• Pin dask and distributed for release (#1399) @galipremsagar
• Pin cupy in wheel tests to supported versions (#1383) @vyasr
• CAGRA (#1375) @tfeher
• add a distance epilogue function to the bfknn call (#1371) @benfred
• Relax UCX pin to allow 1.14 (#1366) @pentschev
• Generate pyproject dependencies with dfg (#1364) @vyasr
• Add nccl to dependencies.yaml (#1361) @benfred
• Add extern template for ivfflat_interleaved_scan (#1360) @ahendriksen
• Stop setting package version attribute in wheels (#1359) @vyasr
• Fix ivf flat specialization header IdxT from uint64_t -> int64_t (#1358) @ahendriksen
• Remove faiss_mr.hpp (#1351) @benfred
• Rename optional helper function (#1345) @viclafargue
• Pass minimum target compile options through raft::raft (#1341) @cjnolet
• Removing FAISS from build (#1340) @cjnolet
• Add dispatch based on compute architecture (#1335) @ahendriksen
• Consolidate pre-compiled specializations into single libraft binary (#1333) @cjnolet
• Update and standardize IVF indexes API (#1328) @viclafargue
• Using int64_t specializations for ivf_pq and refine (#1325) @cjnolet
• Migrate as much as possible to pyproject.toml (#1324) @vyasr
• Pass AWS_SESSION_TOKEN and SCCACHE_S3_USE_SSL vars to conda build (#1321) @ajschmidt8
• Numerical stability fixes for l2 pairwise distance (#1319) @benfred
• Consolidate linter configuration into pyproject.toml (#1317) @vyasr
• IVF-Flat Python wrappers (#1316) @tfeher
• Add stream overloads to ivf_pq serialize/deserialize methods (#1315) @divyegala
• Temporary buffer to view host or device memory in device (#1313) @divyegala
• RAFT skeleton project template (#1312) @cjnolet
• Fix docs build to be pydata-sphinx-theme=0.13.0 compatible (#1311) @galipremsagar
• Update to GCC 11 (#1309) @bdice
• Reduce compile times of distance specializations (#1307) @ahendriksen
• Fix docs upload path (#1305) @AyodeAwe
• Add end-to-end CUDA ann-benchmarks to raft (#1304) @cjnolet
• Make docs builds less verbose (#1302) @AyodeAwe
• Stop using versioneer to manage versions (#1301) @vyasr
• Adding util to get the device id for a pointer address (#1297) @cjnolet
• Enable dfg in pre-commit. (#1293) @vyasr
• Python API for brute-force KNN (#1292) @cjnolet
• support k up to 2048 in faiss select (#1287) @benfred
• CI: Remove specification of manual stage for check_style.sh script. (#1283) @csadorf
• New Sparse Matrix APIs (#1279) @cjnolet
• fix build on cuda 11.5 (#1277) @benfred
• IVF-Flat index splitting (#1271) @lowener
• Remove duplicate librmm runtime dependency (#1264) @ajschmidt8
• build.sh: Add option to log nvcc compile times (#1262) @ahendriksen
• Reduce error handling verbosity in CI tests scripts (#1259) @AjayThorve
• Update shared workflow branches (#1256) @ajschmidt8
• Keeping only compute similarity specializations for uint64_t for now (#1255) @cjnolet
• Fix compile time explosion for minkowski distance (#1254) @ahendriksen
• Unpin dask and distributed for development (#1253) @galipremsagar
• Remove gpuCI scripts. (#1252) @bdice
• IVF-PQ: store cluster data in individual lists and reduce templates (#1249) @achirkin
• Fix inconsistency between the building doc and CMakeLists.txt (#1248) @yong-wang
• Consolidating ANN benchmarks and tests (#1243) @cjnolet
• mdspan view for IVF-PQ API (#1236) @viclafargue
• Remove uint32 distance idx specializations (#1235) @cjnolet
• Add innerproduct to the pairwise distance api (#1226) @benfred
• Move date to build string in conda recipe (#1223) @ajschmidt8
• Replace faiss bfKnn (#1202) @benfred
• Expose KMeans init_plus_plus in pylibraft (#1198) @betatim
• Fix ucx-py version (#1184) @ajschmidt8
• Improve the performance of radix top-k (#1175) @yong-wang
• Add docs build job (#1168) @AyodeAwe
• Remove deprecated headers (#1145) @lowener
• Simplify distance/detail to make is easier to dispatch to different kernel implementations (#1142) @ahendriksen
• Initial port of auto-find-k (#1070) @cjnolet

raft 23.02.00 (9 Feb 2023)

🚨 Breaking Changes

• Remove faiss ANN code from knnIndex (#1121) @benfred
• Use GenPC (Permuted Congruential) as the default random number generator everywhere (#1099) @Nyrio

🐛 Bug Fixes

• Reverting a few commits from 23.02 and speeding up end-to-end build time (#1232) @cjnolet
• Update README.md: fix a missing word (#1185) @achirkin
• balanced-k-means: fix a too large initial memory pool size (#1148) @achirkin
• Catch signal handler change error (#1147) @tfeher
• Squared norm fix follow-up (change was lost in merge conflict) (#1144) @Nyrio
• IVF-Flat bug fix: the squared norm is required for expanded distance calculations (#1141) @Nyrio
• build.sh switch to use RAPIDS magic value (#1132) @robertmaynard
• Fix euclidean_dist in IVF-Flat search (#1122) @Nyrio
• Update handle docstring (#1103) @dantegd
• Pin libcusparse and libcusolver to avoid CUDA 12 (#1095) @wphicks
• Fix race condition in raft::random::discrete (#1094) @Nyrio
• Fixing libraft conda recipes (#1084) @cjnolet
• Ensure that we get the cuda version of faiss. (#1078) @vyasr
• Fix double definition error in ANN refinement header (#1067) @tfeher
• Specify correct global targets names to raft_export (#1054) @robertmaynard
• Fix concurrency issues in k-means++ initialization (#1048) @Nyrio

📖 Documentation

• Adding small comms tutorial to docs (#1204) @cjnolet
• Separating more namespaces into easier-to-consume sections (#1091) @cjnolet
• Paying down some tech debt on docs, runtime API, and cython (#1055) @cjnolet

🚀 New Features

• Add function to convert mdspan to a const view (#1188) @lowener
• Internal library to share headers between test and bench (#1162) @achirkin
• Add public API and tests for hierarchical balanced k-means (#1113) @Nyrio
• Export NCCL dependency as part of raft::distributed. (#1077) @vyasr
• Serialization of IVF Flat and IVF PQ (#919) @tfeher

🛠️ Improvements

• Pin dask and distributed for release (#1242) @galipremsagar
• Update shared workflow branches (#1241) @ajschmidt8
• Removing interruptible from basic handle sync. (#1224) @cjnolet
• pre-commit: Update isort version to 5.12.0 (#1215) @wence-
• Pin wheel dependencies to same RAPIDS release (#1200) @sevagh
• Serializer for mdspans (#1173) @hcho3
• Use CTK 118/cp310 branch of wheel workflows (#1169) @sevagh
• Enable shallow copy of handle_t's resources with different workspace_resource (#1165) @cjnolet
• Protect balanced k-means out-of-memory in some cases (#1161) @achirkin
• Use squeuclidean for metric name in ivf_pq python bindings (#1160) @benfred
• ANN tests: make the min_recall check strict (#1156) @achirkin
• Make cutlass use static ctk (#1155) @sevagh
• Fix various build errors (#1152) @hcho3
• Remove faiss bfKnn call from fused_l2_knn unittest (#1150) @benfred
• Fix unary_op docs and add map_offset as an improved version of write_only_unary_op (#1149) @Nyrio
• Improvement of the math API wrappers (#1146) @Nyrio
• Changing handle_t to device_resources everywhere (#1140) @cjnolet
• Add L2SqrtExpanded support to ivf_pq (#1138) @benfred
• Adding workspace resource (#1137) @cjnolet
• Add raft::void_op functor (#1136) @ahendriksen
• IVF-PQ: tighten the test criteria (#1135) @achirkin
• Fix documentation author (#1134) @bdice
• Add L2SqrtExpanded support to ivf_flat ANN indices (#1133) @benfred
• Improvements in matrix::gather: test coverage, compilation errors, performance (#1126) @Nyrio
• Adding ability to use an existing stream in the pylibraft Handle (#1125) @cjnolet
• Remove faiss ANN code from knnIndex (#1121) @benfred
• Update builds for CUDA 11.8 and Python 3.10 (#1120) @ajschmidt8
• Update workflows for nightly tests (#1119) @ajschmidt8
• Enable Recently Updated Check (#1117) @ajschmidt8
• Build wheels alongside conda CI (#1116) @sevagh
• Allow host dataset for IVF-PQ (#1114) @tfeher
• Decoupling raft handle from underlying resources (#1111) @cjnolet
• Fixing an index error introduced in PR #1109 (#1110) @vinaydes
• Fixing the sample-without-replacement test failures (#1109) @vinaydes
• Remove faiss dependency from fused_l2_knn.cuh, selection_faiss.cuh, ball_cover.cuh and haversine_distance.cuh (#1108) @benfred
• Remove redundant operators in sparse/distance and move others to raft/core (#1105) @Nyrio
• Speedup make_blobs by up to 2x by fixing inefficient kernel launch configuration (#1100) @Nyrio
• Use GenPC (Permuted Congruential) as the default random number generator everywhere (#1099) @Nyrio
• Cleanup faiss includes (#1098) @benfred
• matrix::select_k: move selection and warp-sort primitives (#1085) @achirkin
• Exclude changelog from pre-commit spellcheck (#1083) @benfred
• Add GitHub Actions Workflows. (#1076) @bdice
• Adding uninstall option to build.sh (#1075) @cjnolet
• Use doctest for testing python example docstrings (#1073) @benfred
• Minor cython fixes / cleanup (#1072) @benfred
• IVF-PQ: tweak launch configuration (#1069) @achirkin
• Unpin dask and distributed for development (#1068) @galipremsagar
• Bifurcate Dependency Lists (#1065) @ajschmidt8
• Add support for 64bit svdeig (#1060) @lowener
• switch mma instruction shape to 1684 from current 1688 for 3xTF32 L2/cosine kernel (#1057) @mdoijade
• Make IVF-PQ build index in batches when necessary (#1056) @achirkin
• Remove unused setuputils modules (#1053) @vyasr
• Branch 23.02 merge 22.12 (#1051) @benfred
• Shared-memory-cached kernel for reduce_cols_by_key to limit atomic conflicts (#1050) @Nyrio
• Unify use of common functors (#1049) @Nyrio
• Replace k-means++ CPU bottleneck with a random::discrete prim (#1039) @Nyrio
• Add python bindings for kmeans fit (#1016) @benfred
• Add MaskedL2NN (#838) @ahendriksen
• Move contractions tiling logic outside of Contractions_NT (#837) @ahendriksen

raft 22.12.00 (8 Dec 2022)

🚨 Breaking Changes

• Make ucx linkage explicit and add a new CMake target for it (#1032) @vyasr
• IVF-Flat: make adaptive-centers behavior optional (#1019) @achirkin
• Remove make_mdspan template for memory_type enum (#1005) @wphicks
• ivf-pq performance tweaks (#926) @achirkin

🐛 Bug Fixes

• fusedL2NN: Add input alignment checks (#1045) @achirkin
• Fix fusedL2NN bug that can happen when the same point appears in both x and y (#1040) @Nyrio
• Fix trivial deprecated header includes (#1034) @achirkin
• Suppress ptxas stack size warning in Debug mode (#1033) @tfeher
• Don't use CMake 3.25.0 as it has a FindCUDAToolkit show stopping bug (#1029) @robertmaynard
• Fix for gemmi deprecation (#1020) @lowener
• Remove make_mdspan template for memory_type enum (#1005) @wphicks
• Add except + to cython extern cdef declarations (#1001) @benfred
• Changing Overloads for GCC 11/12 bug (#995) @divyegala
• Changing Overloads for GCC 11/12 bugs (#992) @divyegala
• Fix pylibraft docstring example code (#980) @benfred
• Update raft tests to compile with C++17 features enabled (#973) @robertmaynard
• Making ivf flat gtest invoke mdspanified APIs (#955) @cjnolet
• Updates to kmeans public API to fix cuml (#932) @cjnolet
• Fix logger (vsnprintf consumes args) (#917) @Nyrio
• Adding missing include for device mdspan in mean_squared_error.cuh (#906) @cjnolet

📖 Documentation

• Add links to the docs site in the README (#1042) @benfred
• Moving contributing and developer guides to main docs (#1006) @cjnolet
• Update compiler flags in build docs (#999) @cjnolet
• Updating minimum required gcc version (#993) @cjnolet
• important doc updates for core, cluster, and neighbors (#933) @cjnolet

🚀 New Features

• ANN refinement Python wrapper (#1052) @tfeher
• Add ANN refinement method (#1038) @tfeher
• IVF-Flat: make adaptive-centers behavior optional (#1019) @achirkin
• Add wheel builds (#1013) @vyasr
• Update cuSparse wrappers to avoid deprecated functions (#989) @wphicks
• Provide memory_type enum (#984) @wphicks
• Add Tests for kmeans API (#982) @lowener
• mdspanifying weighted_mean and add raft::stats tests (#910) @lowener
• Implement raft::stats API with mdspan (#802) @lowener

🛠️ Improvements

• Pin dask and distributed for release (#1062) @galipremsagar
• IVF-PQ: use device properties helper (#1035) @achirkin
• Make ucx linkage explicit and add a new CMake target for it (#1032) @vyasr
• Fixing broken doc functions and improving coverage (#1030) @cjnolet
• Expose cluster_cost to python (#1028) @benfred
• Adding lightweight cai_wrapper to reduce boilerplate (#1027) @cjnolet
• Change raft docs theme to pydata-sphinx-theme (#1026) @galipremsagar
• Revert " Pin dask and distributed for release" (#1023) @galipremsagar
• Pin dask and distributed for release (#1022) @galipremsagar
• Replace dots_along_rows with rowNorm and improve coalescedReduction performance (#1011) @Nyrio
• Moving TestDeviceBuffer to pylibraft.common.device_ndarray (#1008) @cjnolet
• Add codespell as a linter (#1007) @benfred
• Fix environment channels (#996) @bdice
• Automatically sync handle when not passed to pylibraft functions (#987) @benfred
• Replace normalize_rows in ann_utils.cuh by a new rowNormalize prim and improve performance for thin matrices (small n_cols) (#979) @Nyrio
• Forward merge 22.10 into 22.12 (#978) @vyasr
• Use new rapids-cmake functionality for rpath handling. (#976) @vyasr
• Update cuda-python dependency to 11.7.1 (#975) @galipremsagar
• IVF-PQ Python wrappers (#970) @tfeher
• Remove unnecessary requirements for raft-dask. (#969) @vyasr
• Expose linalg::dot in public API (#968) @benfred
• Fix kmeans cluster templates (#966) @lowener
• Run linters using pre-commit (#965) @benfred
• linewiseop padded span test (#964) @mfoerste4
• Add unittest for linalg::mean_squared_error (#961) @benfred
• Exposing fused l2 knn to public APIs (#959) @cjnolet
• Remove a left over print statement from pylibraft (#958) @betatim
• Switch to using rapids-cmake for gbench. (#954) @vyasr
• Some cleanup of k-means internals (#953) @cjnolet
• Remove stale labeler (#951) @raydouglass
• Adding optional handle to each public API function (along with example) (#947) @cjnolet
• Improving documentation across the board. Adding quick-start to breathe docs. (#943) @cjnolet
• Add unittest for linalg::axpy (#942) @benfred
• Add cutlass 3xTF32,DMMA based L2/cosine distance kernels for SM 8.0 or higher (#939) @mdoijade
• Calculate max cluster size correctly for IVF-PQ (#938) @tfeher
• Add tests for raft::matrix (#937) @lowener
• Add fusedL2NN benchmark (#936) @Nyrio
• ivf-pq performance tweaks (#926) @achirkin
• Adding fused_l2_nn_argmin wrapper to Pylibraft (#924) @cjnolet
• Moving kernel gramm primitives to raft::distance::kernels (#920) @cjnolet
• kmeans improvements: random initialization on GPU, NVTX markers, no batching when using fusedL2NN (#918) @Nyrio
• Moving raft::spatial::knn -> raft::neighbors (#914) @cjnolet
• Create cub-based argmin primitive and replace argmin_along_rows in ANN kmeans (#912) @Nyrio
• Replace map_along_rows with matrixVectorOp (#911) @Nyrio
• Integrate accumulate_into_selected from ANN utils into linalg::reduce_rows_by_keys (#909) @Nyrio
• Re-enabling Fused L2 NN specializations and renaming cub::KeyValuePair -> raft::KeyValuePair (#905) @cjnolet
• Unpin dask and distributed for development (#886) @galipremsagar
• Adding padded layout 'layout_padded_general' (#725) @mfoerste4

raft 22.10.00 (12 Oct 2022)

🚨 Breaking Changes

• Separating mdspan/mdarray infra into host_* and device_* variants (#810) @cjnolet
• Remove type punning from TxN_t (#781) @wphicks
• ivf_flat::index: hide implementation details (#747) @achirkin

🐛 Bug Fixes

• ivf-pq integration: hotfixes (#891) @achirkin
• Removing cub symbol from libraft-distance instantiation. (#887) @cjnolet
• ivf-pq post integration hotfixes (#878) @achirkin
• Fixing a few compile errors in new APIs (#874) @cjnolet
• Include knn.cuh in knn.cu benchmark source for finding brute_force_knn (#855) @teju85
• Do not use strcpy to copy 2 char (#848) @mhoemmen
• rng_state not including necessary cstdint (#839) @MatthiasKohl
• Fix integer overflow in ANN kmeans (#835) @Nyrio
• Add alignment to the TxN_t vectorized type (#792) @achirkin
• Fix adj_to_csr_kernel (#785) @ahendriksen
• Use rapids-cmake 22.10 best practice for RAPIDS.cmake location (#784) @robertmaynard
• Remove type punning from TxN_t (#781) @wphicks
• Various fixes for build.sh (#771) @vyasr

📖 Documentation

• Fix target names in build.sh help text (#879) @Nyrio
• Document that minimum required CMake version is now 3.23.1 (#841) @robertmaynard

🚀 New Features

• mdspanify raft::random functions uniformInt, normalTable, fill, bernoulli, and scaled_bernoulli (#897) @mhoemmen
• mdspan-ify several raft::random rng functions (#857) @mhoemmen
• Develop new mdspan-ified multi_variable_gaussian interface (#845) @mhoemmen
• Mdspanify permute (#834) @mhoemmen
• mdspan-ify rmat_rectangular_gen (#833) @mhoemmen
• mdspanify sampleWithoutReplacement (#830) @mhoemmen
• mdspan-ify make_regression (#811) @mhoemmen
• Updating raft::linalg APIs to use mdspan (#809) @divyegala
• Integrate KNN implementation: ivf-pq (#789) @achirkin

🛠️ Improvements

• Some fixes for build.sh (#901) @cjnolet
• Revert recent fused l2 nn instantiations (#899) @cjnolet
• Update Python build instructions (#898) @betatim
• Adding ninja and cxx compilers to conda dev dependencies (#893) @cjnolet
• Output non-normalized distances in IVF-PQ and brute-force KNN (#892) @Nyrio
• Readme updates for 22.10 (#884) @cjnolet
• Breaking apart benchmarks into individual binaries (#883) @cjnolet
• Pin dask and distributed for release (#858) @galipremsagar
• Mdspanifying (currently tested) raft::matrix (#846) @cjnolet
• Separating _RAFT_HOST and _RAFT_DEVICE macros (#836) @cjnolet
• Updating cpu job in hopes it speeds up python cpu builds (#828) @cjnolet
• Mdspan-ifying raft::spatial (#827) @cjnolet
• Fixing init.py for handle and stream (#826) @cjnolet
• Moving a few more things around (#822) @cjnolet
• Use fusedL2NN in ANN kmeans (#821) @Nyrio
• Separating test executables (#820) @cjnolet
• Separating mdspan/mdarray infra into host_* and device_* variants (#810) @cjnolet
• Fix malloc/delete mismatch (#808) @mhoemmen
• Renaming pyraft -> raft-dask (#801) @cjnolet
• Branch 22.10 merge 22.08 (#800) @cjnolet
• Statically link all CUDA toolkit libraries (#797) @trxcllnt
• Minor follow-up fixes for ivf-flat (#796) @achirkin
• KMeans benchmarks (cuML + ANN implementations) and fix for IndexT=int64_t (#795) @Nyrio
• Optimize fusedL2NN when data is skinny (#794) @ahendriksen
• Complete the deprecation of duplicated hpp headers (#793) @ahendriksen
• Prepare parts of the balanced kmeans for ivf-pq (#788) @achirkin
• Unpin dask and distributed for development (#783) @galipremsagar
• Exposing python wrapper for the RMAT generator logic (#778) @teju85
• Device, Host, Managed Accessor Types for mdspan (#776) @divyegala
• Fix Forward-Merger Conflicts (#768) @ajschmidt8
• Fea 2208 kmeans use specializations (#760) @cjnolet
• ivf_flat::index: hide implementation details (#747) @achirkin

raft 22.08.00 (17 Aug 2022)

🚨 Breaking Changes

• Update mdspan to account for changes to extents (#751) @divyegala
• Replace csr_adj_graph functions with faster equivalent (#746) @ahendriksen
• Integrate KNN implementation: ivf-flat (#652) @achirkin
• Moving kmeans from cuml to Raft (#605) @lowener

🐛 Bug Fixes

• Relax ivf-flat test recall thresholds (#766) @achirkin
• Restrict the use of ] to CXX 20 only. (#764) @trivialfis
• Update rapids-cmake version for pyraft in update-version.sh (#749) @vyasr

📖 Documentation

• Use documented header template for doxygen (#773) @galipremsagar
• Switch language from None to "en" in docs build (#721) @galipremsagar

🚀 New Features

• Update mdspan to account for changes to extents (#751) @divyegala
• Implement matrix transpose with mdspan. (#739) @trivialfis
• Implement unravel_index for row-major array. (#723) @trivialfis
• Integrate KNN implementation: ivf-flat (#652) @achirkin

🛠️ Improvements

• Use common js and css code (#779) @galipremsagar
• Pin dask & distributed for release (#772) @galipremsagar
• Move cmake to the build section. (#763) @vyasr
• Adding old kmeans impl back in (as kmeans_deprecated) (#761) @cjnolet
• Fix for KMeans raw pointers API (#758) @lowener
• Fix KMeans (#756) @divyegala
• Add inline to nccl_sync_stream() (#750) @seunghwak
• Replace csr_adj_graph functions with faster equivalent (#746) @ahendriksen
• Add wrapper functions for ncclGroupStart() and ncclGroupEnd() (#742) @seunghwak
• Fix variadic template type check for mdarrays (#741) @hlinsen
• RMAT rectangular graph generator (#738) @teju85
• Update conda recipes to UCX 1.13.0 (#736) @pentschev
• Add warp-aggregated atomic increment (#735) @ahendriksen
• fix logic bug in include_checker.py utility (#734) @grlee77
• Support 32bit and unsigned indices in bruteforce KNN (#730) @achirkin
• Ability to use ccache to speedup local builds (#729) @teju85
• Pin max version of cuda-python to 11.7.0 (#728) @Ethyling
• Always add raft::raft_nn_lib and raft::raft_distance_lib aliases (#727) @trxcllnt
• Add several type aliases and helpers for creating mdarrays (#726) @achirkin
• fix nans in naive kl divergence kernel introduced by div by 0. (#724) @mdoijade
• Use rapids-cmake for cuco (#722) @vyasr
• Update Python classifiers. (#719) @bdice
• Fix sccache (#718) @Ethyling
• Introducing raft::mdspan as an alias (#715) @divyegala
• Update cuco version (#714) @vyasr
• Update conda environment pinnings and update-versions.sh. (#713) @bdice
• Branch 22.08 merge branch 22.06 (#712) @cjnolet
• Testing conda compilers (#705) @cjnolet
• Unpin dask & distributed for development (#704) @galipremsagar
• Avoid shadowing CMAKE_ARGS variable in build.sh (#701) @vyasr
• Use unique ptr in print_device_vector (#695) @lowener
• Add missing Thrust includes (#678) @bdice
• Consolidate C++ conda recipes and add libraft-tests package (#641) @Ethyling
• Moving kmeans from cuml to Raft (#605) @lowener

raft 22.06.00 (7 Jun 2022)

🚨 Breaking Changes

• Rng: removed cyclic dependency creating hard-to-debug compiler errors (#639) @MatthiasKohl
• Allow enabling NVTX markers by downstream projects after install (#610) @achirkin
• Rng: expose host-rng-state in host-only API (#609) @MatthiasKohl

🐛 Bug Fixes

• For fixing the cuGraph test failures with PCG (#690) @vinaydes
• Fix excessive memory used in selection test (#689) @achirkin
• Revert print vector changes because of std::vector<bool> (#681) @lowener
• fix race in fusedL2knn smem read/write by adding a syncwarp (#679) @mdoijade
• gemm: fix parameter C mistakenly set as const (#664) @achirkin
• Fix SelectionTest: allow different indices when keys are equal. (#659) @achirkin
• Revert recent cmake updates (#657) @cjnolet
• Don't install component dependency files in raft-header only mode (#655) @robertmaynard
• Rng: removed cyclic dependency creating hard-to-debug compiler errors (#639) @MatthiasKohl
• Fixing raft compile bug w/ RNG changes (#634) @cjnolet
• Get libcudacxx from cuco (#632) @trxcllnt
• RNG API fixes (#630) @MatthiasKohl
• Fix mdspan accessor mixin offset policy. (#628) @trivialfis
• Branch 22.06 merge 22.04 (#625) @cjnolet
• fix issue in fusedL2knn which happens when rows are multiple of 256 (#604) @mdoijade

🚀 New Features

• Restore changes from #653 and #655 and correct cmake component dependencies (#686) @robertmaynard
• Adding handle and stream to pylibraft (#683) @cjnolet
• Map CMake install components to conda library packages (#653) @robertmaynard
• Rng: expose host-rng-state in host-only API (#609) @MatthiasKohl
• mdspan/mdarray template functions and utilities (#601) @divyegala

🛠️ Improvements

• Change build.sh to find C++ library by default (#697) @vyasr
• Pin dask and distributed for release (#693) @galipremsagar
• Pin dask & distributed for release (#680) @galipremsagar
• Improve logging (#673) @achirkin
• Fix minor errors in CMake configuration (#662) @vyasr
• Pulling mdspan fork (from official rapids repo) into raft to remove dependency (#649) @cjnolet
• Fixing the unit test issue(s) in RAFT (#646) @vinaydes
• Build pyraft with scikit-build (#644) @vyasr
• Some fixes to pairwise distances for cupy integration (#643) @cjnolet
• Require UCX 1.12.1+ (#638) @jakirkham
• Updating raft rng host public API and adding docs (#636) @cjnolet
• Build pylibraft with scikit-build (#633) @vyasr
• Add cuda_lib_dir to library_dirs, allow changing UCX/RMM/Thrust/spdlog locations via envvars in setup.py (#624) @trxcllnt
• Remove perf prints from MST (#623) @divyegala
• Enable components installation using CMake (#621) @Ethyling
• Allow nullptr as input-indices argument of select_k (#618) @achirkin
• Update CMake pinning to allow newer CMake versions (#617) @vyasr
• Unpin dask & distributed for development (#616) @galipremsagar
• Improve performance of select-top-k RADIX implementation (#615) @achirkin
• Moving more prims benchmarks to RAFT (#613) @cjnolet
• Allow enabling NVTX markers by downstream projects after install (#610) @achirkin
• Improve performance of select-top-k WARP_SORT implementation (#606) @achirkin
• Enable building static libs (#602) @trxcllnt
• Update ucx-py version (#596) @ajschmidt8
• Fix merge conflicts (#587) @ajschmidt8
• Making cuco, thrust, and mdspan optional dependencies. (#585) @cjnolet
• Some RBC3D fixes (#530) @cjnolet

raft 22.04.00 (6 Apr 2022)

🚨 Breaking Changes

• Moving some of the remaining linalg prims from cuml (#502) @cjnolet
• Fix badly merged cublas wrappers (#492) @achirkin
• Hiding implementation details for lap, clustering, spectral, and label (#477) @cjnolet
• Adding destructor for std comms and using nccl allreduce for barrier in mpi comms (#473) @cjnolet
• Cleaning up cusparse_wrappers (#441) @cjnolet
• Improvents to RNG (#434) @vinaydes
• Remove RAFT memory management (#400) @viclafargue
• LinAlg impl in detail (#383) @divyegala

🐛 Bug Fixes

• Pin cmake in conda recipe to <3.23 (#600) @dantegd
• Fix make_device_vector_view (#595) @lowener
• Update cuco version. (#592) @vyasr
• Fixing raft headers dir (#574) @cjnolet
• Update update-version.sh (#560) @raydouglass
• find_package(raft) can now be called multiple times safely (#532) @robertmaynard
• Allocate sufficient memory for Hungarian if number of batches > 1 (#531) @ChuckHastings
• Adding lap.hpp back (with deprecation) (#529) @cjnolet
• raft-config is idempotent no matter RAFT_COMPILE_LIBRARIES value (#516) @robertmaynard
• Call initialize() in mpi_comms_t constructor. (#506) @seunghwak
• Improve row-major meanvar kernel via minimizing atomicCAS locks (#489) @achirkin
• Adding destructor for std comms and using nccl allreduce for barrier in mpi comms (#473) @cjnolet

📖 Documentation

• Updating docs for 22.04 (#566) @cjnolet

🚀 New Features

• Add benchmarks (#549) @achirkin
• Unify weighted mean code (#514) @lowener
• single-pass raft::stats::meanvar (#472) @achirkin
• Move random package of cuML to RAFT (#449) @divyegala
• mdspan integration. (#437) @trivialfis
• Interruptible execution (#433) @achirkin
• make raft sources compilable with clang (#424) @MatthiasKohl
• Span implementation. (#399) @trivialfis

🛠️ Improvements

• Adding build script for docs (#589) @cjnolet
• Temporarily disable new ops-bot functionality (#586) @ajschmidt8
• Fix commands to get conda output files (#584) @Ethyling
• Link to cuco and add faiss EXCLUDE_FROM_ALL option (#583) @trxcllnt
• exposing faiss::faiss (#582) @cjnolet
• Pin dask and distributed version (#581) @galipremsagar
• removing exclude_from_all from cuco (#580) @cjnolet
• Adding INSTALL_EXPORT_SET for cuco, rmm, thrust (#579) @cjnolet
• Thrust package name case (#576) @trxcllnt
• Add missing thrust includes to transpose.cuh (#575) @zbjornson
• Use unanchored clang-format version check (#573) @zbjornson
• Fixing accidental removal of thrust target from cmakelists (#571) @cjnolet
• Don't add gtest to build export set or generate a gtest-config.cmake (#565) @trxcllnt
• Set main label by default (#559) @galipremsagar
• Add local conda channel while looking for conda outputs (#558) @Ethyling
• Updated dask and distributed to >=2022.02.1 (#557) @rlratzel
• Upload packages using testing label for nightlies (#556) @Ethyling
• Add .github/ops-bot.yaml config file (#554) @ajschmidt8
• Disabling benchmarks building by default. (#553) @cjnolet
• KNN select-top-k variants (#551) @achirkin
• Adding logger (#550) @cjnolet
• clang-tidy support: improved clang run scripts with latest changes (see cugraph-ops) (#548) @MatthiasKohl
• Pylibraft for pairwise distances (#540) @cjnolet
• mdspan PoC for distance make_blobs (#538) @cjnolet
• Include thrust/sort.h in ball_cover.cuh (#526) @akifcorduk
• Increase parallelism in allgatherv (#525) @seunghwak
• Moving device functions to cuh files and deprecating hpp (#524) @cjnolet
• Use dynamic_extent from stdex. (#523) @trivialfis
• Updating some of the ci check scripts (#522) @cjnolet
• Use shfl_xor in warpReduce for broadcast (#521) @akifcorduk
• Fixing Python conda package and installation (#520) @cjnolet
• Adding instructions to install from conda and build using CPM (#519) @cjnolet
• Implement span storage optimization. (#515) @trivialfis
• RNG test fixes and improvements (#513) @vinaydes
• Moving scores and metrics over to raft::stats (#512) @cjnolet
• Random ball cover in 3d (#510) @cjnolet
• Initializing memory in RBC (#509) @cjnolet
• Adjusting conda packaging to remove duplicate dependencies (#508) @cjnolet
• Moving remaining stats prims from cuml (#507) @cjnolet
• Correcting the namespace (#505) @vinaydes
• Passing stream through commsplit (#503) @cjnolet
• Moving some of the remaining linalg prims from cuml (#502) @cjnolet
• Fixing spectral APIs (#496) @cjnolet
• Fix badly merged cublas wrappers (#492) @achirkin
• Fix integer overflow in distances (#490) @RAMitchell
• Reusing shared libs in gpu ci builds (#487) @cjnolet
• Adding fatbin to shared libs and fixing conda paths in cpu build (#485) @cjnolet
• Add CMake install rule for tests (#483) @ajschmidt8
• Adding cpu ci for conda build (#482) @cjnolet
• iUpdating codeowners to use new raft codeowners (#480) @cjnolet
• Hiding implementation details for lap, clustering, spectral, and label (#477) @cjnolet
• Define PTDS via -D to fix cache misses in sccache (#476) @trxcllnt
• Unpin dask and distributed (#474) @galipremsagar
• Replace ccache with sccache (#471) @ajschmidt8
• More README updates (#467) @cjnolet
• CUBLAS wrappers with switchable host/device pointer mode (#453) @achirkin
• Cleaning up cusparse_wrappers (#441) @cjnolet
• Adding conda packaging for libraft and pyraft (#439) @cjnolet
• Improvents to RNG (#434) @vinaydes
• Hiding implementation details for comms (#409) @cjnolet
• Remove RAFT memory management (#400) @viclafargue
• LinAlg impl in detail (#383) @divyegala

raft 22.02.00 (2 Feb 2022)

🚨 Breaking Changes

• Simplify raft component CMake logic, and allow compilation without FAISS (#428) @robertmaynard
• One cudaStream_t instance per raft::handle_t (#291) @divyegala

🐛 Bug Fixes

• Removing extra logging from faiss mr (#463) @cjnolet
• Pin dask & distributed versions (#455) @galipremsagar
• Replace RMM CUDA Python bindings with those provided by CUDA-Python (#451) @shwina
• Fix comms memory leak (#436) @seunghwak
• Fix C++ doxygen documentation (#426) @achirkin
• Fix clang-format style errors (#425) @achirkin
• Fix using incorrect macro RAFT_CHECK_CUDA in place of RAFT_CUDA_TRY (#415) @achirkin
• Fix CUDA_CHECK_NO_THROW compatibility define (#414) @zbjornson
• Disabling fused l2 knn from bfknn (#407) @cjnolet
• Disabling expanded fused l2 knn to unblock cuml CI (#404) @cjnolet
• Reverting default knn distance to L2Unexpanded for now. (#403) @cjnolet

📖 Documentation

• README and build fixes before release (#459) @cjnolet
• Updates to Python and C++ Docs (#442) @cjnolet

🚀 New Features

• error macros: determining buffer size instead of fixed 2048 chars (#420) @MatthiasKohl
• NVTX range helpers (#416) @achirkin

🛠️ Improvements

• Splitting fused l2 knn specializations (#461) @cjnolet
• Update cuCollection git tag (#447) @seunghwak
• Remove libcudacxx patch needed for nvcc 11.4 (#446) @robertmaynard
• Unpin dask and distributed (#440) @galipremsagar
• Public apis for remainder of matrix and stats (#438) @divyegala
• Fix bug in producer-consumer buffer exchange which occurs in UMAP test on GV100 (#429) @mdoijade
• Simplify raft component CMake logic, and allow compilation without FAISS (#428) @robertmaynard
• Update ucx-py version on release using rvc (#422) @Ethyling
• Disabling fused l2 knn again. Not sure how this got added back. (#421) @cjnolet
• Adding no throw macro variants (#417) @cjnolet
• Remove IncludeCategories from .clang-format (#412) @codereport
• fix nan issues in L2 expanded sqrt KNN distances (#411) @mdoijade
• Consistent renaming of CHECK_CUDA and *_TRY macros (#410) @cjnolet
• Faster matrix-vector-ops (#401) @achirkin
• Adding dev conda environment files. (#397) @cjnolet
• Update to UCX-Py 0.24 (#392) @pentschev
• Branch 21.12 merge 22.02 (#386) @cjnolet
• Hiding implementation details for sparse API (#381) @cjnolet
• Adding distance specializations (#376) @cjnolet
• Use FAISS with RMM (#363) @viclafargue
• Add Fused L2 Expanded KNN kernel (#339) @mdoijade
• Update .clang-format to be consistent with all other RAPIDS repos (#300) @codereport
• One cudaStream_t instance per raft::handle_t (#291) @divyegala

raft 21.12.00 (9 Dec 2021)

🚨 Breaking Changes

• Use 64 bit CuSolver API for Eigen decomposition (#349) @lowener

🐛 Bug Fixes

• Fixing bad host->device copy (#375) @cjnolet
• Fix coalesced access checks in matrix_vector_op (#372) @achirkin
• Port libcudacxx patch from cudf (#370) @dantegd
• Fixing overflow in expanded distances (#365) @cjnolet

📖 Documentation

• Getting doxygen to run (#371) @cjnolet

🛠️ Improvements

• Upgrade clang to 11.1.0 (#394) @galipremsagar
• Fix Changelog Merge Conflicts for branch-21.12 (#390) @ajschmidt8
• Pin max dask & distributed (#388) @galipremsagar
• Removing conflict w/ CUDA_CHECK (#378) @cjnolet
• Update RAFT test directory (#359) @viclafargue
• Update to UCX-Py 0.23 (#358) @pentschev
• Hiding implementation details for random, stats, and matrix (#356) @divyegala
• README updates (#351) @cjnolet
• Use 64 bit CuSolver API for Eigen decomposition (#349) @lowener
• Hiding implementation details for distance primitives (dense + sparse) (#344) @cjnolet
• Unpin dask & distributed in CI (#338) @galipremsagar

raft 21.10.00 (7 Oct 2021)

🚨 Breaking Changes

• Miscellaneous tech debts/cleanups (#286) @viclafargue

🐛 Bug Fixes

• Accounting for rmm::cuda_stream_pool not having a constructor for 0 streams (#329) @divyegala
• Fix wrong lda parameter in gemv (#327) @achirkin
• Fix matrixVectorOp to verify promoted pointer type is still aligned to vectorized load boundary (#325) @viclafargue
• Pin rmm to branch-21.10 and remove warnings from kmeans.hpp (#322) @dantegd
• Temporarily pin RMM while refactor removes deprecated calls (#315) @dantegd
• Fix more warnings (#311) @harrism

📖 Documentation

• Fix build doc (#316) @lowener

🚀 New Features

• Add Hamming, Jensen-Shannon, KL-Divergence, Russell rao and Correlation distance metrics support (#306) @mdoijade

🛠️ Improvements

• Pin max dask and distributed versions to 2021.09.1 (#334) @galipremsagar
• Make sure we keep the rapids-cmake and raft cal version in sync (#331) @robertmaynard
• Add broadcast with const input iterator (#328) @seunghwak
• Fused L2 (unexpanded) kNN kernel for NN <= 64, without using temporary gmem to store intermediate distances (#324) @mdoijade
• Update with rapids cmake new features (#320) @robertmaynard
• Update to UCX-Py 0.22 (#319) @pentschev
• Fix Forward-Merge Conflicts (#318) @ajschmidt8
• Enable CUDA device code warnings as errors (#307) @harrism
• Remove max version pin for dask & distributed on development branch (#303) @galipremsagar
• Warnings are errors (#299) @harrism
• Use the new RAPIDS.cmake to fetch rapids-cmake (#298) @robertmaynard
• ENH Replace gpuci_conda_retry with gpuci_mamba_retry (#295) @dillon-cullinan
• Miscellaneous tech debts/cleanups (#286) @viclafargue
• Random Ball Cover Algorithm for 2D Haversine/Euclidean (#213) @cjnolet

raft 21.08.00 (4 Aug 2021)

🚨 Breaking Changes

• expose epsilon parameter to allow precision to to be specified (#275) @ChuckHastings

🐛 Bug Fixes

• Fix support for different input and output types in linalg::reduce (#296) @Nyrio
• Const raft handle in sparse bfknn (#280) @cjnolet
• Add cuco::cuco to list of linked libraries (#279) @trxcllnt
• Use nested include in destination of install headers to avoid docker permission issues (#263) @dantegd
• Update UCX-Py version to 0.21 (#255) @pentschev
• Fix mst knn test build failure due to RMM device_buffer change (#253) @mdoijade

🚀 New Features

• Add chebyshev, canberra, minkowksi and hellinger distance metrics (#276) @mdoijade
• Move FAISS ANN wrappers to RAFT (#265) @cjnolet
• Remaining sparse semiring distances (#261) @cjnolet
• removing divye from codeowners (#257) @divyegala

🛠️ Improvements

• Pinning cuco to a specific commit hash for release (#304) @rlratzel
• Pin max dask & distributed versions (#301) @galipremsagar
• Overlap epilog compute with ldg of next grid stride in pairwise distance & fusedL2NN kernels (#292) @mdoijade
• Always add faiss library alias if it's missing (#287) @trxcllnt
• Use NVIDIA/cuCollections repo again (#284) @trxcllnt
• Use the 21.08 branch of rapids-cmake as rmm requires it (#278) @robertmaynard
• expose epsilon parameter to allow precision to to be specified (#275) @ChuckHastings
• Fix 21.08 forward-merge conflicts (#274) @ajschmidt8
• Add lds and sts inline ptx instructions to force vector instruction generation (#273) @mdoijade
• Move ANN to RAFT (additional updates) (#270) @cjnolet
• Sparse semirings cleanup + hash table & batching strategies (#269) @divyegala
• Revert "pin dask versions in CI (#260)" (#264" (#264)) @ajschmidt8
• Pass stream to device_scalar::value() calls. (#259) @harrism
• Update get_rmm.cmake to better support CalVer (#258) @harrism
• Add Grid stride pairwise dist and fused L2 NN kernels (#250) @mdoijade
• Fix merge conflicts (#236) @ajschmidt8

raft 21.06.00 (9 Jun 2021)

🐛 Bug Fixes

• Update UCX-Py version to 0.20 (#254) @pentschev
• cuco git tag update (again) (#248) @seunghwak
• Revert PR #232 for 21.06 release (#246) @dantegd
• Python comms to hold onto server endpoints (#241) @cjnolet
• Fix Thrust 1.12 compile errors (#231) @trxcllnt
• Make sure we use CalVer when checking out rapids-cmake (#230) @robertmaynard
• Loss of Precision in MST weight alteration (#223) @divyegala

🛠️ Improvements

• cuco git tag update (#243) @seunghwak
• Update CHANGELOG.md links for calver (#233) @ajschmidt8
• Add Grid stride pairwise dist and fused L2 NN kernels (#232) @mdoijade
• Updates to enable HDBSCAN (#208) @cjnolet

raft 0.19.0 (21 Apr 2021)

🐛 Bug Fixes

• Exposing spectral random seed property (#193) @cjnolet
• Fix pointer arithmetic in spmv smem kernel (#183) @lowener
• Modify default value for rowMajorIndex and rowMajorQuery in bf-knn (#173) @viclafargue
• Remove setCudaMallocWarning() call for libfaiss[@v1.7.0 (#167) @trxcllnt](https://github.com/v1.7.0 (#167) @trxcllnt)
• Add const to KNN handle (#157) @hlinsen

🚀 New Features

• Moving optimized L2 1-nearest neighbors implementation from cuml (#158) @cjnolet

🛠️ Improvements

• Fixing codeowners (#194) @cjnolet
• Adjust Hellinger pairwise distance to vaoid NaNs (#189) @lowener
• Add column major input support in contractions_nt kernels with new kernel policy for it (#188) @mdoijade
• Dice formula correction (#186) @lowener
• Scaling knn graph fix connectivities algorithm (#181) @cjnolet
• Fixing RAFT CI & a few small updates for SLHC Python wrapper (#178) @cjnolet
• Add Precomputed to the DistanceType enum (for cuML DBSCAN) (#177) @Nyrio
• Enable matrix::copyRows for row major input (#176) @tfeher
• Add Dice distance to distancetype enum (#174) @lowener
• Porting over recent updates to distance prim from cuml (#172) @cjnolet
• Update KNN (#171) @viclafargue
• Adding translations parameter to brute_force_knn (#170) @viclafargue
• Update Changelog Link (#169) @ajschmidt8
• Map operation (#168) @viclafargue
• Updating sparse prims based on recent changes (#166) @cjnolet
• Prepare Changelog for Automation (#164) @ajschmidt8
• Update 0.18 changelog entry (#163) @ajschmidt8
• MST symmetric/non-symmetric output for SLHC (#162) @divyegala
• Pass pre-computed colors to MST (#154) @divyegala
• Streams upgrade in RAFT handle (RMM backend + create handle from parent's pool) (#148) @afender
• Merge branch-0.18 into 0.19 (#146) @dantegd
• Add device_send, device_recv, device_sendrecv, device_multicast_sendrecv (#144) @seunghwak
• Adding SLHC prims. (#140) @cjnolet
• Moving cuml sparse prims to raft (#139) @cjnolet

raft 0.18.0 (24 Feb 2021)

Breaking Changes 🚨

• Make NCCL root initialization configurable. (#120) @drobison00

Bug Fixes 🐛

• Add idx_t template parameter to matrix helper routines (#131) @tfeher
• Eliminate CUDA 10.2 as valid for large svd solving (#129) @wphicks
• Update check to allow svd solver on CUDA>=10.2 (#125) @wphicks
• Updating gpu build.sh and debugging threads CI issue (#123) @dantegd

New Features 🚀

• Adding additional distances (#116) @cjnolet

Improvements 🛠️

• Update stale GHA with exemptions & new labels (#152) @mike-wendt
• Add GHA to mark issues/prs as stale/rotten (#150) @Ethyling
• Prepare Changelog for Automation (#135) @ajschmidt8
• Adding Jensen-Shannon and BrayCurtis to DistanceType for Nearest Neighbors (#132) @lowener
• Add brute force KNN (#126) @hlinsen
• Make NCCL root initialization configurable. (#120) @drobison00
• Auto-label PRs based on their content (#117) @jolorunyomi
• Add gather & gatherv to raft::comms::comms_t (#114) @seunghwak
• Adding canberra and chebyshev to distance types (#99) @cjnolet
• Gpuciscripts clean and update (#92) @msadang

RAFT 0.17.0 (10 Dec 2020)

New Features

• PR #65: Adding cuml prims that break circular dependency between cuml and cumlprims projects
• PR #101: MST core solver
• PR #93: Incorporate Date/Nagi implementation of Hungarian Algorithm
• PR #94: Allow generic reductions for the map then reduce op
• PR #95: Cholesky rank one update prim

Improvements

• PR #108: Remove unused old-gpubuild.sh
• PR #73: Move DistanceType enum from cuML to RAFT
• pr #92: Cleanup gpuCI scripts
• PR #98: Adding InnerProduct to DistanceType
• PR #103: Epsilon parameter for Cholesky rank one update
• PR #100: Add divyegala as codeowner
• PR #111: Cleanup gpuCI scripts
• PR #120: Update NCCL init process to support root node placement.

Bug Fixes

• PR #106: Specify dependency branches to avoid pip resolver failure
• PR #77: Fixing CUB include for CUDA < 11
• PR #86: Missing headers for newly moved prims
• PR #102: Check alignment before binaryOp dispatch
• PR #104: Fix update-version.sh
• PR #109: Fixing Incorrect Deallocation Size and Count Bugs

RAFT 0.16.0 (Date TBD)

New Features

• PR #63: Adding MPI comms implementation
• PR #70: Adding CUB to RAFT cmake

Improvements

• PR #59: Adding csrgemm2 to cusparse_wrappers.h
• PR #61: Add cusparsecsr2dense to cusparse_wrappers.h
• PR #62: Adding get_device_allocator to handle.pxd
• PR #67: Remove dependence on run-time type info

Bug Fixes

• PR #56: Fix compiler warnings.
• PR #64: Remove cublas_try from cusolver_wrappers.h
• PR #66: Fixing typo get_stream to getStream in handle.pyx
• PR #68: Change the type of recvcounts & displs in allgatherv from size_t[] to size_t* and int[] to size_t*, respectively.
• PR #69: Updates for RMM being header only
• PR #74: Fix std_comms::comm_split bug
• PR #79: remove debug print statements
• PR #81: temporarily expose internal NCCL communicator

RAFT 0.15.0 (Date TBD)

New Features

• PR #12: Spectral clustering.
• PR #7: Migrating cuml comms -> raft comms_t
• PR #18: Adding commsplit to cuml communicator
• PR #15: add exception based error handling macros
• PR #29: Add ceildiv functionality
• PR #44: Add get_subcomm and set_subcomm to handle_t

Improvements

• PR #13: Add RMM_INCLUDE and RMM_LIBRARY options to allow linking to non-conda RMM
• PR #22: Preserve order in comms workers for rank initialization
• PR #38: Remove #include <cudar_utils.h> from raft/mr/
• PR #39: Adding a virtual destructor to raft::handle_t and raft::comms::comms_t
• PR #37: Clean-up CUDA related utilities
• PR #41: Upgrade to cusparseSpMV(), alg selection, and rectangular matrices.
• PR #45: Add Ampere target to cuda11 cmake
• PR #47: Use gtest conda package in CMake/build.sh by default

Bug Fixes

• PR #17: Make destructor inline to avoid redeclaration error
• PR #25: Fix bug in handle_t::get_internal_streams
• PR #26: Fix bug in RAFT_EXPECTS (add parentheses surrounding cond)
• PR #34: Fix issue with incorrect docker image being used in local build script
• PR #35: Remove #include <nccl.h> from raft/error.hpp
• PR #40: Preemptively fixed future CUDA 11 related errors.
• PR #43: Fixed CUDA version selection mechanism for SpMV.
• PR #46: Fix for cpp file extension issue (nvcc-enforced).
• PR #48: Fix gtest target names in cmake build gtest option.
• PR #49: Skip raft comms test if raft module doesn't exist

RAFT 0.14.0 (Date TBD)

New Features

• Initial RAFT version
• PR #3: defining raft::handle_t, device_buffer, host_buffer, allocator classes

Bug Fixes

• PR #5: Small build.sh fixes