- Add nvcomp support to older CUDA 11 versions on aarch64. (#255) @bdice
- Unify
KVIKIO_CUFILE_FOUND(#243) @madsbk - Disable the batch API when in compatibility mode (#239) @madsbk
- Fix libcufile dependency. (#237) @bdice
- Revert CUDA 12.0 CI workflows to branch-23.08. (#252) @bdice
- Add support for nvCOMP batch API (#249) @Alexey-Kamenev
- Use cuda-version to constrain cudatoolkit. (#247) @bdice
- Make C++ & Python teams owners of
legate/(#246) @jakirkham - Use nvcomp conda package. (#245) @bdice
- Adding code owners (#244) @madsbk
- Clean up dependency lists (#241) @vyasr
- Clean up isort configs (#240) @vyasr
- Update to CMake 3.26.4 (#238) @vyasr
- use rapids-upload-docs script (#234) @AyodeAwe
- Migrate as much as possible to pyproject.toml, stop using versioneer to manage versions, update dependencies.yaml. (#232) @bdice
- Remove documentation build scripts for Jenkins (#230) @ajschmidt8
- Use the Zarr's new
getitems()API (#131) @madsbk
- Drop Python 3.8 and run Python 3.9 tests/builds (#206) @shwina
- Use the new registry and mapper API of Legate (#202) @madsbk
- Add sccache s3 controls (#226) @robertmaynard
- fixed import of
ArrayLikeandDTypeLike(#219) @madsbk load_library(): fixed the mode argument, which was ignored by mistake (#199) @madsbk
- Legate HDF5 using kerchunk (#222) @madsbk
- run docs nightly too (#221) @AyodeAwe
- C++ bindings to the Batch API (#220) @madsbk
- mypy: bump to v1.3.0 (#214) @madsbk
- Update cupy dependency (#213) @vyasr
- Enable sccache hits for local builds (#210) @AyodeAwe
- Revert to branch-23.06 for shared-action-workflows (#209) @shwina
- Zarr+nvCOMP (#208) @madsbk
- Drop Python 3.8 and run Python 3.9 tests/builds (#206) @shwina
- isort clean up (#205) @madsbk
- Only look for libcufile.so.0 (#203) @wence-
- Use the new registry and mapper API of Legate (#202) @madsbk
- Remove usage of rapids-get-rapids-version-from-git (#201) @jjacobelli
- Legate Zarr (#198) @madsbk
- Add API to get compatibility mode status in a FileHandle object (#197) @vuule
- Update clang-format to 16.0.1. (#196) @bdice
- Use ARC V2 self-hosted runners for GPU jobs (#195) @jjacobelli
- Optimize small reads and writes (#190) @madsbk
- Remove underscore in build string. (#188) @bdice
- Fallback to use the CUDA primary context (#189) @madsbk
- posix_io: fix error message and allow
nbytes == 0on write (#184) @madsbk - Support of stream ordered device memory allocations (async mallocs) (#181) @madsbk
- Implement
build.sh(#185) @madsbk - Legate Support (#183) @madsbk
- Fix docs build to be
pydata-sphinx-theme=0.13.0compatible (#180) @galipremsagar - Update to GCC 11 (#179) @bdice
- Fix GHA build workflow (#177) @AjayThorve
- Remove Jenkins/
gpuCIreferences (#174) @ajschmidt8 - Update
ops-bot.yamlconfig (#173) @ajschmidt8 - nvcomp xfail compression ratios (#167) @madsbk
- Move date to build string in
condarecipe (#165) @ajschmidt8 - Add gh actions (#164) @AjayThorve
- Manually open and close the cuFile driver (#160) @madsbk
- nvcomp tests: update hardcoded lengths (#156) @madsbk
- Update changelog for releases 22.06-22.12. (#157) @bdice
- Add nvcomp to kvikio-exports export set (#155) @trxcllnt
- Use pre-commit for style checks. (#154) @bdice
- Enable copy_prs. (#151) @bdice
- Don't use CMake 3.25.0 as it has a FindCUDAToolkit show stopping bug (#146) @robertmaynard
- dlopen: now trying "libcufile.so.1", "libcufile.so.0", "libcufile.so" (#141) @madsbk
- Update nvcomp's expected sizes when testing (#134) @madsbk
is_host_memory(): returnstruewhenCUDA_ERROR_NOT_INITIALIZED(#133) @madsbk
- Use rapidsai CODE_OF_CONDUCT.md (#144) @bdice
- Dependency clean up (#137) @madsbk
- Overload
numpy.fromfile()andcupy.fromfile()(#135) @madsbk
- Rename
reset_num_threads()andreset_task_size()(#123) @madsbk
- Fixing
kvikio_dev_cuda11.5.ymland clean up (#122) @madsbk
- Document that minimum required CMake version is now 3.23.1 (#132) @robertmaynard
- Use Zarr v2.13.0a2 (#129) @madsbk
- Allow cupy 11 (#128) @galipremsagar
- Set version when dlopen() cuda and cufile (#127) @madsbk
- Fall back to compat mode if we cannot open the file with
O_DIRECT(#126) @madsbk - Update versioneer to v0.22 (#124) @madsbk
- Rename
reset_num_threads()andreset_task_size()(#123) @madsbk - document channel_priority for conda (#121) @dcherian
- Update nvComp bindings to 2.3.3 (#120) @thomcom
- Use rapids-cmake 22.10 best practice for RAPIDS.cmake location (#118) @robertmaynard
- Standalone Downstream C++ Build Example (#29) @madsbk
- Fix typo in GDS availability check (#78) @jakirkham
- CI: install cuDF to test nvCOMP (#108) @madsbk
- Require
pythoninrun(#101) @jakirkham - Check stub error (#86) @madsbk
- FindcuFile now searches in the current CUDA Toolkit location (#85) @robertmaynard
- Fix typo in GDS availability check (#78) @jakirkham
- Defer loading of
custom.js(#114) @galipremsagar - Add stable channel install instruction (#107) @dcherian
- Use documented header template for
doxygen(#105) @galipremsagar - Fix issues with day & night modes in docs (#95) @galipremsagar
- add Notes related to page-aligned read/write methods to the Python docstrings (#91) @grlee77
- minor docstring fixes (#89) @grlee77
- Conda environment file (#103) @madsbk
- CI: build without cuFile (#100) @madsbk
- Support host memory (#82) @madsbk
- Mark detail functions as inline (#69) @vyasr
- Embed Cython docstrings in generated files. (#58) @vyasr
- Add new files to update version script (#53) @charlesbluca
- Fix conda recipes (#74) @Ethyling
- Use conda compilers (#71) @Ethyling
- dlopen
libcuda.so(#70) @madsbk - Use CMake provided targets for the cuda driver and dl libraries (#68) @robertmaynard
- Use namespace kvikio::detail (#67) @madsbk
- Build kvikio using libkvikio from CPU job in GPU job (#64) @Ethyling
- Use conda to build python packages during GPU tests (#62) @Ethyling
convert_size2off(): fix different signedness (#60) @madsbk- Keep the rapids-cmake version insync with calver (#59) @robertmaynard
- compat_mode per
FileHandle(#54) @madsbk - Add
ops-bot.yamlconfig file (#52) @ajschmidt8 - python: bump versions to v22.06 (#51) @madsbk