NVIDIA/MatX v0.9.3

v0.9.3

Repository: NVIDIA/MatX

Tag: v0.9.3

Published: 2025-09-26T23:30:05Z

Prerelease: no

Release notes: New operators: find_peaks, zipvec Key Updates:

• C2R FFT transforms
• Indexing speedup for accessing tensors

What's Changed

• Add qualifier to maybe unused variables by @cliffburdick in https://github.com/NVIDIA/MatX/pull/1027
• Add CTK 12.9.1 / Ubuntu 24.04 container recipe by @tmartin-gh in https://github.com/NVIDIA/MatX/pull/1028
• Added find_peaks operator by @cliffburdick in https://github.com/NVIDIA/MatX/pull/1029
• Add missing include by @miscco in https://github.com/NVIDIA/MatX/pull/1031
• Removing legacy docs folder by @cliffburdick in https://github.com/NVIDIA/MatX/pull/1032
• Updated CCCL to 3.0.0 to prepare for CTK 13.0 by @cliffburdick in https://github.com/NVIDIA/MatX/pull/1030
• Fix error in nvc++ by @cliffburdick in https://github.com/NVIDIA/MatX/pull/1036
• Fixed std::accumulate starting value by @cliffburdick in https://github.com/NVIDIA/MatX/pull/1035
• Updated developer docs for EPT by @cliffburdick in https://github.com/NVIDIA/MatX/pull/1038
• Fixed issue where op=transform was double-calling transform by @cliffburdick in https://github.com/NVIDIA/MatX/pull/1037
• Make cache entries per-thread since most CUDA library handles are not thread-safe by @cliffburdick in https://github.com/NVIDIA/MatX/pull/1040
• Add pad operator for padding input operators along one dimension by @tbensonatl in https://github.com/NVIDIA/MatX/pull/1041
• Remove unreachable return by @cliffburdick in https://github.com/NVIDIA/MatX/pull/1042
• Add zipvec operator by @tbensonatl in https://github.com/NVIDIA/MatX/pull/1033
• Fixed const issues seen in user's code by @cliffburdick in https://github.com/NVIDIA/MatX/pull/1044
• Added negative file tests by @cliffburdick in https://github.com/NVIDIA/MatX/pull/1045
• Add conditional CUDA 13+ support for select vector types by @cliffburdick in https://github.com/NVIDIA/MatX/pull/1047
• Update CUDA macro by @cliffburdick in https://github.com/NVIDIA/MatX/pull/1048
• Use index_t with get_grid_dims to support 32-bit builds by @tbensonatl in https://github.com/NVIDIA/MatX/pull/1050
• qr_econ unreachable fix by @cliffburdick in https://github.com/NVIDIA/MatX/pull/1049
• Missing return value in QR even though code is unreachable by @cliffburdick in https://github.com/NVIDIA/MatX/pull/1051
• Avoid deprecated thrust iterators by @miscco in https://github.com/NVIDIA/MatX/pull/1055
• Added support for C2R FFTs via irfft and irfft2 by @cliffburdick in https://github.com/NVIDIA/MatX/pull/1054
• Rename version_config.h -> matx/version_config.h by @valgur in https://github.com/NVIDIA/MatX/pull/1052
• Refactor Storage system to use duck-typed allocators by @cliffburdick in https://github.com/NVIDIA/MatX/pull/1046
• Added min/max headers where needed by @cliffburdick in https://github.com/NVIDIA/MatX/pull/1056
• Optimize tensor indexing for ranks 1-4 with explicit stride calculations by @cliffburdick in https://github.com/NVIDIA/MatX/pull/1057
• Add cusparse explicitly to link libraries by @agirault in https://github.com/NVIDIA/MatX/pull/1058
• Add shared_ptr constructor to Storage class by @cliffburdick in https://github.com/NVIDIA/MatX/pull/1059

New Contributors

• @valgur made their first contribution in https://github.com/NVIDIA/MatX/pull/1052
• @agirault made their first contribution in https://github.com/NVIDIA/MatX/pull/1058

Full Changelog: https://github.com/NVIDIA/MatX/compare/v0.9.2...v0.9.3