Pull Requests bitsandbytes-foundation/bitsandbytes

Update AMD targets Documentation ROCm Build

#1832 by sstamenk was merged 2025-12-12 17:57

Fix: Python 3.14 compatibility with PyTorch 2.9 Cross Platform Dependencies

#1831 by matthewdouglas was merged 2025-12-11 18:19 v0.49.0

Update installation doc Documentation

#1830 by matthewdouglas was merged 2025-12-11 15:12 v0.49.0

Add release for DGX Spark cuda121 Build CUDA

#1829 by mfuntowicz was merged 2025-12-11 15:16 v0.49.0

CPU: workaround avx512 4bit dequantize accuracy issue for large blocksize x64 CPU

#1828 by matthewdouglas was merged 2025-12-10 21:19 v0.49.0

CUDA/ROCm: Remove dead code ROCm CUDA

#1827 by matthewdouglas was merged 2025-12-09 18:33 v0.49.0

Update workflow dependencies CI/CD CUDA

#1824 by matthewdouglas was merged 2025-12-08 17:31 v0.49.0

ROCm: Add gfx1150/gfx1151 to build targets ROCm Build

#1822 by matthewdouglas was merged 2025-12-03 18:56 v0.49.0

CUDA 13: aggressive compression of binary size Build CUDA

#1820 by matthewdouglas was merged 2025-12-03 01:08 v0.49.0

ROCm: reduce size of builds ROCm Build CI/CD

#1819 by matthewdouglas was merged 2025-12-02 22:44 v0.49.0

Enable publishing of macOS wheel macOS Build

#1818 by matthewdouglas was merged 2025-12-02 21:04 v0.49.0

Cleanup: remove FastBinarySearch Build Dependencies x64 CPU

#1817 by matthewdouglas was merged 2025-12-02 19:04 v0.49.0

Update README

#1816 by matthewdouglas was merged 2025-12-02 17:44

Hf kernel Intel x64 CPU

#1814 by jiqing-feng was merged 2025-12-09 16:04 v0.49.0

Remove old nightly workflow CI/CD

#1812 by matthewdouglas was merged 2025-11-21 22:07

CI: Run tests on PRs, refactor nightly test workflow CI/CD

#1811 by matthewdouglas was merged 2025-11-21 21:59

Replace NULL with nullptr in pythonInterface.cpp

#1809 by yuguo68 was merged 2025-11-18 16:30 v0.49.0

CI: Enable tests on Linux x86-64 with CUDA 13 CI/CD CUDA

#1808 by matthewdouglas was merged 2025-11-14 17:34

ROCm: Add build for ROCm 7.1 ROCm

#1807 by matthewdouglas was merged 2025-11-14 17:34 v0.49.0

CUDA: Drop compilation compatibility with Maxwell CUDA

#1806 by matthewdouglas was merged 2025-11-14 16:09 v0.49.0

Enable more tests on AMD for warp size 32 ROCm

#1805 by sstamenk was merged 2025-11-14 16:00 v0.49.0

Cpu fused kernel x64 CPU

#1804 by jiqing-feng was merged 2025-11-26 16:20 v0.49.0

CI: skip rebuilding CPU lib when building/installing wheels Build CI/CD

#1803 by matthewdouglas was merged 2025-11-13 18:52 v0.49.0

fix build error: "no case matching constant switch condition" ROCm

#1802 by yuguo68 was merged 2025-11-13 15:41 v0.49.0

Remove deprecated code

#1798 by matthewdouglas was merged 2025-11-04 19:07 v0.49.0

Tests: Run CPU tests against PyTorch 2.9 CI/CD

#1797 by matthewdouglas was merged 2025-11-04 15:13

Fix indexing overflow issue for blockwise quantization on AMD ROCm

#1796 by sstamenk was merged 2025-11-03 21:56 v0.49.0

Drop Python 3.9 support

#1795 by matthewdouglas was merged 2025-11-03 21:36 v0.49.0

mps: add nf4 dequantize/quantize kernel macOS

#1790 by bghira was closed 2025-11-12 17:42

Cpu C++ kernel x64 CPU

#1789 by jiqing-feng was merged 2025-11-13 15:17 v0.49.0