Enable publishing of macOS wheel (#1818)

* Enable publishing of macOS wheel

* Update macOS target to 14.0+

* Update doc

Author: matthewdouglas

.github/scripts/set_platform_tag.py

@@ -9,7 +9,7 @@ def get_platform_tag(architecture):
     if system == "Linux":
         tag = "manylinux_2_24_x86_64" if architecture == "x86_64" else "manylinux_2_24_aarch64"
     elif system == "Darwin":
-        tag = "macosx_13_1_x86_64" if architecture == "x86_64" else "macosx_13_1_arm64"
+        tag = "macosx_14_0_arm64"
     elif system == "Windows":
         tag = "win_amd64" if architecture == "x86_64" else "win_arm64"
     else:

.github/workflows/python-package.yml

@@ -266,8 +266,7 @@ jobs:
           # `pip install https://github.com/bitsandbytes-foundation/bitsandbytes/releases/download/continuous-release_main/bitsandbytes-1.33.7.preview-py3-none-manylinux_2_24_x86_64.whl`
           STABLE_PLACEHOLDER_VERSION="1.33.7.preview"
 
-          # exclude macos wheels for now
-          find tmp/ -type f -name '*.whl' ! -name '*macos*' -print0 | while IFS= read -r -d '' wheel; do
+          find tmp/ -type f -name '*.whl' -print0 | while IFS= read -r -d '' wheel; do
             wheel_filename=$(basename "$wheel")
 
             # Strip off the original version
@@ -318,9 +317,11 @@ jobs:
             if [[ "$fname" == *"manylinux_2_24_x86_64"* ]]; then
               echo "### Linux (x86_64)" >> body.md
             elif [[ "$fname" == *"manylinux_2_24_aarch64"* ]]; then
-              echo "### Linux (ARM/aarch64)" >> body.md
+              echo "### Linux (aarch64)" >> body.md
             elif [[ "$fname" == *"win_amd64"* ]]; then
               echo "### Windows (x86_64)" >> body.md
+            elif [[ "$fname" == *"macosx"* ]]; then
+              echo "### macOS 14+ (arm64)" >> body.md
             else
               echo "### Other platform" >> body.md
             fi
@@ -340,7 +341,7 @@ jobs:
           > pip install https://.../bitsandbytes-1.33.7-preview-py3-none-manylinux_2_24_x86_64.whl
           Collecting bitsandbytes==1.33.7rc0
           ...
-          Successfully installed bitsandbytes-0.46.0.dev0
+          Successfully installed bitsandbytes-0.49.0.dev0
           ```
           ENDOFMARKDOWN
 
@@ -405,9 +406,6 @@ jobs:
           pattern: "bdist_wheel_*"
           merge-multiple: true
 
-      - name: Remove macOS wheels
-        run: rm dist/*macos*
-
       - name: Publish to PyPI
         uses: pypa/gh-action-pypi-publish@release/v1
         with:

CMakeLists.txt

@@ -37,7 +37,7 @@ set_property(CACHE COMPUTE_BACKEND PROPERTY STRINGS cpu cuda hip mps xpu)
 option(PTXAS_VERBOSE "Pass through -v flag to PTX Assembler" OFF)
 
 if(APPLE)
-  set(CMAKE_OSX_DEPLOYMENT_TARGET 13.1)
+  set(CMAKE_OSX_DEPLOYMENT_TARGET 14.0)
 endif()
 
 set(BNB_OUTPUT_NAME "bitsandbytes")

docs/source/installation.mdx

@@ -2,8 +2,7 @@
 
 Welcome to the installation guide for the `bitsandbytes` library! This document provides step-by-step instructions to install `bitsandbytes` across various platforms and hardware configurations.
 
-We provide official support for NVIDIA GPUs, CPUs, Intel XPUs, and Intel Gaudi platforms. We also have experimental support for
-additional platforms such as AMD ROCm.
+We provide official support for NVIDIA GPUs, CPUs, Intel XPUs, and Intel Gaudi. We also have experimental support for additional platforms such as AMD ROCm and Apple Silicon.
 
 ## Table of Contents
 
@@ -100,7 +99,7 @@ pip install -e .   # `-e` for "editable" install, when developing BNB (otherwise
 
 Compilation from source on Windows systems require Visual Studio with C++ support as well as an installation of the CUDA Toolkit.
 
-To compile from source, you need CMake >= **3.22.1** and Python >= **3.9** installed. You should also install CUDA Toolkit by following the [CUDA Installation Guide for Windows](https://docs.nvidia.com/cuda/cuda-installation-guide-microsoft-windows/index.html) guide from NVIDIA. The current minimum supported CUDA Toolkit version that we support is **11.8**.
+To compile from source, you need CMake >= **3.22.1** and Python >= **3.10** installed. You should also install CUDA Toolkit by following the [CUDA Installation Guide for Windows](https://docs.nvidia.com/cuda/cuda-installation-guide-microsoft-windows/index.html) guide from NVIDIA. The current minimum supported CUDA Toolkit version that we support is **11.8**.
 
 ```bash
 git clone https://github.com/bitsandbytes-foundation/bitsandbytes.git && cd bitsandbytes/
@@ -163,6 +162,7 @@ The currently distributed `bitsandbytes` packages are built with the following c
 | **Linux x86-64**   | GCC 11.4             | AVX2                 |
 | **Linux aarch64**  | GCC 11.4             |                      |
 | **Windows x86-64** | MSVC 19.43+ (VS2022) | AVX2                 |
+| **macOS arm64**    | Apple Clang 17       |                      |
 
 The Linux build has a minimum glibc version of 2.24.
 
@@ -243,4 +243,11 @@ pip install --force-reinstall https://github.com/bitsandbytes-foundation/bitsand
 pip install --force-reinstall https://github.com/bitsandbytes-foundation/bitsandbytes/releases/download/continuous-release_main/bitsandbytes-1.33.7.preview-py3-none-win_amd64.whl
 ```
 </hfoption>
+<hfoption id="macOS">
+
+```bash
+# Note: if you don't want to reinstall our dependencies, append the `--no-deps` flag!
+pip install --force-reinstall https://github.com/bitsandbytes-foundation/bitsandbytes/releases/download/continuous-release_main/bitsandbytes-1.33.7.preview-py3-none-macosx_14_0_arm64.whl
+```
+</hfoption>
 </hfoptions>

pyproject.toml

@@ -31,7 +31,7 @@ classifiers = [
     "Intended Audience :: Developers",
     "Intended Audience :: Science/Research",
     "Operating System :: POSIX :: Linux",
-    # "Operating System :: MacOS",
+    "Operating System :: MacOS",
     "Operating System :: Microsoft :: Windows",
     "Programming Language :: C++",
     "Programming Language :: Python :: Implementation :: CPython",