Add PyTorch 2.12 support

.github/workflows/building.yml

@@ -16,23 +16,25 @@ jobs:
       matrix:
         os: [ubuntu-22.04, macos-14, windows-2022, ubuntu-22.04-arm]
         python-version: ['3.10', '3.11', '3.12', '3.13', '3.14']
-        torch-version: [2.11.0]
-        cuda-version: ['cpu', 'cu126', 'cu128', 'cu130']
+        torch-version: [2.12.0]
+        cuda-version: ['cpu', 'cu126', 'cu130', 'cu132']
         exclude:
           - os: windows-2022
             cuda-version: 'cu130'
+          - os: windows-2022
+            cuda-version: 'cu132'
           - os: macos-14
             cuda-version: 'cu126'
-          - os: macos-14
-            cuda-version: 'cu128'
           - os: macos-14
             cuda-version: 'cu130'
+          - os: macos-14
+            cuda-version: 'cu132'
           - os: ubuntu-22.04-arm
             cuda-version: 'cu126'
-          - os: ubuntu-22.04-arm
-            cuda-version: 'cu128'
           - os: ubuntu-22.04-arm
             cuda-version: 'cu130'
+          - os: ubuntu-22.04-arm
+            cuda-version: 'cu132'
 
     steps:
       - uses: actions/checkout@v4

.github/workflows/cuda/cu132-Linux-env.sh

@@ -0,0 +1,8 @@
+#!/bin/bash
+
+CUDA_HOME=/usr/local/cuda-13.2
+LD_LIBRARY_PATH=${CUDA_HOME}/lib64:${LD_LIBRARY_PATH}
+PATH=${CUDA_HOME}/bin:${PATH}
+
+export FORCE_CUDA=1
+export TORCH_CUDA_ARCH_LIST="7.5;8.0;8.6;9.0;10.0;12.0+PTX"

.github/workflows/cuda/cu132-Linux.sh

@@ -0,0 +1,17 @@
+#!/bin/bash
+
+OS=ubuntu2204
+
+wget -nv https://developer.download.nvidia.com/compute/cuda/repos/${OS}/x86_64/cuda-${OS}.pin
+sudo mv cuda-${OS}.pin /etc/apt/preferences.d/cuda-repository-pin-600
+
+wget -nv https://developer.download.nvidia.com/compute/cuda/13.2.0/local_installers/cuda-repo-${OS}-13-2-local_13.2.0-595.45.04-1_amd64.deb
+
+sudo dpkg -i cuda-repo-${OS}-13-2-local_13.2.0-595.45.04-1_amd64.deb
+sudo cp /var/cuda-repo-${OS}-13-2-local/cuda-*-keyring.gpg /usr/share/keyrings/
+
+sudo apt-get -qq update
+sudo apt install cuda-nvcc-13-2 cuda-libraries-dev-13-2
+sudo apt clean
+
+rm -f *.deb

.github/workflows/testing.yml

@@ -20,7 +20,7 @@ jobs:
       matrix:
         os: [ubuntu-latest, macos-latest, windows-2022]
         python-version: ['3.10']
-        torch-version: [2.11.0]
+        torch-version: [2.12.0]
 
     steps:
       - uses: actions/checkout@v4

README.md

@@ -34,28 +34,28 @@ Note that only `value` comes with autograd support, as `index` is discrete and t
 
 We provide pip wheels for all major OS/PyTorch/CUDA combinations, see [here](https://data.pyg.org/whl).
 
-#### PyTorch 2.11
+#### PyTorch 2.12
 
-To install the binaries for PyTorch 2.11, simply run
+To install the binaries for PyTorch 2.12, simply run
 
 ```
-pip install torch-scatter torch-sparse -f https://data.pyg.org/whl/torch-2.11.0+${CUDA}.html
+pip install torch-scatter torch-sparse -f https://data.pyg.org/whl/torch-2.12.0+${CUDA}.html
 ```
 
-where `${CUDA}` should be replaced by either `cpu`, `cu126`, `cu128`, or `cu130` depending on your PyTorch installation.
+where `${CUDA}` should be replaced by either `cpu`, `cu126`, `cu130`, or `cu132` depending on your PyTorch installation.
 
-|             | `cpu` | `cu126` | `cu128` | `cu130` |
+|             | `cpu` | `cu126` | `cu130` | `cu132` |
 |-------------|-------|---------|---------|---------|
 | **Linux**   | ✅    | ✅      | ✅      | ✅      |
-| **Windows** | ✅    | ✅      | ✅      | ✅      |
+| **Windows** | ✅    | ✅      |         |         |
 | **macOS**   | ✅    |         |         |         |
 
-#### PyTorch 2.10
+#### PyTorch 2.11
 
-To install the binaries for PyTorch 2.10, simply run
+To install the binaries for PyTorch 2.11, simply run
 
 ```
-pip install torch-scatter torch-sparse -f https://data.pyg.org/whl/torch-2.10.0+${CUDA}.html
+pip install torch-scatter torch-sparse -f https://data.pyg.org/whl/torch-2.11.0+${CUDA}.html
 ```
 
 where `${CUDA}` should be replaced by either `cpu`, `cu126`, `cu128`, or `cu130` depending on your PyTorch installation.
@@ -66,12 +66,12 @@ where `${CUDA}` should be replaced by either `cpu`, `cu126`, `cu128`, or `cu130`
 | **Windows** | ✅    | ✅      | ✅      | ✅      |
 | **macOS**   | ✅    |         |         |         |
 
-#### PyTorch 2.9
+#### PyTorch 2.10
 
-To install the binaries for PyTorch 2.9, simply run
+To install the binaries for PyTorch 2.10, simply run
 
 ```
-pip install torch-scatter torch-sparse -f https://data.pyg.org/whl/torch-2.9.0+${CUDA}.html
+pip install torch-scatter torch-sparse -f https://data.pyg.org/whl/torch-2.10.0+${CUDA}.html
 ```
 
 where `${CUDA}` should be replaced by either `cpu`, `cu126`, `cu128`, or `cu130` depending on your PyTorch installation.
@@ -82,7 +82,7 @@ where `${CUDA}` should be replaced by either `cpu`, `cu126`, `cu128`, or `cu130`
 | **Windows** | ✅    | ✅      | ✅      | ✅      |
 | **macOS**   | ✅    |         |         |         |
 
-**Note:** Binaries of older versions are also provided for PyTorch 1.4.0, PyTorch 1.5.0, PyTorch 1.6.0, PyTorch 1.7.0/1.7.1, PyTorch 1.8.0/1.8.1, PyTorch 1.9.0, PyTorch 1.10.0/1.10.1/1.10.2, PyTorch 1.11.0, PyTorch 1.12.0/1.12.1, PyTorch 1.13.0/1.13.1, PyTorch 2.0.0/2.0.1, PyTorch 2.1.0/2.1.1/2.1.2, PyTorch 2.2.0/2.2.1/2.2.2, PyTorch 2.3.0/2.3.1, PyTorch 2.4.0/2.4.1, PyTorch 2.5.0/2.5.1, PyTorch 2.6.0, PyTorch 2.7.0/2.7.1, and PyTorch 2.8.0 (following the same procedure).
+**Note:** Binaries of older versions are also provided for PyTorch 1.4.0, PyTorch 1.5.0, PyTorch 1.6.0, PyTorch 1.7.0/1.7.1, PyTorch 1.8.0/1.8.1, PyTorch 1.9.0, PyTorch 1.10.0/1.10.1/1.10.2, PyTorch 1.11.0, PyTorch 1.12.0/1.12.1, PyTorch 1.13.0/1.13.1, PyTorch 2.0.0/2.0.1, PyTorch 2.1.0/2.1.1/2.1.2, PyTorch 2.2.0/2.2.1/2.2.2, PyTorch 2.3.0/2.3.1, PyTorch 2.4.0/2.4.1, PyTorch 2.5.0/2.5.1, PyTorch 2.6.0, PyTorch 2.7.0/2.7.1, PyTorch 2.8.0, and PyTorch 2.9.0 (following the same procedure).
 For older versions, you need to explicitly specify the latest supported version number or install via `pip install --no-index` in order to prevent a manual installation from source.
 You can look up the latest supported version number [here](https://data.pyg.org/whl).