Triton fork for Windows support
See windows-fix branch for the code.
Based on wkpark, mantaionut, eaplatanios, and more development in the community. Thank you all!
Free software should run on non-free platforms, as per Richard Stallman.
- Forked from the official main branch (> 3.0.0) as of today
- Can build the package locally and jit Python functions
- When I run Flux or CogVideoX in ComfyUI on Windows, it's almost as fast as on WSL on the same machine (although the memory usage is hard to profile in WSL)
- Only MSVC is supported, from my experience it's much more stable than GCC and Clang when working with CUDA on Windows
- Only CUDA is supported, help wanted to support AMD
- TODO: Fix all tests
- TODO: Auto find the paths of MSVC, Windows SDK, and CUDA when jitting in
python/triton/runtime/build.py - TODO: Build wheels using cibuildwheel
First, build LLVM according to the instructions of the official triton: https://github.com/triton-lang/triton?tab=readme-ov-file#building-with-a-custom-llvm
Also, download JSON according to the setup.py:
https://github.com/nlohmann/json/releases/download/v3.11.3/include.zip
Set their paths:
$Env:JSON_SYSPATH = "C:/json"
$Env:LLVM_SYSPATH = "C:/llvm-project/build"Only offline build is supported:
$Env:TRITON_OFFLINE_BUILD = "1"I recommend to use ccache:
$Env:TRITON_BUILD_WITH_CCACHE = "1"Set the binary, include, and library paths: (Help wanted to auto find these in CMake)
$Env:Path =
"C:\Python310;" +
"C:\Python310\Scripts;" +
"C:\Program Files (x86)\Microsoft Visual Studio\2022\BuildTools\Common7\IDE\CommonExtensions\Microsoft\CMake\CMake\bin;" +
"C:\Program Files (x86)\Microsoft Visual Studio\2022\BuildTools\VC\Tools\MSVC\14.41.34120\bin\Hostx64\x64;" +
"C:\Program Files (x86)\Windows Kits\10\bin\10.0.26100.0\x64"
$Env:INCLUDE =
"C:\Program Files (x86)\Microsoft Visual Studio\2022\BuildTools\VC\Tools\MSVC\14.41.34120\include;" +
"C:\Program Files (x86)\Windows Kits\10\Include\10.0.26100.0\shared;" +
"C:\Program Files (x86)\Windows Kits\10\Include\10.0.26100.0\ucrt;" +
"C:\Program Files (x86)\Windows Kits\10\Include\10.0.26100.0\um;" +
"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.5\include;" +
"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v12.5\extras\CUPTI\include"
$Env:LIB =
"C:\Program Files (x86)\Microsoft Visual Studio\2022\BuildTools\VC\Tools\MSVC\14.41.34120\lib\x64;" +
"C:\Program Files (x86)\Windows Kits\10\Lib\10.0.26100.0\ucrt\x64;" +
"C:\Program Files (x86)\Windows Kits\10\Lib\10.0.26100.0\um\x64"Then, make an editable build using pip:
pip install --no-build-isolation --verbose -e python- To implement
dlopen, dlfcn-win32 is added tothirdparty/and linked in CMake for building the package, and inthird_party/nvidia/backend/driver.canddriver.pyit's rewritten withLoadLibraryfor jitting - In
lib/Analysis/Utility.cppandlib/Dialect/TritonGPU/Transforms/Utility.cpp, explicit namespaces are added to support the resolution behaviors of MSVC - In
python/src/interpreter.ccthe GCC built-in__ATOMICmemory orders are replaced withstd::memory_order