Environment:Sgl project Sglang CUDA Runtime
Sgl_project_Sglang_CUDA_Runtime is the CUDA C++ extension build environment for SGLang, used to compile custom CUDA kernels via PyTorch's `torch.utils.cpp_extension` and the `sgl-kernel` package.
Requirements
- NVIDIA CUDA Toolkit 12.3+ (12.8+ for Blackwell)
- NVCC compiler
- C++17-compatible host compiler (GCC 9+ or Clang 11+)
- PyTorch 2.9.1+ with CUDA support
- `torch.utils.cpp_extension` module
- `ninja` build system (recommended for parallel compilation)
- `setuptools` >= 61.0
Required By
See Also
Page Connections
Double-click a node to navigate. Hold to expand connections.
Principle
Implementation
Heuristic
Environment