wep21/rules_tensorrt_rtx — reverse-engineered prompt
Reverse engineered prompt
Build me a Bazel rules package that lets a C++ project use NVIDIA TensorRT RTX without the developer manually downloading and wiring up the SDK.
I want it to work through MODULE.bazel with a simple bazel_dep and module extension call, where the user can choose the TensorRT RTX version and CUDA version if needed. It should fetch the right prebuilt artifacts for Linux x86_64 and Windows x86_64, expose easy C++ deps for TensorRT RTX and the ONNX parser, and include a tiny example cc_binary that links against them so people can copy it.
Please include the Bazel files, extension code, platform build definitions, and a clear README showing exactly what to paste into MODULE.bazel and BUILD.bazel. Keep it small and focused, basically a reusable wrapper around TensorRT RTX for Bazel projects. Look up current Bazel module extension docs if you need to.
Want more depth? Deep Reverse