From d83943b9b7e3769330de71d5ff23322b5794ac79 Mon Sep 17 00:00:00 2001 From: Canberk Pitirli Date: Wed, 28 Oct 2026 21:35:15 +0300 Subject: [PATCH] docs(cuda-oxide): put the Rust row inside the prerequisites table Signed-off-by: Canberk Pitirli --- cuda-oxide/cuda-oxide-book/getting-started/installation.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/cuda-oxide/cuda-oxide-book/getting-started/installation.md b/cuda-oxide/cuda-oxide-book/getting-started/installation.md index 7557eb8e13..81dafff56f 100644 --- a/cuda-oxide/cuda-oxide-book/getting-started/installation.md +++ b/cuda-oxide/cuda-oxide-book/getting-started/installation.md @@ -13,6 +13,7 @@ This section walks through everything you need to get `cargo oxide run vecadd` w | **CUDA Toolkit** | 13.0+ | `nvcc`, `cuda.h` and `curand.h` must be available | | **LLVM** | 21+ | Must include the NVPTX backend | | **Clang** | 21+ | `clang-21` — needed by `bindgen` for host `cuda-bindings` | +| **Rust** | 1.99.1 (pinned) | Pinned in `cuda-oxide/rust-toolchain.toml` | The host runtime does not link `libcuda` at build time. The shared `cuda-bindings` crate loads it at the first driver call, so a binary starts @@ -20,7 +21,6 @@ without a driver and fails on that call with `CUDA_ERROR_NOT_INITIALIZED`; the error message names the library files the loader tried. A driver whose CUDA major version is older than the toolkit's (a 12.x driver with a 13.x build) fails the same way with a "driver too old" message. -| **Rust** | 1.99.1 (pinned) | Pinned in `cuda-oxide/rust-toolchain.toml` | :::{note} cuda-oxide currently targets **Linux only**. Windows is not supported.