--- title: NVIDIA Labs releases cuda-oxide 0.1, an experimental... description: NVIDIA Labs has released cuda-oxide 0.1, an experimental open-source Rust-to-CUDA compiler that allows writing SIMT GPU kernels in standard Rust, compiling... canonical: https://daily.dev/posts/nvidia-labs-releases-cuda-oxide-0-1-an-experimental-rust-to-cuda-compiler-tx8t5seck twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: NVIDIA Labs releases cuda-oxide 0.1, an experimental Rust-to-CUDA compiler | daily.dev og:description: NVIDIA Labs has released cuda-oxide 0.1, an experimental open-source Rust-to-CUDA compiler that allows writing SIMT GPU kernels in standard Rust, compiling... og:url: https://daily.dev/posts/nvidia-labs-releases-cuda-oxide-0-1-an-experimental-rust-to-cuda-compiler-tx8t5seck og:image: https://api.daily.dev/og/posts/Tx8T5sEck.png og:image:alt: NVIDIA Labs releases cuda-oxide 0.1, an experimental Rust-to-CUDA compiler og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # NVIDIA Labs releases cuda-oxide 0.1, an experimental Rust-to-CUDA compiler **[Collections](https://daily.dev/sources/collections)** · 2 min read · 5 upvotes · 1 comments ## Summary NVIDIA Labs has released cuda-oxide 0.1, an experimental open-source Rust-to-CUDA compiler that allows writing SIMT GPU kernels in standard Rust, compiling them to PTX without DSLs or foreign bindings. It works as a custom rustc codegen backend, routing Rust MIR through Pliron IR and LLVM IR to PTX. The compiler supports generics, closures, warp operations, shared memory, TMA, atomics, and thread block clusters. A sample GEMM kernel achieves 868 TFLOPS on an NVIDIA B200 (~58% of cuBLAS). The project requires Rust nightly, CUDA Toolkit 12.x+, and LLVM 21+, and is explicitly in alpha with an unstable API. ## Content NVIDIA Labs has released [cuda-oxide 0.1](https://github.com/NVlabs/cuda-oxide), an experimental open-source compiler that lets you write SIMT GPU kernels in standard Rust and compile them directly to PTX — no DSLs, no foreign language bindings. ## How it works cuda-oxide is a custom `rustc` codegen backend. Functions marked with `#[kernel]` go through a pipeline: Rust MIR → Pliron IR → LLVM IR → PTX. A `#[cuda_module]` macro embeds the resulting device artifacts directly into the host binary, so host and device code live in the same source file. The compiler supports generics, closures with captures, and the full range of GPU intrinsics you'd expect: warp operations, shared memory, TMA, atomics, and thread block clusters. There's also an async execution layer built around `DeviceOperation` graphs for GPU programming. ## Performance The project ships 46 examples, including a GEMM kernel that hits 868 TFLOPS on an NVIDIA B200 — about 58% of cuBLAS's speed-of-light. Not parity, but respectable for an alpha. ## Requirements and status You'll need Rust nightly, CUDA Toolkit 12.x or later, and LLVM 21+. The project is explicitly in alpha: bugs are expected, and the API will break. That said, it's under active development from NVIDIA Labs, which gives it more institutional backing than most experimental GPU Rust projects. For anyone who's wrestled with CUDA C++ or tried to make existing Rust GPU crates work, this is worth watching. Writing kernels in idiomatic Rust with real ownership semantics and generics is a meaningful quality-of-life improvement — if the project matures. ## Similar posts on daily.dev - [CUDA-Oxide 0.2 Brings Early Improvements To Pure Rust CUDA Kernels](https://daily.dev/posts/cuda-oxide-0-2-brings-early-improvements-to-pure-rust-cuda-kernels-glwfh6z2x) · Phoronix · 1 upvotes · 0 comments - [Offloading Rust To GPUs Proves Capable Of High Performance With Memory Safety](https://daily.dev/posts/offloading-rust-to-gpus-proves-capable-of-high-performance-with-memory-safety-6ojskq9d9) · Phoronix · 0 upvotes · 0 comments --- Tags: [#rust](https://daily.dev/tags/rust), [#nvidia](https://daily.dev/tags/nvidia), [#gpu](https://daily.dev/tags/gpu), [#compiler](https://daily.dev/tags/compiler), [#cuda](https://daily.dev/tags/cuda) [View this post on daily.dev](https://daily.dev/posts/nvidia-labs-releases-cuda-oxide-0-1-an-experimental-rust-to-cuda-compiler-tx8t5seck) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"NVIDIA Labs releases cuda-oxide 0.1, an experimental Rust-to-CUDA compiler","url":"https://daily.dev/posts/nvidia-labs-releases-cuda-oxide-0-1-an-experimental-rust-to-cuda-compiler-tx8t5seck","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/nvidia-labs-releases-cuda-oxide-0-1-an-experimental-rust-to-cuda-compiler-tx8t5seck"},"datePublished":"2026-05-09T01:49:49.176Z","dateModified":"2026-05-11T17:51:27.835Z","description":"NVIDIA Labs has released cuda-oxide 0.1, an experimental open-source Rust-to-CUDA compiler that allows writing SIMT GPU kernels in standard Rust, compiling...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/b888d84ca8a3e26ba56a8e85fb75f3f5?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/b888d84ca8a3e26ba56a8e85fb75f3f5?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"Collections","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"Collections","logo":"https://media.daily.dev/image/upload/s--fk_6ycEi--/f_auto,q_auto/v1780996001/logos/collections?_a=BAMAMiWQ0","url":"https://daily.dev/sources/collections"},"commentCount":1,"discussionUrl":"https://daily.dev/posts/nvidia-labs-releases-cuda-oxide-0-1-an-experimental-rust-to-cuda-compiler-tx8t5seck","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":5},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":1}],"keywords":"rust,nvidia,gpu,compiler,cuda","timeRequired":"PT2M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Collections","item":"https://daily.dev/sources/collections"},{"@type":"ListItem","position":3,"name":"NVIDIA Labs releases cuda-oxide 0.1, an experimental Rust-to-CUDA compiler"}]} ```