# NVIDIA/stdexec

`std::execution`, the standard C++ framework for asynchronous and parallel programming.

Repository: https://github.com/NVIDIA/stdexec
Canonical: https://ross.abutalabs.com/products/stdexec
Homepage: https://nvidia.github.io/stdexec/
Language: C++
License: Apache-2.0
License Family: permissive
Last push: 2026-08-23T03:31:50+00:00

## Health v2 (maintenance only)
Score: 77/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 99, release rhythm 35, longevity 100
- inputs: {"age_days": 1941, "days_push": 10, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2417, forks 268 (observed 2026-08-28T04:06:50.380992+00:00)

## What it is
stdexec is NVIDIA's reference implementation of std::execution, the C++26 standard framework for asynchronous and parallel programming based on the sender/receiver model (P2300). It lets developers compose lazy asynchronous pipelines that run on threads, thread pools, GPUs, or custom schedulers with structured concurrency guarantees.

## Use cases
- compose asynchronous work as lazy sender pipelines in C++
- run parallel computations on a thread pool with structured concurrency
- schedule GPU work from C++ with a standard async model
- replace callbacks and std::async with composable async algorithms
- integrate coroutines with asynchronous execution via co_await on senders
- prototype C++26 std::execution APIs before compilers ship them

## When to choose
- you want a modern, composable, zero-overhead async model in C++20+
- you need structured cancellation and error handling for concurrent tasks
- you are targeting heterogeneous execution across CPUs and GPUs
- you want to future-proof code against the C++26 execution standard

## When to avoid
- you need a stable, frozen API for production-critical code
- your compiler cannot handle heavy C++20 template metaprogramming
- you only need simple threading with std::thread or std::async
- you require a small dependency footprint in legacy C++ codebases

## Facets
- artifact type: library
- maturity: experimental
- function: concurrency, developer-tools
- domain: developer-tools, cross-platform, gpu-computing
- platform: cpp, windows
- tags: senders, structured-concurrency, async, parallel-programming, header-only, c++26, p2300, linux, macos, gpu

## Member repositories
- NVIDIA/stdexec (main) score 77

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:50.380992+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T02:33:07.449105+00:00, confidence not recorded.
  - readme: https://github.com/NVIDIA/stdexec (fetched 2026-08-28T04:06:50.380992+00:00, sha 058acbf94162)
  - homepage: https://nvidia.github.io/stdexec/ (fetched 2026-08-29T10:13:34.430308+00:00, sha d1698e1a7533)
- Data as of 2026-08-30T08:39:29.467469+00:00.