Asynchronous Operations

2026-10-01

2 min read time

Applies to Linux and Windows

This page covers the asynchronous operations extended API, providing non-blocking memory copy functions synchronized via barriers or pipelines.

cuda::memcpy_async	Asynchronously copies one range to another	libhipcxx 1.1.0 / CCCL 2.0.0 / CUDA 11.0
cuda::memcpy_async_tx	Asynchronously copies one range to another with manual transaction accounting	libhipcxx 1.2.0 / CCCL 2.0.0 / CUDA 11.1

Note

Asynchronous operations like memcpy_async <libcudacxx-extended-api-asynchronous-operations-memcpy-async> are non-blocking operations performed as-if by a new thread of execution.