Skip to main content
Ctrl+K
NVIDIA RMM - Home NVIDIA RMM - Home

NVIDIA RMM

  • GitHub
Home
rmm
cucimcudf-javacudfcugraphcumlcuvsdask-cudadask-cudfkvikiolibcudflibcumllibkvikiolibrapidsmpflibucxxnvforestraftrapids-cmakerapidsmpfrmmucxx
nightly (26.08)
nightly (26.08)stable (26.06)legacy (26.04)
NVIDIA RMM - Home NVIDIA RMM - Home

NVIDIA RMM

  • GitHub

Table of Contents

Contents

  • User Guide
    • RMM 26.06 Migration Guide
  • C++ API
    • Memory Resources
      • Memory Resources
      • Memory Resource Adaptors
    • Data Containers
    • Thrust Integrations
    • CUDA Device Management
    • CUDA Streams
    • Errors
    • Utilities
    • Namespaces
  • Python API
    • rmm (top-level)
    • rmm.mr (Memory Resources)
    • rmm.allocators
    • rmm.statistics
    • rmm.pylibrmm
    • rmm.librmm

NVIDIA RMM Documentation#

NVIDIA RMM (RAPIDS Memory Manager) is a library for allocating and managing GPU memory in C++ and Python.

Contents

  • User Guide
    • Installation
    • Migration Guides
    • Using RMM
    • Using RMM with third-party libraries
    • Memory statistics and profiling
  • C++ API
    • Memory Resources
    • Data Containers
    • Thrust Integrations
    • CUDA Device Management
    • CUDA Streams
    • Errors
    • Utilities
    • Namespaces
    • Indices and tables
  • Python API
    • rmm (top-level)
    • rmm.mr (Memory Resources)
    • rmm.allocators
    • rmm.statistics
    • rmm.pylibrmm
    • rmm.librmm

next

User Guide

NVIDIA NVIDIA
Privacy Policy | Your Privacy Choices | Terms of Service | Accessibility | Corporate Policies | Product Security | Contact

Copyright © 2018-2026, NVIDIA Corporation.