Skip to main content
Ctrl+K
CUDA Core Compute Libraries - Home CUDA Core Compute Libraries - Home

CUDA Core Compute Libraries

  • GitHub
CUDA Core Compute Libraries - Home CUDA Core Compute Libraries - Home

CUDA Core Compute Libraries

  • GitHub

Table of Contents

  • CUDA C++ Core Libraries
    • libcu++
      • Overview
      • Setup
        • Getting libcu++
        • Building & Testing libcu++
      • Standard API
        • Algorithms Library
        • C Library
        • Concepts Library
        • Container Library
        • Execution Library
        • Numerics Library
        • Ranges Library
        • Synchronization Library
        • Time Library
        • Type Support Library
        • Utility Library
      • Extended API
        • Macros
        • Bit
        • Execution model
        • Exception Handling
        • Memory model
        • Thread Groups
        • Synchronization Primitives
        • Asynchronous Operations
        • Memory access properties
        • Functional
        • Fancy Iterators
        • Type traits
        • Vector Tuple Protocol
        • Numeric
        • Random
        • Memory
        • Memory Resources
        • Math
        • Mdspan
        • Tensor Memory Accelerator (TMA)
        • Warp
        • Utility
        • Work stealing
      • Runtime
        • CUDA Runtime interactions
        • Streams
        • Events
        • Algorithm
        • Devices
        • Hierarchy
        • Launch
        • Buffer
        • Memory Pools
        • Legacy resources
      • PTX API
        • Examples
        • PTX Instructions
        • PTX Pragmas
      • CUDA Tile support
      • API reference
    • CUB
      • Overview
      • Thread-level Primitives
      • Warp-Wide “Collective” Primitives
        • Warp-wide Primitives
      • Block-Wide “Collective” Primitives
        • Block-wide Primitives
      • Device-Wide Primitives
        • Device-wide Primitives
      • Execution Environments
      • Determinism
      • Benchmarks
      • Tunings
      • Automated Tuning Infrastructure
      • API reference
      • Developer Overview
        • Thread-level
        • Warp-level
        • Block-scope
        • Device-scope
        • NVTX
        • CUB Tests
    • Thrust
      • Overview
      • Algorithms
        • Copying
        • Merging
        • Prefix sums
        • Reductions
        • Reordering
        • Searching
        • Set Operations
        • Sorting
        • Transformations
      • Containers
        • thrust::host_vector
        • thrust::device_vector
        • thrust::universal_host_pinned_vector
        • thrust::universal_vector
      • Function Objects
        • Function Object Adaptors
        • Placeholder Objects
        • Function Objects
      • Iterators
        • thrust::make_transform_input_output_iterator
        • thrust::make_constant_iterator
        • thrust::make_counting_iterator
        • thrust::make_permutation_iterator
        • thrust::make_tabulate_output_iterator
        • thrust::make_strided_iterator
        • thrust::make_transform_output_iterator
        • thrust::make_discard_iterator
        • thrust::make_zip_iterator
        • thrust::make_transform_iterator
        • thrust::make_shuffle_iterator
        • thrust::iterator_traversal_t
        • thrust::iterator_system_t
        • thrust::iterator_reference_t
        • thrust::iterator_pointer_t
        • thrust::iterator_traits
        • thrust::iterator_difference_t
        • thrust::iterator_value_t
        • thrust::bidirectional_host_iterator_tag
        • thrust::random_access_host_iterator_tag
        • thrust::output_host_iterator_tag
        • thrust::input_host_iterator_tag
        • thrust::forward_host_iterator_tag
        • thrust::retag
        • thrust::reinterpret_tag
        • thrust::bidirectional_traversal_tag
        • thrust::forward_traversal_tag
        • thrust::incrementable_traversal_tag
        • thrust::no_traversal_tag
        • thrust::random_access_traversal_tag
        • thrust::single_pass_traversal_tag
      • Memory Management
        • Allocators
        • Memory Resources
        • Memory Management
        • thrust::device_pointer_cast
        • thrust::device_free
        • thrust::device_delete
        • thrust::device_malloc
        • thrust::malloc
        • thrust::return_temporary_buffer
        • thrust::swap
        • thrust::device_new
        • thrust::free
        • thrust::raw_pointer_cast
        • thrust::raw_reference_cast
        • thrust::get_temporary_buffer
      • Numerics
        • Complex Numbers
        • thrust::operator<<
        • thrust::sqrt
        • thrust::sinh
        • thrust::asin
        • thrust::cosh
        • thrust::acosh
        • thrust::abs
        • thrust::log10
        • thrust::atanh
        • thrust::cos
        • thrust::exp
        • thrust::asinh
        • thrust::pow
        • thrust::norm
        • thrust::operator/
        • thrust::arg
        • thrust::log
        • thrust::operator*
        • thrust::proj
        • thrust::tan
        • thrust::operator>>
        • thrust::conj
        • thrust::polar
        • thrust::operator==
        • thrust::operator-
        • thrust::acos
        • thrust::operator!=
        • thrust::tanh
        • thrust::atan
        • thrust::sin
        • thrust::operator+
      • Parallel Execution Policies
        • Parallel Execution Policies
        • thrust::device
        • thrust::host
      • Random Number Generators
        • Random Number Engines with Predefined Parameters
        • Random Number Distributions
        • Random Number Engines
        • Random Number Engine Adaptors
      • System
        • Systems
        • System Diagnostics
      • Utility
        • Pair
        • Tuple
        • Type Traits
      • API reference
      • Developer Overview
        • Developer CMake Options
        • Thrust systems
    • CUDA Experimental
      • Overview
      • Containers library
        • cuda::experimental::uninitialized_buffer
      • Graphs library
        • cuda::experimental::graph
        • cuda::experimental::graph_builder
        • cuda::experimental::graph_builder_ref
        • cuda::experimental::graph_memory_resource
        • cuda::experimental::graph_node_ref
        • cuda::experimental::stf::graphed_interface_of
        • cuda::experimental::stf::graphed_interface_of< mdspan< T, P… > >
        • cuda::experimental::stf::graphed_interface_of< scalar_view< T > >
        • cuda::experimental::stf::graphed_interface_of< void_interface >
      • Places
      • CUDASTF
        • Implementation of the matrix class
        • Lower-level API
      • API reference
    • Tensor Memory Accelerator (TMA)
    • Determinism
    • CCCL configuration macros
    • CCCL 2.x ‐ CCCL 3.0 migration guide
    • CCCL Development Guide
      • CCCL Internal Macros
      • CCCL Testing Utilities
      • Build and Bisect Utilities
      • Symbol Visibility
        • Host Stub Visibility Issue
        • Device Kernel Visibility Issue
        • Linking TUs compiled with different architectures
      • General Debugging
    • Contributing to the CUDA Core Compute Libraries
      • Code of Conduct
    • License
  • CCCL Python Libraries
    • Setup and Installation
    • cuda.compute: Parallel Computing Primitives
      • Overview
      • cuda.compute Developer Overview
    • cuda.stf._experimental: Sequential Task Flow
    • Resources
    • API Reference
      • cuda.compute API Reference
      • cuda.stf._experimental API Reference
  • Contributor Docs
    • How Tos
      • Install, Build, Test
      • Build and run targeted tests
      • Launching a container
      • Reproducing CI locally
      • Override matrix
      • Adding a build target
  • Maintainer Docs
    • How Tos
      • How to commit fixes to release branches (backport process)
      • Run compile-time benchmarks
      • How To Plan Work with GitHub Issues
      • Adding CI coverage
      • Adding a new devcontainer toolchain
    • References
      • Git Methodology
      • CodeRabbit
      • Compile-time benchmark reference
  • Infrastructure
    • CI Reference
      • CI overview
      • Change detection
      • matrix.yaml reference
      • CI scripts
      • Artifact system
      • GitHub Actions workflows
      • Skip tags
    • CMake Reference
      • Preset reference
      • CMake helper modules
      • Architecture flags
    • Devcontainer Reference
      • Devcontainer overview
      • launch.sh reference
  • CUDA C++ Core Libraries
  • License

License#

The CUDA Core Compute Libraries use a mix of several licenses, mostly for historical reasons.

libcu++ uses the Apache License v2.0 with LLVM Exceptions. CUB mostly uses the 3-Clause BSD License. Thrust mostly the Apache License v2.0.

Newly developed code will be under the Apache License v2.0 with LLVM Exceptions.

previous

Code of Conduct

next

CCCL Python Libraries

Copyright © 2026, NVIDIA Corporation.

Created using Sphinx 9.1.0.