A community-oriented list of libraries, tools, and resources aimed to assist with NUMA-aware software development.

Non-Uniform Memory Access (NUMA) refers to multiprocessor systems whose memory is divided into multiple memory nodes.

To add, remove or change things on the list: please submit a pull request to the GitHub repository.

• libnuma - The libnuma shared library to control NUMA policy for processes or shared memory on Linux
  • numa-rs - Rust bindings for libnuma
  • go-numa - Go bindings for libnuma
  • jnuma - Java bindings for libnuma
  • py-numa - Python bindings for libnuma
  • ocaml-numa - OCAML bindings for libnuma
• libNumaAPI - A cross-platform API wrapper for NUMA architecture
• hwloc - A portable API to detect and exploit the topology of parallel architectures
• Windows NUMA API - Official Microsoft Win32 NUMA API

• numactl - A program to run other programs with a specific NUMA policy
• numatool - A wrapper to linux kernels move_pages system call to balance processes' pages across NUMA nodes
• libtorque - A threaded, continuations-based I/O event library for manycore NUMA machines
• pgasus - A C++ parallel programming framework for NUMA systems, based on PGAS semantics
• NUMASK - A skip list designed to exploit the characteristics of NUMA architectures to improve performance
• golang-numa - A golang utility library for NUMA-aware code
• NumaAllocator - A simple C++ header NUMA memory allocator
• numanji - A rustlang Local-affinity first NUMA-aware allocator with optional fallback
• Tesson - A tool for NUMA-aware sharding with Docker
• Umpire - An application-focused API for memory management on NUMA & GPU architectures
• memkind - A heap manager which enables control of memory characteristics and a partitioning of the heap between kinds of memory

• numastat - A program display NUMA allocation statistics
• NUMACC - A golang-based tool to check CPU affinity and NUMA configuration for containers and pods
• NumaTOP - An observation tool for runtime memory locality characterization and analysis of processes and threads running on a NUMA system
• SnuMAP - A NUMA performance profiler
• irqstat - A /proc/interrupts watcher designed for NUMA systems
• Numa-Trace - A Pin tool to track NUMA memory accesses
• NUMAPROF - A NUMA memory profliler based on Pintool to track remote memory accesses
• Numalize - A memory tracing tool to detect communication and page usage of NUMA applications
• NumaMMA - A lightweight NUMA memory profiler/analyzer and a NUMA application execution engin
• numap - A Linux library for memory profiling based on hardware performance monitoring unit
• pcm-numa- A tool to monitor local and remote memory accesses on a NUMA system

• Comm|Scope - A NUMA-aware multi-CPU multi-GPU CUDA data transfer benchmarks
• Numafac - Scripts to calculate the NUMA factor of NUMA machines, based on the stream and lmbench3 benchmarks
• nurdma - Understanding RDMA behavior in NUMA systems
• numabench - A NUMA benchmarking tool
• numa-bench - A NUMA-aware memory allocation benchmark

• Skyhooks - An experimental NUMA-aware, lock-free heap memory allocator
• rw-numa-locks - A NUMA-aware reader-writer locks
• NUMA_Black-Box - Black-box concurrent data structures for NUMA architectures

This list was compiled by Domagoj Margan with help and resources from the systems community.