NumRS2 - High-Performance Numerical Computing for Rust

NumRS2 is a high-performance numerical computing library for Rust, designed as a Rust-native alternative to NumPy. It provides N-dimensional arrays, linear algebra operations, and comprehensive mathematical functions with a focus on performance, safety, and ease of use.

Version 0.4.1 is a production-hardening pass: Array<T> is now Arc-backed copy-on-write (O(1) clone); a new kernels dispatch layer backs matmul, elementwise ops, and reductions (matmul up to ~18.8x faster at 512^3 than the prior loop); the distributed feature's collectives and TSQR/Cholesky linear algebra now run over a real TCP transport instead of returning fabricated local data; lapack (core linear algebra) is a default feature; and a large batch of NumPy-parity additions landed (ufunc reduce/accumulate/at, N-D FFT wrappers, 9 NumPy>=1.22 quantile methods, masked-array completion, polynomial classes, SeedSequence/Philox/SFC64 random generators). See CHANGELOG.md for the full, verified list, including a Known Upstream Issues section for the scirs2-core/scirs2-fft bugs this release works around rather than silently inherits. The core array, linear algebra, SIMD, and autodiff paths are production-ready; some advanced modules (e.g. quantum and parts of the reinforcement-learning suite) remain experimental, and the distributed feature's DistributedArray-based linear algebra convenience functions are intentionally unimplemented in favor of the real DistributedMatrix-based versions (see CHANGELOG.md).

✨ Architecture Highlights

🏗️ Enhanced Design

Trait-based architecture for extensibility and generic programming
Hierarchical error system with rich context and recovery suggestions
Memory management with pluggable allocators (Arena, Pool, NUMA-aware)
Comprehensive documentation with migration guides and best practices

🔧 Core Features

N-dimensional arrays with efficient memory layout, broadcasting, and Arc-backed copy-on-write cloning (O(1) Clone; the first mutation after a clone unshares)
Advanced linear algebra with BLAS/LAPACK integration and matrix decompositions (enabled by default via the lapack feature)
SIMD optimization with automatic vectorization and CPU feature detection, backed by a shared, dtype-dispatched kernels layer for matmul/elementwise/reduction hot paths
Thread safety with parallel processing support via scirs2_core::parallel_ops (COOLJAPAN policy: no direct rayon dependency)
Distributed computing (distributed feature) over a real point-to-point TCP transport: collectives (broadcast/reduce/allreduce/gather/scatter/...), Tall-Skinny QR, and block-cyclic Cholesky, tested with an in-process multi-rank harness
Python interoperability for easy migration from NumPy (python feature, maturin wheels)

Main Features

N-dimensional Array: Core Array type with efficient memory layout and NumPy-compatible broadcasting
Advanced Linear Algebra:
- Matrix operations, decompositions, solvers through BLAS/LAPACK integration
- Sparse matrices (COO, CSR, CSC, DIA formats) with format conversions
- Iterative solvers (CG, GMRES, BiCGSTAB) for large systems
- Randomized algorithms (randomized SVD, random projections, range finders)
Numerical Optimization: BFGS, L-BFGS, Trust Region, Nelder-Mead, Levenberg-Marquardt, constrained optimization
Root-Finding: Bisection, Brent, Newton-Raphson, Secant, Halley, fixed-point iteration
Numerical Differentiation: Gradient, Jacobian, Hessian with Richardson extrapolation
Automatic Differentiation: Forward and reverse mode AD with higher-order derivatives
Data Interoperability:
- Apache Arrow integration for zero-copy data exchange
- Feather format support for fast columnar storage
- IPC streaming for inter-process communication
- Python bindings via PyO3 for NumPy compatibility
Expression Templates: Lazy evaluation and operation fusion for performance
Advanced Indexing: Fancy indexing, boolean masking, and conditional selection
Polynomial Functions: Interpolation, evaluation, and arithmetic operations
Fast Fourier Transform: Optimized FFT implementation with 1D/2D transforms, real FFT specialization, frequency shifting, and various windowing functions
SIMD Acceleration: Enhanced vectorized operations via SciRS2-Core with AVX2/AVX512/NEON support
Parallel Computing: Advanced multi-threaded execution with adaptive chunking and work-stealing
GPU Acceleration: Optional GPU-accelerated array operations via the WGPU backend (cross-platform: runs on Vulkan, Metal, DirectX 12, and WebGPU)
Mathematical Functions: Comprehensive set of element-wise mathematical operations
Statistical Analysis: Descriptive statistics, probability distributions, and more
Random Number Generation: Modern interface for various distributions with fast generation and NumPy-compatible API
SciRS2 Integration: Integration with SciRS2 for advanced statistical distributions and scientific computing functionality
Fully Type-Safe: Leverage Rust's type system for compile-time guarantees

Optional Features

The default feature set is ["matrix_decomp", "scirs", "lapack"]. NumRS2 includes the following features, all definable in your Cargo.toml (see Cargo.toml's [features] table for the authoritative list):

Enabled by default

matrix_decomp: Matrix decomposition functions (SVD, QR, LU, etc.)
scirs: Mandatory SciRS2 ecosystem integration; kept only for backward-compatible opt-out syntax and must not be disabled
lapack: LAPACK-dependent linear algebra (det/inv/svd/eig/qr/cholesky), through scirs2-linalg (OxiBLAS-backed, pure Rust)

Off by default

validation: Additional runtime validation checks for array operations
unstable / fast: Development-time feature combinations (minimal build for fast iteration)
gpu: GPU acceleration for array operations using WGPU (Vulkan/Metal/DX12/WebGPU; no native CUDA/ROCm)
python: Python bindings via PyO3 for NumPy interoperability (wheel builds via maturin)
distributed: Pure-Rust distributed computing (no MPI) — real point-to-point TCP transport, collectives, TSQR, block-cyclic Cholesky
wasm: WebAssembly bindings (dlmalloc global allocator on wasm32 targets)
arrow: Apache Arrow integration for zero-copy data exchange with Python/Polars/DataFusion
parquet / netcdf / matlab / messagepack / bson: additional pure-Rust I/O formats (or io-all for all five at once)
visualization: Plotters-based 2D/3D/matrix/stats/performance plots (pure Rust, via plotters's own bitmap/SVG backends and the oxifont-bundled COOLJAPAN font)
ci-safe: matrix_decomp + validation only, for CI environments without external dependencies

To enable a feature:

[dependencies]
numrs2 = { version = "0.4.1", features = ["distributed"] }

Or, when building:

cargo build --features distributed
cargo build --all-features

🚀 Performance Optimizations

NumRS2 leverages SciRS2-Core (v0.6.5) for cutting-edge performance optimizations:

Unified SIMD Operations: All SIMD code goes through SciRS2-Core's SimdUnifiedOps trait
Adaptive Algorithm Selection: AutoOptimizer automatically chooses between scalar, SIMD, or GPU implementations
Platform Detection: Automatic detection of AVX2, AVX512, NEON, and GPU capabilities
Parallel Operations: Optimized parallel processing with intelligent work distribution
Memory-Efficient Chunking: Process large datasets without memory bottlenecks

See the optimization example for usage details.

SciRS2 Integration

The SciRS2 integration provides additional advanced statistical distributions:

Noncentral Chi-square: Extends the standard chi-square with a noncentrality parameter
Noncentral F: Extends the standard F distribution with a noncentrality parameter
Von Mises: Circular normal distribution for directional statistics
Maxwell-Boltzmann: Used for modeling particle velocities in physics
Truncated Normal: Normal distribution with bounded support
Multivariate Normal with Rotation: Allows rotation of the coordinate system

For examples, see scirs_integration_example.rs

GPU Acceleration

The GPU acceleration feature provides:

GPU-accelerated array operations for significant performance improvements
Seamless CPU/GPU interoperability with the same API
Support for various operations: arithmetic, matrix multiplication, element-wise functions, etc.
GPU acceleration via the WGPU backend (cross-platform: runs on Vulkan, Metal, DirectX 12, and WebGPU). NumRS2 does not ship native CUDA/ROCm/Metal/OpenCL backends; all GPU execution goes through WGPU.

For examples, see gpu_example.rs

🎯 Key Features

Numerical Optimization (scipy.optimize equivalent)

BFGS & L-BFGS: Quasi-Newton methods for large-scale optimization
Trust Region: Robust optimization with dogleg path
Nelder-Mead: Derivative-free simplex method
Levenberg-Marquardt: Nonlinear least squares
Constrained optimization: Projected gradient, penalty methods

Root-Finding Algorithms (scipy.optimize.root_scalar)

Bracketing methods: Bisection, Brent, Ridder, Illinois
Open methods: Newton-Raphson, Secant, Halley
Fixed-point iteration for implicit equations

Numerical Differentiation

Gradient, Jacobian, and Hessian computation
Forward, backward, central differences
Richardson extrapolation for high accuracy

SIMD & Dispatch Infrastructure

A shared, dtype-dispatched kernels layer (src/kernels/) backs matmul, elementwise ops, and reductions with explicit SIMD/parallel thresholds and deterministic accumulation order, replacing a set of previously per-call-site, hand-rolled dispatch decisions
AVX2/AVX512 (x86_64) and ARM NEON vectorization via scirs2-core, with automatic threshold-based dispatch between scalar, SIMD, and parallel implementations
Support for both f32 and f64 numeric types

Production-Hardening Status (0.4.1)

Array<T> is Arc-backed copy-on-write: O(1) Clone, unshare-on-first-mutation, enforced to exactly one Arc::make_mut / Arc::try_unwrap call site by a CI policy check (scripts/ci-local.sh)
No production unwrap(), zero clippy warnings under the current lint configuration (cargo clippy --all-targets)
~262,000 lines of Rust across 529 files in src/ (tokei src); see CHANGELOG.md for the per-release test-count baseline (nextest + doctests)
Core array, linear algebra, SIMD, autodiff, and (as of this release) distributed transport paths are production-hardened; some advanced modules (e.g. quantum and parts of the reinforcement-learning suite) remain experimental — see TODO.md for the current, short list of known deferred items and NumPy deviations

Enhanced Modules

Linear algebra: Extended iterative solvers (CG, GMRES, BiCGSTAB, FGMRES, MINRES), plus distributed TSQR / block-cyclic Cholesky under the distributed feature
Statistics: quantile/percentile (9 NumPy>=1.22 methods + 4 legacy), histogramdd with density=, masked-array reductions
Polynomial operations: Chebyshev/Legendre/Hermite/HermiteE/Laguerre/Power/Series classes plus NumPy polynomial-module compatibility
Special functions: Spherical harmonics, Jacobi elliptic, Lambert W, and more

Example

use numrs2::prelude::*;

fn main() -> Result<()> {
    // Create arrays
    let a = Array::from_vec(vec![1.0, 2.0, 3.0, 4.0]).reshape(&[2, 2]);
    let b = Array::from_vec(vec![5.0, 6.0, 7.0, 8.0]).reshape(&[2, 2]);
    
    // Basic operations with broadcasting
    let c = a.add(&b);
    let d = a.multiply_broadcast(&b)?;
    
    // Matrix multiplication
    let e = a.matmul(&b)?;
    println!("a @ b = {}", e);
    
    // Linear algebra operations
    let (u, s, vt) = a.svd_compute()?;
    println!("SVD components: U = {}, S = {}, Vt = {}", u, s, vt);
    
    // Eigenvalues and eigenvectors
    let symmetric = Array::from_vec(vec![2.0, 1.0, 1.0, 2.0]).reshape(&[2, 2]);
    let (eigenvalues, eigenvectors) = symmetric.eigh("lower")?;
    println!("Eigenvalues: {}", eigenvalues);
    
    // Polynomial interpolation
    let x = Array::linspace(0.0, 1.0, 5)?;
    let y = Array::from_vec(vec![0.0, 0.1, 0.4, 0.9, 1.6]);
    let poly = PolynomialInterpolation::lagrange(&x, &y)?;
    println!("Interpolated value at 0.5: {}", poly.evaluate(0.5));
    
    // FFT operations
    let signal = Array::from_vec(vec![1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0]);
    // Window the signal before transforming
    let windowed_signal = signal.apply_window("hann")?;
    // Compute FFT
    let spectrum = windowed_signal.fft()?;
    // Shift frequencies to center the spectrum
    let centered = spectrum.fftshift_complex()?;
    println!("FFT magnitude: {}", spectrum.power_spectrum()?);
    
    // Statistical operations
    let data = Array::from_vec(vec![1.0, 2.0, 3.0, 4.0, 5.0]);
    println!("mean = {}", data.mean()?);
    println!("std = {}", data.std()?);
    
    // Sparse array operations
    let mut sparse = SparseArray::new(&[10, 10]);
    sparse.set(&[0, 0], 1.0)?;
    sparse.set(&[5, 5], 2.0)?;
    println!("Density: {}", sparse.density());
    
    // SIMD-accelerated operations
    let result = numrs2::simd::simd_sqrt(&data);
    println!("SIMD result: {}", result);

    // Random number generation
    let rng = random::default_rng();
    let uniform = rng.random::<f64>(&[3])?;
    let normal = rng.normal(0.0, 1.0, &[3])?;
    println!("Random uniform [0,1): {}", uniform);
    println!("Random normal: {}", normal);

    Ok(())
}

Performance

NumRS is designed with performance as a primary goal:

Rust's Zero-Cost Abstractions: Compile-time optimization without runtime overhead
BLAS/LAPACK Integration: Industry-standard libraries for linear algebra operations
SIMD Vectorization: Parallel processing at the CPU instruction level with automatic CPU feature detection
Memory Layout Optimization: Cache-friendly data structures and memory alignment
Data Placement Strategies: Optimized memory placement for better cache utilization
Adaptive Parallelization: Smart thresholds to determine when parallel execution is beneficial
Scheduling Optimization: Intelligent selection of work scheduling strategies based on workload
Fine-grained Parallelism: Advanced workload partitioning for better load balancing
Modern Random Generation: Advanced thread-safe RNG with PCG64 algorithm for high-quality randomness

Expression Templates

NumRS2 provides a powerful expression templates system for lazy evaluation and performance optimization:

SharedArray - Reference-Counted Arrays

use numrs2::prelude::*;

// Create shared arrays with natural operator syntax
let a: SharedArray<f64> = SharedArray::from_vec(vec![1.0, 2.0, 3.0, 4.0]);
let b: SharedArray<f64> = SharedArray::from_vec(vec![10.0, 20.0, 30.0, 40.0]);

// Cheap cloning (O(1) - just increments reference count)
let a_clone = a.clone();

// Natural operator overloading
let sum = a.clone() + b.clone();         // [11.0, 22.0, 33.0, 44.0]
let product = a.clone() * b.clone();     // [10.0, 40.0, 90.0, 160.0]
let scaled = a.clone() * 2.0;            // [2.0, 4.0, 6.0, 8.0]
let result = (a.clone() + b.clone()) * 2.0 - 5.0;  // Chained operations

SharedExpr - Lifetime-Free Lazy Evaluation

use numrs2::expr::{SharedExpr, SharedExprBuilder};

// Build expressions lazily - no computation until eval()
let c: SharedArray<f64> = SharedArray::from_vec(vec![1.0, 2.0, 3.0, 4.0]);
let expr = SharedExprBuilder::from_shared_array(c);
let squared = expr.map(|x| x * x);   // Expression built, not evaluated
let result = squared.eval();         // [1.0, 4.0, 9.0, 16.0] - evaluated here

Common Subexpression Elimination (CSE)

use numrs2::expr::{CachedExpr, ExprCache};

// Automatic caching of repeated computations
let cache: ExprCache<f64> = ExprCache::new();
let cached_expr = CachedExpr::new(sum_expr.into_expr(), cache.clone());

let result1 = cached_expr.eval();  // Computes and caches
let result2 = cached_expr.eval();  // Uses cached result

Memory Access Pattern Optimization

use numrs2::memory_optimize::access_patterns::*;

// Detect memory layout for optimization
let layout = detect_layout(&[100, 100], &[100, 1]);  // CContiguous

// Get optimization hints for array shapes
let hints = OptimizationHints::default_for::<f64>(10000);
println!("Block size: {}", hints.block_size);
println!("Use parallel: {}", hints.use_parallel);

// Cache-aware iteration for large arrays
let block_iter = BlockedIterator::new(10000, 64);
for block in block_iter {
    // Process block.start..block.end with cache efficiency
}

// Cache-aware operations
cache_aware_transform(&src, &mut dst, |x| x * 2.0);
cache_aware_binary_op(&a, &b, &mut result, |x, y| x + y);

See the expression templates example for a comprehensive demonstration.

Installation

Add this to your Cargo.toml:

[dependencies]
numrs2 = "0.4.1"

For BLAS/LAPACK support, ensure you have the necessary system libraries:

Note: NumRS2 uses OxiBLAS, a pure Rust BLAS/LAPACK implementation with no C dependencies. You do NOT need to install system BLAS/LAPACK libraries.

To use LAPACK functionality (pure Rust via OxiBLAS):

cargo build --features lapack
cargo test --features lapack

OxiBLAS provides:

Pure Rust implementation with SIMD optimizations (AVX2/NEON)
No external C dependencies required
80-172% of OpenBLAS performance (competitive or faster on Apple M3)
Complete BLAS Level 1/2/3 and LAPACK operations

Implementation Details

NumRS2 is built on top of the SciRS2 ecosystem and pure Rust libraries:

SciRS2 ecosystem (scirs2-core, scirs2-linalg, scirs2-stats, etc. v0.6.5): Provides the foundation for n-dimensional arrays, linear algebra, statistics, survival analysis, causal inference, bioinformatics, and combinatorics
OxiBLAS (pure Rust BLAS/LAPACK): Powers high-performance linear algebra routines with no C dependencies
Oxicode: Pure Rust serialization for data persistence
scirs2_core::parallel_ops: Rayon-backed parallel computation, reached only through the SciRS2 re-export (COOLJAPAN policy: no direct rayon dependency)
oxiarc-lz4 / oxiarc-archive: Pure Rust compression for the distributed transport and NPZ archives
num-traits / num-complex: Provides generic numeric traits and complex number support for numerical operations

Features

NumRS2 provides a comprehensive suite of numerical computing capabilities:

Core Functionality

N-dimensional arrays with efficient memory layout and broadcasting
Linear algebra operations with BLAS/LAPACK integration
Matrix decompositions (SVD, QR, Cholesky, LU, Schur, COD)
Eigenvalue and eigenvector computation
Mathematical functions with numerical stability optimizations

Performance Optimizations

SIMD acceleration with automatic CPU feature detection
Parallel processing with adaptive scheduling and load balancing
Memory optimization with cache-friendly data structures
Vectorized operations for improved computational efficiency

Advanced Features

Fast Fourier Transform with 1D/2D transforms and windowing functions
Polynomial operations and interpolation methods
Sparse matrix support for memory-efficient computations
Random number generation with multiple distribution support
Statistical analysis functions and descriptive statistics

Integration & Interoperability

GPU acceleration support via WGPU (optional, gpu feature)
Distributed computing over a real TCP transport (optional, distributed feature)
SciRS2 integration — the foundation for arrays, linear algebra, statistics, and FFT (mandatory; the scirs feature flag exists only for backward-compatible syntax and cannot be disabled)
Memory-mapped arrays for large dataset handling
Serialization support for data persistence (NPY/NPZ, Arrow, Parquet, NetCDF-3, MATLAB, MessagePack, BSON — see Optional Features)

📖 Documentation

📚 Comprehensive Guides

Architecture Guide - System design and core concepts
Migration Guide - Upgrading from previous versions
Trait System Guide - Generic programming with NumRS2
Error Handling Guide - Robust error management
Memory Management Guide - Optimizing memory usage
GPU Acceleration Guide - WGPU-backed GPU operations
Python Bindings Guide - PyO3 bindings and maturin wheel builds
WebAssembly Guide - Browser/Node.js WASM builds

🔗 Additional Resources

Official API Documentation - Complete API reference
NumPy Migration Guide - Guide for NumPy users transitioning to NumRS2
Benchmarking Guide - Running and interpreting NumRS2's benchmark suite
Development Roadmap - Current status and known deferred items
Changelog - Full, per-release list of additions, changes, and fixes
Contributing Guide - How to contribute to NumRS2

🔧 Local CI

GitHub Actions is restricted by project policy to publish workflows only (.github/workflows/pypi-publish.yml). There is no hosted CI build; instead, run the same checks CI would locally:

./scripts/ci-local.sh          # fmt-check, build, clippy, test, doctest, deny, wasm-check, policy
./scripts/ci-local.sh clippy deny   # or a subset of steps

This script is the authoritative pre-merge gate — it enforces the no-unwrap() policy, the <2000-line-per-file policy, the Arc copy-on-write invariant (exactly one Arc::make_mut / one Arc::try_unwrap, both in src/array/core.rs), and cargo deny check bans.

Module-specific documentation:

Random Module Guide - Random number generation
Statistics Module Guide - Statistical functions
Linear Algebra Guide - Linear algebra operations
Polynomial Guide - Polynomial operations
FFT Guide - Fast Fourier Transform

Testing Documentation:

Testing Guide - Guide for NumRS testing approach
Property-based testing for mathematical operations
- Property tests for linear algebra operations
- Property tests for special functions
- Statistical validation of random distributions
Reference testing
- Reference tests for random distributions
- Reference tests for linear algebra operations
- Reference tests for special functions
Benchmarking
- Linear algebra benchmarks
- Special functions benchmarks

Examples

Check out the examples/ directory for more usage examples:

basic_usage.rs: Core array operations and manipulations
matrix_decomp_example.rs: Linear algebra decompositions and solvers
simd_example.rs: SIMD-accelerated computations
memory_optimize_example.rs: Memory layout optimization for cache efficiency
parallel_optimize_example.rs: Parallelization optimization techniques
random_distributions_example.rs: Comprehensive examples of random number generation
distributed_basics.rs / distributed_computing.rs: Distributed arrays, collectives, and TSQR under the distributed feature
gpu_acceleration.rs / gpu_batching.rs: GPU-accelerated operations under the gpu feature
See the examples README for the full, current list (62 example programs)

Development

NumRS is in active development. See TODO.md for upcoming features and development roadmap.

Testing

NumRS requires the approx crate for testing. Tests can be run after installation with:

cargo test

For running property-based and statistical tests for the random module:

cargo test --test test_random_statistical
cargo test --test test_random_properties
cargo test --test test_random_advanced

Contributing

NumRS2 is a community-driven project, and we welcome contributions from everyone. There are many ways to contribute:

Code: Implement new features or fix bugs
Documentation: Improve guides, docstrings, or examples
Testing: Write tests or improve existing ones
Reviewing: Review pull requests from other contributors
Performance: Identify bottlenecks or implement optimizations
Examples: Create example code showing library usage

If you're interested in contributing, please read our Contributing Guide for detailed instructions on how to get started.

For significant changes, please open an issue to discuss your ideas first.

Sponsorship

NumRS2 is developed and maintained by COOLJAPAN OU (Team Kitasan).

If you find NumRS2 useful, please consider sponsoring the project to support continued development of the Pure Rust ecosystem.

https://github.com/sponsors/cool-japan

Your sponsorship helps us:

Maintain and improve the COOLJAPAN ecosystem
Keep the entire ecosystem (OxiBLAS, OxiFFT, SciRS2, etc.) 100% Pure Rust
Provide long-term support and security updates

License

This project is licensed under the Apache License 2.0 - see the LICENSE file for details.

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

numrs2-0.4.1.tar.gz (2.9 MB view details)

Uploaded Aug 29, 2026 Source

Built Distributions

If you're not sure about the file name format, learn more about wheel file names.

The dropdown lists show the available interpreters, ABIs, and platforms. Enable javascript to be able to filter the list of wheel files.

numrs2-0.4.1-pp311-pypy311_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (1.1 MB view details)

Uploaded Aug 29, 2026 PyPymanylinux: glibc 2.17+ x86-64

numrs2-0.4.1-pp311-pypy311_pp73-manylinux_2_17_aarch64.manylinux2014_aarch64.whl (1.0 MB view details)

Uploaded Aug 29, 2026 PyPymanylinux: glibc 2.17+ ARM64

numrs2-0.4.1-cp315-cp315t-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (1.1 MB view details)

Uploaded Aug 29, 2026 CPython 3.15tmanylinux: glibc 2.17+ x86-64

numrs2-0.4.1-cp315-cp315-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (1.1 MB view details)

Uploaded Aug 29, 2026 CPython 3.15manylinux: glibc 2.17+ x86-64

numrs2-0.4.1-cp314-cp314t-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (1.1 MB view details)

Uploaded Aug 29, 2026 CPython 3.14tmanylinux: glibc 2.17+ x86-64

numrs2-0.4.1-cp314-cp314t-manylinux_2_17_aarch64.manylinux2014_aarch64.whl (1.0 MB view details)

Uploaded Aug 29, 2026 CPython 3.14tmanylinux: glibc 2.17+ ARM64

numrs2-0.4.1-cp314-cp314-win_amd64.whl (1.0 MB view details)

Uploaded Aug 29, 2026 CPython 3.14Windows x86-64

numrs2-0.4.1-cp314-cp314-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (1.1 MB view details)

Uploaded Aug 29, 2026 CPython 3.14manylinux: glibc 2.17+ x86-64

numrs2-0.4.1-cp314-cp314-manylinux_2_17_aarch64.manylinux2014_aarch64.whl (1.0 MB view details)

Uploaded Aug 29, 2026 CPython 3.14manylinux: glibc 2.17+ ARM64

numrs2-0.4.1-cp314-cp314-macosx_11_0_arm64.whl (980.4 kB view details)

Uploaded Aug 29, 2026 CPython 3.14macOS 11.0+ ARM64

numrs2-0.4.1-cp314-cp314-macosx_10_12_x86_64.whl (1.1 MB view details)

Uploaded Aug 29, 2026 CPython 3.14macOS 10.12+ x86-64

numrs2-0.4.1-cp313-cp313-win_amd64.whl (1.0 MB view details)

Uploaded Aug 29, 2026 CPython 3.13Windows x86-64

numrs2-0.4.1-cp313-cp313-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (1.1 MB view details)

Uploaded Aug 29, 2026 CPython 3.13manylinux: glibc 2.17+ x86-64

numrs2-0.4.1-cp313-cp313-manylinux_2_17_aarch64.manylinux2014_aarch64.whl (1.0 MB view details)

Uploaded Aug 29, 2026 CPython 3.13manylinux: glibc 2.17+ ARM64

numrs2-0.4.1-cp313-cp313-macosx_11_0_arm64.whl (980.8 kB view details)

Uploaded Aug 29, 2026 CPython 3.13macOS 11.0+ ARM64

numrs2-0.4.1-cp313-cp313-macosx_10_12_x86_64.whl (1.1 MB view details)

Uploaded Aug 29, 2026 CPython 3.13macOS 10.12+ x86-64

numrs2-0.4.1-cp312-cp312-win_amd64.whl (1.0 MB view details)

Uploaded Aug 29, 2026 CPython 3.12Windows x86-64

numrs2-0.4.1-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (1.1 MB view details)

Uploaded Aug 29, 2026 CPython 3.12manylinux: glibc 2.17+ x86-64

numrs2-0.4.1-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl (1.0 MB view details)

Uploaded Aug 29, 2026 CPython 3.12manylinux: glibc 2.17+ ARM64

numrs2-0.4.1-cp312-cp312-macosx_11_0_arm64.whl (980.6 kB view details)

Uploaded Aug 29, 2026 CPython 3.12macOS 11.0+ ARM64

numrs2-0.4.1-cp312-cp312-macosx_10_12_x86_64.whl (1.1 MB view details)

Uploaded Aug 29, 2026 CPython 3.12macOS 10.12+ x86-64

numrs2-0.4.1-cp311-cp311-win_amd64.whl (1.0 MB view details)

Uploaded Aug 29, 2026 CPython 3.11Windows x86-64

numrs2-0.4.1-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (1.1 MB view details)

Uploaded Aug 29, 2026 CPython 3.11manylinux: glibc 2.17+ x86-64

numrs2-0.4.1-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl (1.0 MB view details)

Uploaded Aug 29, 2026 CPython 3.11manylinux: glibc 2.17+ ARM64

numrs2-0.4.1-cp311-cp311-macosx_11_0_arm64.whl (987.4 kB view details)

Uploaded Aug 29, 2026 CPython 3.11macOS 11.0+ ARM64

numrs2-0.4.1-cp311-cp311-macosx_10_12_x86_64.whl (1.1 MB view details)

Uploaded Aug 29, 2026 CPython 3.11macOS 10.12+ x86-64

numrs2-0.4.1-cp310-cp310-win_amd64.whl (1.0 MB view details)

Uploaded Aug 29, 2026 CPython 3.10Windows x86-64

numrs2-0.4.1-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (1.1 MB view details)

Uploaded Aug 29, 2026 CPython 3.10manylinux: glibc 2.17+ x86-64

numrs2-0.4.1-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl (1.0 MB view details)

Uploaded Aug 29, 2026 CPython 3.10manylinux: glibc 2.17+ ARM64

numrs2-0.4.1-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (1.1 MB view details)

Uploaded Aug 29, 2026 CPython 3.9manylinux: glibc 2.17+ x86-64

numrs2-0.4.1-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl (1.0 MB view details)

Uploaded Aug 29, 2026 CPython 3.9manylinux: glibc 2.17+ ARM64

numrs2-0.4.1-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl (1.0 MB view details)

Uploaded Aug 29, 2026 CPython 3.8manylinux: glibc 2.17+ ARM64

File details

Details for the file numrs2-0.4.1.tar.gz.

File metadata

Download URL: numrs2-0.4.1.tar.gz
Upload date: Aug 29, 2026
Size: 2.9 MB
Tags: Source
Uploaded using Trusted Publishing? Yes
Uploaded via: twine/7.0.0 CPython/3.13.14

File hashes

Hashes for numrs2-0.4.1.tar.gz
Algorithm	Hash digest
SHA256	`a5915bf54233e18d12b10952d376636b28fd1f406186159432ba75897f88a9c2`
MD5	`0c5d3764453856a52bf16ded37d2c215`
BLAKE2b-256	`4e54ca6d4e446608620b5f0d1055b622536da6676167215944be6d48eb6864bc`

Algorithm	Hash digest
SHA256	`63200d39968e27f29517bc89f45bcd914af62236ab5523071207cf0b162b7648`
MD5	`80afa1082a78791f3e2e94e18122694a`
BLAKE2b-256	`fed6147bc4f202672fc469c92a310d464c8d508529eecac4b7d49bc56ba71529`

Algorithm	Hash digest
SHA256	`a254308cb0268964f4f902855e4bea8f4e1e92d224c6ba54d4e242238f3d825e`
MD5	`563788246a06421c9e723e22b6ccdfbb`
BLAKE2b-256	`6aae095cccf9df6ddaeed0d01d6851d61355f3cccc095d67d196544df031bbc9`

Algorithm	Hash digest
SHA256	`1a1cd640f3a175772a86289e341a6a9389939149465603e473e66ee6175598ad`
MD5	`f6546d0841bf332c80daf029420cbb06`
BLAKE2b-256	`ce91a956595bf8a3f447cf56ce5618972424499ef75d435c3b63d251bb0369b6`

Algorithm	Hash digest
SHA256	`d13b580b4cce6216aad898eaf2921b3ebf7040ed4fcce79ac93ff7d9ce4e01fa`
MD5	`d9ada09ddaf35b5930dc025423a956aa`
BLAKE2b-256	`e928ed37da7640683aa618959f54bf1a3f5e0edbe8ca904b4983c76179d1ea8f`

Algorithm	Hash digest
SHA256	`46287efa25d4609075569e6b2470c69bd88258dc7df982162634018314b9a522`
MD5	`327879aa7a70d8ff551abb4563e53de4`
BLAKE2b-256	`88a1c95fe7b194b27d26102baecac87f193457972f029e5ad0091f1df247c6c0`

Algorithm	Hash digest
SHA256	`8cb889965961ad78d91deba9931138253fa6a04ef7928b28f8e030b0ffa6794e`
MD5	`553c584656482643b3af77aecb0b3d28`
BLAKE2b-256	`dbeb87b3f96466d3bd24ddc1b06d0f086aef72836fc4c5ab121ce5ea3ebe48a6`

Algorithm	Hash digest
SHA256	`9fcf671ab397eabc3db550300c9d6aa15189f80649b487311909caee11a4337d`
MD5	`a6fc92cba8244de80fe2f97cf79b94ca`
BLAKE2b-256	`bfccf8af64bed3f88b0864ed933f3c2bec9b3ec5eb8e57375b1b06457f7098fe`

Algorithm	Hash digest
SHA256	`fbfcc17d2536c3fd6b36708fefd4bc6483b15c9ad4d1d9c392563096b41e0b81`
MD5	`f8ec9e2c4a72c85a57e1304d382c45ae`
BLAKE2b-256	`0c0f8b1b9beeca1d99fefa8f13b11328d54a92f36056e7802541a38bb32e3e5f`

Algorithm	Hash digest
SHA256	`2dfdddb2587e552e690a8d1b47968655c74019deb4ac9f2f691388e200c51c90`
MD5	`79cb940b19a3b6e58328da3113871afa`
BLAKE2b-256	`7f88aa62cd6e3fda61470d9b5e1e195d9522594ead548c983e893e334a65f00c`

Algorithm	Hash digest
SHA256	`aba3b11670e3c8082014559190ce780ddb4bcc3d19c4b1594866b7de02241aeb`
MD5	`e35950dc64e6f6c92dde66ced62ac67b`
BLAKE2b-256	`803fc8b90a3cb2841313e60c8bed6c2e83c49f3eacdbb9b9b7c6954e96774cb5`

Algorithm	Hash digest
SHA256	`1bcef735dfc3f2b5b323ff7ce90211863f5e7b0ab0292b5156940d43a1cf0b30`
MD5	`335bdc504e119f11935f715c2d703516`
BLAKE2b-256	`aecd24fd55cb8bd11bf495d53ccc550ccc19ed70c4b74cec8888a25e895f354b`

Algorithm	Hash digest
SHA256	`37530cb69c93f1d7fba926ea67be105bc7b196b556ca33843c8599e218dfc867`
MD5	`c4ad59898763b4375be1efeb7d365473`
BLAKE2b-256	`8ce53353b413c484863035155c0a574ae186c732e274d25136d80af6d01ce5ee`

Algorithm	Hash digest
SHA256	`fa2bb07e9788c71150b8b39aad6706b7353e90c574c71e3231e99ff8bf102a20`
MD5	`a27d105e7cfebff2431adad0128361bd`
BLAKE2b-256	`244944678f38688d8fbd9a295056dc31ed70901df3cbfb363cb6564d7d10eab4`

Algorithm	Hash digest
SHA256	`28bddd3c351154a18499d62fe218664cec90fd9d90bfd828be5415355ff8768b`
MD5	`4e77a1c7836a1a43369b69bfaea8b9c5`
BLAKE2b-256	`add02292adf7fd577e55089c153e3b0f117dfa648d65ee16bd6bd4d171324a82`

Algorithm	Hash digest
SHA256	`072f8eaeb1159e5bf6bb7bff2169dde8c08305011c88a8090ea9adb5249d10df`
MD5	`7b1da7def7cf51c80b08c6e2ab12d202`
BLAKE2b-256	`efa8d430cf596fc011004b9e73e6e7b221d3bfdec93dcd91224d639fc95e478e`

Algorithm	Hash digest
SHA256	`7d53bccc4df14c0e9a1c10cc36edc1d4d2b0a3631b78278fef404539e0b93aed`
MD5	`3024486af513da4e6363239f680eecfd`
BLAKE2b-256	`fc18a66b583e9a684ae7ccd48c191c34726f51b4fbd6ff51078ab491376984c6`

Algorithm	Hash digest
SHA256	`0489ae134999125f2f178deca240939c30fb1514ba15ee77e7dc10df3a86fb0f`
MD5	`7f641a8cf4fc6e70e02e74b11c1ded99`
BLAKE2b-256	`d8b465957b7db1e0fddf0efca93d89a181cfe22bc3acf19d486924175cde2b82`

Algorithm	Hash digest
SHA256	`93b07d823a0465c52957848bf46abf0f2c6586f1e718bfe5b86d98815ddd64d8`
MD5	`c21ac3e656725140a3e20ef0d9f25bf2`
BLAKE2b-256	`2bd65b057b9ab1748978478ba564d69def582c591e7143f5559e1075b4df8a56`

Algorithm	Hash digest
SHA256	`2beee53ab0cc237df18032371b55dec26a5cc7063c19f3ab6a4225760fb4c72b`
MD5	`ef43508d249fdb3566e7b645677d9eb4`
BLAKE2b-256	`004cd267f610845096d11385dbde77d457f19a065faa28396f0b28c5c38fd4e5`

Algorithm	Hash digest
SHA256	`54ca79ef70f5a46f4eaa3d59896b8b7ae5db0583c1327769826af8ec7668546f`
MD5	`7ca5978cd06312145f2b2d8989c148af`
BLAKE2b-256	`153b5f4c7e39b9ec1565d72e3bd5f67c864541c7e744134afd7cf5342057924c`

numrs2 0.4.1

NumRS2 - High-Performance Numerical Computing for Rust

✨ Architecture Highlights

🏗️ Enhanced Design

🔧 Core Features

Main Features

Optional Features

🚀 Performance Optimizations

SciRS2 Integration

GPU Acceleration

🎯 Key Features

Example

Performance

Expression Templates

SharedArray - Reference-Counted Arrays

SharedExpr - Lifetime-Free Lazy Evaluation

Common Subexpression Elimination (CSE)

Memory Access Pattern Optimization

Installation

Implementation Details

Features

Core Functionality

Performance Optimizations

Advanced Features

Integration & Interoperability

📖 Documentation

📚 Comprehensive Guides

🔗 Additional Resources

🔧 Local CI

Examples

Development

Testing

Contributing

Sponsorship

License

Project links

Key dates

1 maintainer

Credits

GitHub Statistics

License

Requires

Provides Extra

Tags

Classifiers

Download files

Source Distribution

Built Distributions

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

Algorithm	Hash digest
SHA256	`1939b02e75726d2545787bd12505153182546a4d65ad851fb9b23c8df18df9fc`
MD5	`aa950015f40c1df622adee0c6460ce01`
BLAKE2b-256	`2f86c696b877780ba47f1af2b4b47f7f4891245c1e8dd372ec14305243d1351f`

Algorithm	Hash digest
SHA256	`eee56cc668ba3b3c278e498c73a0724f53db5fecbff434906275f04efb1212d2`
MD5	`7b670eb977371cf4035327a4a8cc5704`
BLAKE2b-256	`7174eec567e28e5327ae98eeeb635a277bedf4a7bcf95bdb3cb645f978c009fd`

Algorithm	Hash digest
SHA256	`2b57b4e9b2e80fb4e3377f519e794f7e5bee711369bf8ded2cfb0b53ae44ae34`
MD5	`b2814cb390bc835be62f37f47111351b`
BLAKE2b-256	`b1132e8f5afc0bd063475a048ffdfc928ab1b9d9dcdfa601a7f1867bfc568939`

Algorithm	Hash digest
SHA256	`671d2a4f1fa31bc33b8d5ea45faed52afbd86c5c285b79135e71b03afda28bdf`
MD5	`4a10baf90dd13254a975048282b16f03`
BLAKE2b-256	`2f7ef14dc4a90de6cceb90c22e0d1fc739b58ae07916488156b29c65c0872152`

Algorithm	Hash digest
SHA256	`930e22d7a5679a796b596482518f74115f253657546aaf984f9692e2cce3fcc8`
MD5	`3bd9f8c6fb02388ba1abf04d55af6677`
BLAKE2b-256	`32c86a2383145d648625ea3e22dd50ef4fc3a574300aa18cc102b39dc64ed502`

Algorithm	Hash digest
SHA256	`12ed75238f4298f0f08d7ba44f80f2973f36068a32007684be4a6ef1d1d58eec`
MD5	`cc7e8f7ba72f2b40efa7bf7b6d2912e3`
BLAKE2b-256	`edba49e91f8e66d6629d1ed257c0d0f9c1c171b2ea95c407687e1d73ac274aa0`

Algorithm	Hash digest
SHA256	`816b1e7b11051531f7e0649e309dd348ca649ba61a5b2f62f0dab797e0b533bf`
MD5	`bb880ba9e43a2baa976c543a7e7c3658`
BLAKE2b-256	`ae5cfc83f17b1cbe4258951ab829f7a6e5d037b87eee21e2f61695344ab48823`

Algorithm	Hash digest
SHA256	`35ae42086ead85387438ddc0c9b40370101bb18dd13b76682ef2adfacd84e62c`
MD5	`d3acbd170bd1e9629442c0227036ec09`
BLAKE2b-256	`41f62a8cc8e6f90b824b77418a1d83f2789cf3e27703f511971aa94e42e3c658`

Algorithm	Hash digest
SHA256	`bdba61cc73426e1cd23c47e14ac9e4ddb48992ffa5a2d1a609c261026c7d9ca9`
MD5	`21bf09089a80df4454b0154877c6537b`
BLAKE2b-256	`300b23f714f9191647bea2c2341b58e25176f14d82b6e6d49974571fb6c8a4f5`

Algorithm	Hash digest
SHA256	`abe5832bf92bfc135e31bd5b1e02e35778f314a0f344d68c1415c794da8ec062`
MD5	`9fca2081ccb1da94ab79906dbff14e6c`
BLAKE2b-256	`c43cc6009fac7430adf6ac66f8f3ddc3fc45ce021bf8c0f18b1eeda6c6a5415e`