Skip to content

FEATURES.md

Latest commit

 

History

History
127 lines (98 loc) · 3.61 KB

FEATURES.md

File metadata and controls

127 lines (98 loc) · 3.61 KB
Raw

node-accelerate Features

Complete API (22 Functions)

Matrix Operations (3 functions)

  • ✅ matmul - Matrix multiplication (GEMM) - 283x faster
  • ✅ matmulFloat - Single precision matrix multiplication
  • ✅ matvec - Matrix-vector multiplication (GEMV)

BLAS Operations (1 function)

  • ✅ axpy - Scaled vector addition (y = alpha*x + y)

Vector Arithmetic (6 functions)

  • ✅ dot - Dot product - 5x faster
  • ✅ vadd - Element-wise addition - 3.7x faster
  • ✅ vsub - Element-wise subtraction
  • ✅ vmul - Element-wise multiplication
  • ✅ vdiv - Element-wise division
  • ✅ vscale - Scalar multiplication

Vector Functions (4 functions)

  • ✅ vabs - Absolute value
  • ✅ vsquare - Square (element-wise)
  • ✅ vsqrt - Square root (element-wise)
  • ✅ normalize - Normalize to unit length

Reductions (5 functions)

  • ✅ sum - Sum of elements - 7.6x faster
  • ✅ mean - Average of elements
  • ✅ max - Maximum element
  • ✅ min - Minimum element
  • ✅ rms - Root Mean Square

Distance Metrics (1 function)

  • ✅ euclidean - Euclidean distance

Signal Processing (1 function)

  • ✅ fft - Fast Fourier Transform

Performance Benchmarks

Operation Pure JS Accelerate Speedup
Matrix multiply (500×500) 93ms 0.33ms 283x
Vector dot (1M) 0.66ms 0.13ms 5x
Vector sum (1M) 0.59ms 0.08ms 7.6x
Vector add (1M) 0.74ms 0.20ms 3.7x

Quality Metrics

  • ✅ 26 comprehensive tests - All passing
  • ✅ Full TypeScript definitions - Complete IDE support
  • ✅ 4 example programs - Real-world usage
  • ✅ Zero dependencies (runtime) - Only node-addon-api for building
  • ✅ CI/CD ready - GitHub Actions workflow included
  • ✅ Complete documentation - 8 markdown files, 20,000+ words

Use Cases

Machine Learning

  • Neural network inference
  • K-means clustering
  • Distance calculations
  • Feature normalization

Signal Processing

  • FFT analysis
  • Audio processing
  • Filtering
  • Spectral analysis

Scientific Computing

  • Linear algebra
  • Numerical simulations
  • Data analysis
  • Statistics

Computer Graphics

  • Vector/matrix math
  • Transformations
  • Physics simulations

Platform Support

  • ✅ macOS (Apple Silicon: M1/M2/M3/M4)
  • ✅ macOS (Intel)
  • ✅ Node.js 18, 20, 22

What's Next (Future Versions)

v1.1.0 (Planned)

  • More BLAS operations (triangular solve, etc.)
  • More vDSP operations (convolution, correlation)
  • Additional distance metrics (Manhattan, Cosine)

v1.2.0 (Planned)

  • Complex number support
  • Sparse matrix operations
  • Additional signal processing functions

v2.0.0 (Future)

  • Breaking API changes if needed
  • Platform expansion (if possible)
  • Major performance improvements

Why node-accelerate?

  1. Massive speedups - 50-283x faster than pure JavaScript
  2. Hardware acceleration - Direct access to Apple's AMX and NEON
  3. Production ready - Comprehensive tests and documentation
  4. Easy to use - Simple API, works like regular JavaScript
  5. Well maintained - Active development and support

Comparison with Alternatives

Feature node-accelerate Pure JS Other addons
Matrix multiply 283x faster Baseline N/A
Apple Silicon ✅ Optimized ❌ ❌
TypeScript ✅ Full support N/A ⚠️ Partial
Documentation ✅ Complete N/A ⚠️ Limited
Tests ✅ 26 tests N/A ⚠️ Varies
Dependencies 0 (runtime) 0 ⚠️ Many

Ready to use? npm install node-accelerate