go-ndarray documentation

A pure-Go (no cgo) NumPy-style n-dimensional array for float64 — the numpy equivalent for Go. Creation routines, strided views that share data, broadcasting elementwise ops and ufuncs, reductions, manipulation and linear algebra, all with cgo disabled.

Ruby has no cgo-free ndarray (Numo::NArray, NMatrix are C extensions) and gonum's optimized assembly is amd64-only. go-ndarray pairs a portable scalar core with multicore fan-out + go-asmgen SIMD, beating single-threaded NumPy on the parallelizable core. 100% coverage, differentially checked against numpy 2.2.

import nd "github.com/go-ndarray/ndarray"

a := nd.Arange(0, 6).Reshape(2, 3)
a.Sum()              // 15
a.Transpose().Shape() // [3 2]

API surface

Area	Functions / methods
Creation	Zeros, Ones, Full, Arange, Linspace, Eye, FromData
Views & shape	Slice (All/R/Rng/From/To/Step), Reshape, Ravel, Transpose, Copy
Elementwise	Add/Sub/Mul/Div (+*Scalar, +*Into), Map, Neg, Abs
Ufuncs	Sqrt, Exp, Log/Log2/Log10, Sin/Cos/Tan, Floor/Ceil/Round, Square, Power
Reductions	Sum, Mean, Max/Min, Prod, ArgMax/ArgMin, CumSum/CumProd, Clip, Where (+ per-axis)
Manipulation	Flatten, ExpandDims, Squeeze, Concatenate, Stack, VStack, HStack
Linear algebra	MatMul, Dot, Inner, Outer

Performance & architectures

The hot paths are multicore + SIMD: a go-asmgen sum kernel and a panel-packed, cache-blocked GEMM with a SIMD-FMA micro-kernel (amd64 SSE2, arm64 NEON), plus packed elementwise / sqrt / max-min kernels. The result beats single-threaded NumPy on Add/Mul/Sum/Sqrt/Max and the blocked GEMM. MatMul reaches tuned-BLAS parity at 1024² (≈1.00× of single-threaded vecLib, ~373 GFLOP/s; ≈0.99× vs multi-threaded OpenBLAS) and beats the pure-Go gonum 4–10× at every size, while Dot (1-D) wins at parity (~0.98×). Where tuned BLAS still leads at small n, the benchmark page says so.

Where to go next

• Roadmap — the plan and what ships today.
• Performance — honest benchmarks versus NumPy 2.2 / OpenBLAS.

Source: github.com/go-ndarray/ndarray · also the cgo-free ndarray backend behind go-embedded-ruby's NDArray class.