xtensor-stack/xsimd

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

2740 stars · 306 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

xsimd is a C++ library providing unified wrappers over SIMD intrinsics (SSE, AVX, AVX512, NEON, SVE, WebAssembly, VSX, RISC-V) so the same code runs across architectures. It lets you manipulate batches of numbers with standard arithmetic operators and includes optimized parallelized mathematical functions.

Use cases

• write portable SIMD-accelerated numeric code across x86 and ARM
• speed up math-heavy loops with vectorized batches
• use SIMD intrinsics without writing per-architecture code
• accelerate mathematical functions like exp, sin, log on batches
• target WebAssembly SIMD from C++
• build a numeric library that supports multiple SIMD ISAs

When to choose

• you need explicit SIMD vectorization with a portable C++ API
• you want to support many ISAs (x86, ARM, RISC-V, WASM) from one codebase
• you are writing performance-critical math or signal processing code in C++

When to avoid

• you want the compiler to auto-vectorize without explicit SIMD code
• you need GPU acceleration rather than CPU SIMD
• your project is not C++17-capable

Facets

library · maturity stable

math graphics developer-tools performance developer-tools cpp cross-platform wasm simd vectorization sse avx avx512 neon sve risc-v wasm vsx cpp17 header-only mathematical-functions algorithms gpu

Member repositories

For agents

markdown · JSON · MCP: product_card(name="xtensor-stack/xsimd")