p12tic/libsimdpp

Health v2 · maintenance only

Adoption not part of the score

1303 stars · 129 forks observed · 2026-08-28

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

A portable, header-only, zero-overhead C++ SIMD library providing a single interface over SIMD instruction sets across x86, ARM, PowerPC, and MIPS architectures. It supports compiling the same source for multiple instruction sets with runtime dynamic dispatch to select the most efficient implementation.

Use cases

• write portable SIMD code across x86 and ARM
• runtime-dispatch between SSE, AVX2, and AVX512 code paths
• vectorize image or signal processing kernels
• avoid writing raw SIMD intrinsics per architecture
• compile one codebase for NEON and Altivec targets

When to choose

• you need fine-grained control and predictable generated code close to intrinsics
• you must target multiple CPU architectures and instruction sets from one source
• you want a header-only dependency with zero overhead

When to avoid

• you want a high-level abstraction that auto-vectorizes for you
• you need GPU or heterogeneous compute
• you only target a single instruction set and can use intrinsics directly

Facets

library · maturity stable

sdk developer-tools developer-tools performance cross-platform cpp windows cross-platform simd header-only sse avx2 avx512 neon altivec msa intrinsics runtime-dispatch linux macos

Member repositories

For agents

markdown · JSON · MCP: product_card(name="p12tic/libsimdpp")