Everyone should know SIMD

Logged at IST: 2026-07-23 07:25 IST

What it is: Mitchell Hashimoto linking to his practical SIMD introduction, “Everyone Should Know SIMD,” using a Ghostty loop as the worked example.

Gist: Hashimoto argues that everyday SIMD should not feel scary: the common “process N values at a time” case follows a repeatable five-step shape: broadcast constants, loop one vector-width chunk at a time, do the parallel operation, reduce or store the vector result, then finish with the scalar tail. His Ghostty example turns a scalar codepoint scan into generic Zig vector code, yielding up to 4x, 8x, or 16x lane-level throughput depending on NEON, AVX2, or AVX-512, and about a 5x real end-to-end speedup on his AVX2 desktop.

Newsletter angle: Strong systems and developer-tools item because it makes SIMD approachable as a practical hot-loop pattern, not a specialist-only bag of CPU-specific tricks.