rlx-fastmath v0.2.20 — Accurate vectorized f32 exp / log / tanh in pure Rust (NEON on aarch64): <= 1.33 ulp over every f32, faster than the platform libm

Sign in to join the discussion — your thebay.events account works here.