Recently, the two major Web engines (WebKit and Chromium) adopted fast SIMD routines to scan HTML content. The key insight is to use vectorized classification (Langdale and Lemire, 2019): you load blocks of characters and identify the characters you seek…