GitHub has dramatically improved its code search performance by case-folding source code at memory speed using branch-free loops and byte-space arithmetic. This technique allows the platform to process every byte of code at rates exceeding 45 GiB/s on just a single core.