GoKawiilLwan author details integer-to-string optimizations in 2014 blog post
The 2014 post traces how the Lwan web server's integer-to-string conversion evolved from using snprintf() to a naive divide-and-modulus algorithm that built digits in reverse and then flipped the string. The author then replaced it with a version that writes digits directly into a buffer from the end backwards, avoiding the separate string-reversal step and using a buffer sized via a trick based on multiplying sizeof(int) by three.
GoKawiil's interpretation of the reporting above, not reported fact.
Although integer-to-string conversion is rarely a performance bottleneck on its own, the author notes it surfaced repeatedly during profiling of Lwan's HTTP header generation, suggesting such micro-optimizations can matter in high-throughput server code. The post illustrates a broader lesson for systems programmers: small algorithmic tweaks, like eliminating an unnecessary reversal pass, can yield measurable speedups in hot code paths.
- Lwan originally used snprintf() for integer-to-string conversion before moving to faster custom algorithms.
- A naive divide-by-10 algorithm required reversing the resulting string, which the author found inefficient.
- A revised version writes digits backwards directly into a buffer, removing the reversal step and improving speed.
Source: tia.mat.br, 2026-10-18
Published there as: “Integer to String Conversion (2014)”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.