Faster Base64 Encoding and Decoding Using AVX2 Instructions
This work addresses performance bottlenecks for web developers handling billions of daily base64 operations, though it is incremental as it applies existing SIMD techniques to a specific domain.
The paper tackled the problem of inefficient base64 encoding and decoding, which is widely used in web development, by achieving speed improvements of approximately 10x for encoding and 7x for decoding compared to state-of-the-art implementations.
Web developers use base64 formats to include images, fonts, sounds and other resources directly inside HTML, JavaScript, JSON and XML files. We estimate that billions of base64 messages are decoded every day. We are motivated to improve the efficiency of base64 encoding and decoding. Compared to state-of-the-art implementations, we multiply the speeds of both the encoding (~10x) and the decoding (~7x). We achieve these good results by using the single-instruction-multiple-data (SIMD) instructions available on recent Intel processors (AVX2). Our accelerated software abides by the specification and reports errors when encountering characters outside of the base64 set. It is available online as free software under a liberal license.