Transcoding billions of Unicode characters per second with SIMD instructions. (13th October 2021)

• Record Type: Journal Article
• Authors: Lemire, Daniel; Muła, Wojciech
• Is Part Of: Software, practice & experience. Volume 52:Number 2(2022)
• Page Start: 555
• Page End: 575
• Publication Date: 2021-10-13
• Subjects: character encoding -- text processing -- Unicode -- vectorization; Computer software -- Periodicals; Computer programming -- Periodicals; Computer programs -- Periodicals; 005.3
• DOI: 10.1002/spe.3036
• Languages: English
• ISSNs: 0038-0644

Abstract

Abstract: In software, text is often represented using Unicode formats (UTF‐8 and UTF‐16). We frequently have to convert text from one format to the other, a process called transcoding. Popular transcoding functions are slower than state‐of‐the‐art disks and networks. These transcoding functions make little use of the single‐instruction‐multiple‐data (SIMD) instructions available on commodity processors. By designing transcoding algorithms for SIMD instructions, we multiply the speed of transcoding on current systems ( × 64 and ARM). To ensure reproducibility, we make our software freely available as an open source library.