Reranko05 commented on issue #49800: URL: https://github.com/apache/arrow/issues/49800#issuecomment-4276523911
Hi @dmitry-chirkov-dremio, this was mainly for larger inputs, where processing each byte starts to add up. So far I’ve looked at the current implementation to understand the decode path and confirm it’s scalar. I also took a quick look through the repo history and from what I could see, the base64 implementation hasn’t really had performance-focused changes apart from recent correctness fixes and the LUT optimization. I didn’t find anything around SIMD or vectorized decoding for this path. So this was more of an initial observation from working in that area rather than something deeply explored yet. I can dig further if there’s actually meaningful benefit. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
