On Friday, 30 August 2019 11:27:00 PDT Jakub Narolewski wrote: > As for compression performance, zstd is still not as fast as lz4's > LZ4_compress_fast - at least not in my use case.
Indeed. Your data shows you're compressing less than 1.5k of data, which incidentally fits into a single TCP or UDP frame. So the gain in compression at level 0 (234 bytes versus 520) is not worth the time it takes, since it won't be transmitted faster. Zstd's big avantage is the search algorithm. Since your data is so small, it's no surprise it's no better than LZ4. That result is the opposite of what I found when replacing Icecream's compression: the data transferred is in the order of megabytes, so the network time is the biggest contributor. But thanks for the info, it's good to know. PS: Have you tried LZO? That's what Icecream used before I changed. -- Thiago Macieira - thiago.macieira (AT) intel.com Software Architect - Intel System Software Products _______________________________________________ Interest mailing list Interest@qt-project.org https://lists.qt-project.org/listinfo/interest
