WIP: compress benchmark

Open squarejesse opened this issue 5 months ago • 3 comments

Jul 11 '25 14:07 squarejesse

Benchmark                                    (sampleDataSize)  Mode  Cnt   Score    Error  Units
ZstdImplementationBenchmark.lubenCompress             8388608  avgt    3  14.237 ± 10.332  ms/op
ZstdImplementationBenchmark.lubenDecompress           8388608  avgt    3   4.338 ±  3.174  ms/op
ZstdImplementationBenchmark.okioCompress              8388608  avgt    3  13.639 ±  1.751  ms/op
ZstdImplementationBenchmark.okioDecompress            8388608  avgt    3   4.441 ±  0.432  ms/op

Jul 11 '25 16:07 squarejesse

Is this going to get way better with a dictionary in place?

https://github.com/facebook/zstd.git

I guess that is coming later? https://datatracker.ietf.org/doc/html/rfc8878#name-use-of-dictionaries

Jul 12 '25 09:07 yschimke

Yep, I like dictionaries. Especially for smaller documents. Nothing to do yet though.

Jul 22 '25 03:07 swankjesse