Repository navigation
Consider adding benchmarks from "programming language and compiler benchmarks" site #2167
Description
Activity
Seems like are likely measuring startup as dynamic PGO is frequently slower.
@joperezr @stephentoub in case you're interested, the above site has kindly added regex-redux at the link above. They used CSharp#5, the fastest one that doesn't use PCRE (which IMO is cheating).
Eyeballing, the Rust- C# - Java ratios for these 6.0 numbers aren't that different to the 5.0 numbers still up at https://benchmarksgame-team.pages.debian.net/benchmarksgame/performance/regexredux.html
6.0 numbers aren't that different to the 5.0 numbers
That's what we'd expect; there was little-to-no work done for 6.0 in anything that would meaningfully improve this benchmark. I expect they will improve in 7.0.
Reacted by Dan Moseley
https://programming-language-benchmarks.vercel.app/csharp-vs-java
Some are versions from the familiar CLBG, but some are new and Java numbers suggest we have room for improvement.
cc @stephentoub @AndyAyersMS in case either are interested.
The above are 6.0 numbers. Note that the official CLBG maintainer has said he cannot update the numbers for a few months, so we don't have 6.0 numbers from there.