abacef commented on PR #7819: URL: https://github.com/apache/arrow-rs/pull/7819#issuecomment-3321832059
@alamb would you be able to run a benchmark with the env `RUSTFLAGS='-C target-cpu=native'` for the benchmark `json-writer`? On my machine (2020 x86_64) I do not get any regressions with newer CPU flags, but I do get a 15% regression with the unchanged code paths with the default cpu flags. I am hopeful your GCP x86_64 benchmarking machine will find similar results. Would this be a convincing proof to resolve conversations about the benchmarks? Or do we specifically want to take into consideration most people compiling using the default CPU flags (I believe Rust defaults to x86 CPUs from 2000)? -- 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]
