Closed rui-mo closed 1 week ago
Name | Link |
---|---|
Latest commit | 77eda917232a991c8a34d2b89c8c4d6b86f7f24e |
Latest deploy log | https://app.netlify.com/sites/meta-velox/deploys/6650838991df410008b730b8 |
@PHILO-HE Would you like to take a review? Thanks.
Do you need to update the document?
@jinchengchenghh Updated the documentation. Thanks.
@mbasmanova Could you help review this change? Thanks!
@mbasmanova has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator.
@rui-mo
I wonder if I can follow-up below TODO in a separate PR to provide custom cast signatures for Presto and Spark.
That would be great. Thanks.
@mbasmanova has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator.
@rui-mo I'm seeing "Conbench performance report — Found 2 regressions". Would you take a look?
CC: @assignUser @kgpai
@mbasmanova merged this pull request in facebookincubator/velox@9446f67b14e9690ff3e5b0f85e46abb788e6560a.
Conbench analyzed the 1 benchmark run on commit 9446f67b
.
There were no benchmark performance regressions. 🎉
The full Conbench report has more details.
@mbasmanova Thanks for your review. It appears that no regression is mentioned in the most recent report. Please kindly contact me if I'm missing something.
@rui-mo Perf testing can be flaky at times. This must be one of these cases. CC: @assignUser
@mbasmanova @rui-mo Yeah looks like at the time this was committed the times where very close together so false positives can happen. But while checking this out I noticed that PR a day later clearly made the results worse: https://velox-conbench.voltrondata.run/compare/benchmark-results/0665116962307f828000737782ce0ccc...06651355ec2377e8800073bd529feb0a/
Fixes https://github.com/facebookincubator/velox/issues/9820.