Closed BridgeAR closed 6 years ago
Yes and no. The current approach was copied over from pino, as benchmark is almost useless because of I/O. Moreover, benchmark impacts heavily the V8 call stack, and it skew the results.
Overall, I'd say 👍 for this repository. However, I'm not really satisfied with both.
Resolved with v.2
Closing
As far as I see it the current benchmarks are not ideal. It would be great to have something similar like the Node.js benchmarks where they are over and over and a significance is shown as well. Something similar is done when using https://www.npmjs.com/package/benchmark.
Should we switch to that?