I was looking at these benchmarks too the other day. And I also noticed the high error rate on Phoenix’s benchmarks. It seems strange to me, like their server wasn’t setup right. It seems implausible that Phoenix could perform so poorly. The GitHub repo is here - I thought about pulling it down and running the benchmarks myself. For one thing, the version of Phoenix is out of date, so we could make a pull request to update those dependencies.
2 Likes
