Can someone explain to me the Php-raw outperforming Go on db queries using the PEAK control? They mention something about the NUMA architecture potentially being at fault?
I’m not going to question the facts (numbers tell there are differences), but why as a programmer should I care? This equation isn’t that simple.
It’s important to know the upperbounds of performance for your chosen frameworks. Just think of these results as another data point when deciding on which framework to pick.
Is there anywhere where you can see which version of each language/compiler was used? GHC 7.8.2 should produce much better results for concurrent connections for example, but I can’t see if they’re using it or not.
They seem to be using GHC 7.4.1. The test suite is on GitHub, and for Haskell/Yesod there’s a readme file detailing the environment.