4 ms·
I'm thinking that ultimately, you're testing two different things here. Language speed is part of it, but also the performance of your web server library. I'm
by moomin 15y ago
I'm thinking that ultimately, you're testing two different things here. Language speed is part of it, but also the performance of your web server library.
I'm not saying it's unfair or meaningless, but it's worth bearing in mind that, for instance, the node.js web server is mostly written in native code, while the clojure web server is written in Clojure. For a large program, I imagine that node would get slower and slower, and Clojure and Scala would get faster and faster.
- gtrak 15y agoIt's hard to know what you're actually testing without profiling. My guess is it's only the overhead of netty and lamina at play here, and not clojure itself. There are profiling facilities to know for sure, and the author should try to check that out. I can get like 8k request/s from clojure with jetty last time I tried.