2 ms·
To me it seems pretty misleading to publish a benchmark with a language setting that basically cannot be used in real world applications. Makes the whole benchm
by marcinzm 6y ago
To me it seems pretty misleading to publish a benchmark with a language setting that basically cannot be used in real world applications. Makes the whole benchmark pretty much pointless except as a game.
- msplash9 6y agoThat's specifically the purpose of the benchmarking right ? to be misleading and get attention. I guess this benchmark achieved that. Java subs are celebrating it like they won a war. Java is a great language and will remain one of the top languages in the near future. But these silly benchmarks don't serve any purpose. I am expecting to see this benchmark thrown around a lot from now on any language discussion. Whoever is doing these benchmarks should be more responsible and sensible.
- stingraycharles 6y agoIt’s not necessary a “language setting” — a Java “ExecutorService” is an abstraction which allows one to dispatch futures / async jobs. In this case, the DirectExecutor is an implementation of this service that, doesn’t do anything async at all, avoiding context switches and all other overhead that comes with it. If anything, to me it’s a clear example of why benchmarks like these are silly as they benchmark very uninteresting aspects that typically are never real world bottlenecks.