To make our programs run faster we make them concurrent. But then we forgot to make the concurrency work well.
What's more, it's a bit of surprise to find out that adding more threads to the pool, not only doesn't help, but slows down the application even more.
And so, we usually decide alternatively to stick to the defaults in our configuration parameters - which is typically not bad decision, but not optimal either.
Talk meant for JVM developers to revise threading basic cognition and see its consequences for building concurrent programs. It will first cover the subject at operating strategy lever and then show how, despite of the fact process has no power to influence OS scheduler, we can inactive compose our code to utilise processor cores in optimal way. Pragmatically selected explanation will be followed by real planet examples, comparing different thread pooling approaches (global thread pool vs. dedicates ones), illustrated by charts showing execution performance benchmarks. All this to supply a useful applicable cognition in nutshell that can be applied in your task even on next working day.
Lecture took place on Wednesday, 22nd June at 16:00 in area 3
Krzysztof Płachno:
Developer traveling mostly via JVM lands, armed in Java and Scala, fighting with Web App dragons and pursuing large Data quests. And due to the fact that all reasonably tale dream yet comes to an end, so needed I to leave my cozy programming comfort region and get my hands dirty with operations stuff. And thus, willy-nilly, I became DevOps, which in fact turned out not to be that bad at all. Always willing to realize things in item as well as to see them in wide image view where they fit among others. Strong believer that technology is only a tool and that yet our goal is to solve end users problems.
#IT #Development #SoftwareDevelopment
