Welcome to Oracle Communities
Register and Participate in Oracle's online communities. Learn from thousands of experts, get answers to your questions and share knowledge with peers.
Choose a layout and drag widgets onto your Overview Page to customize it. Widgets placed on the page below can be configured by selecting the symbol.
Performance tuning Twitter services with Graal and Machine Learning
Running Twitter services on Graal has been very successful and saved Twitter a lot of money in data center costs. But in order to run more efficiently and reduce costs, even more, Twitter is using a machine learning framework called Autotune to tune Graal inlining parameters. This session shows how much performance improvement was gained by autotuning Graal.
Chris Thalinger is a software engineer who has been working on Java Virtual Machines for more than 14 years. His primary expertise is in compiler technology, with a focus on just-in-time compilation. Initially involved with the CACAO and GNU classpath projects, his focus shifted to OpenJDK as soon as Sun made the JDK open source. Ever since, Chris has worked on the HotSpot JVM at Sun, Oracle, and now at Twitter.
* US$300 free credits available in select countries and valid for up to 30 days. You can run a maximum of 8 instances across all available services and up to 5 TB of storage during the promotional period. Oracle Cloud credits are consumed at discounted rates during the 30-day promotional period. The capacity limits listed under each service are only estimates and reflect the maximum capacity you can get if you consume your entire credits on one service during the promotional period. The maximum capacity per service you get can be less if you consume multiple services. The list of services and capacity estimates are subject to change. Your credit balance is available to you on the Oracle My Services dashboard.