Multi-Language Runtime
summary
IBM refactors the J9 VM into a language‑agnostic toolkit and threads components—allocator, Health Center tracing, J9 garbage collector and the Testarossa JIT—into Ruby MRI and CPython. Method profiling, verbose GC output and a visualizer appear with minimal changes; Rails runs end‑to‑end. Microbenchmarks show several workloads >2× on x86 and all >1×; Power and mainframe targets work too. CPython GC and C‑extension compatibility remain the main integration challenges.
AI-written summary · about
chapters (16)
- 0:00Introduction
- 0:50Story behind this talk
- 3:30IBMs perspective
- 6:50Existing runtimes
- 7:43JVMs
- 8:52IBM Concerns
- 11:36Secret Path
- 13:31Secret Experiments
- 16:10Method Profiling
- 18:55Garbage Collection
- 22:49Garbage Collection Memory Visualization
- 23:27JustinTime Compilers
- 26:39JustinTime Background
- 28:07JustinTime Results
- 30:19JDK8 Release
- 31:28IBMs Vision
related talks
32mExperiments in sharing Java VM technology with CRuby
48mThe Next Generation of Java Runtimes
48mOne VM To Rule Them All
47mWhy JRuby Works
47mBeyond JVM: How the Platform is Evolving for new Languages & Features
47mJRuby: What, Why, How... Try It Now
31mExpand your Ruby Potential with JRuby
31mSimplifying and Optimizing Life on the Edge with Graal