TornadoVM: Transparent Hardware Acceleration for Java…and Beyond!
summary
TornadoVM is an OpenJDK plug‑in that compiles JVM methods to run on multi‑core CPUs, GPUs and FPGAs using OpenCL, NVIDIA PTX and SPIR‑V (Level Zero) backends, performs automatic task migration and provides per‑vendor IR optimization. Demoed with a Mandelbrot kernel it drops a 300 ms Java run to ~59 ms on an integrated GPU (about a 50x speedup reported across kernels; DFT showed ~300x in some tests). Runtime exposes profiling (kernel vs transfer times), is open source (GPLv2 + CPE) and ships examples on GitHub and Docker.
AI-written summary · about
chapters (12)
- 0:00Introduction to TornadoVM
- 3:03Motivation for heterogeneous hardware
- 8:27Overview of TornadoVM project
- 13:47TornadoVM software stack
- 25:11TornadoVM compilation process
- 32:12Mandelbrot demo walkthrough
- 36:01Device inspection and execution
- 41:17Performance evaluation results
- 44:52Profiling and bottlenecks
- 46:38Support for other programming languages
- 47:21Industry applications and use cases
- 49:28Conclusion and future directions
related talks
46mTornadoVM: Supercharge your Java with GPU acceleration
41mTornadoVM
43mFrom CPU to GPU and FPGAs: Supercharging Java Applications with TornadoVM
44mWrite Once, Accelerate Everywhere: GPU-Ready Java with TornadoVM
17mDynamic and Static Code Analysis for Java Programs on Heterogeneous Hardware
43mHow to make your electrons do more in Java?
49mJVM in the Age of AI: Babylon, Valhalla, TornadoVM and friends
33mOptimizing Java on the EU processor platform