Explore the potential of GPU acceleration in Java applications through this 49-minute Devoxx conference talk. Discover how GPUs can be leveraged beyond traditional rendering tasks to optimize high-performance enterprise and technical computing applications, including big data and analytics workloads. Learn about explicit GPU programming and JIT compiler optimization techniques for offloading work to GPUs. Gain insights into GPU programming principles, software stacks, and developer tools available for Java. Watch a demonstration of GPU acceleration and get a glimpse of future developments. Delve into topics such as floating threads, use cases for GPUs, AI applications, GPU characteristics, CUDA code integration, JIT compilation, benchmarking, and challenges in GPU programming for Java. By the end of this talk, acquire the knowledge needed to harness the full potential of GPUs in your own applications.
Overview
Syllabus
Introduction
Disclaimers
Topics
Where is Java used
Bring GPUs to Java
Floating threads
Changing number of threads
Rememba clever
Use cases
Where GPUs are being used
AI vs Poker
What makes a GPU good
What type of GPU
HPC cards
GPU
CPU
Good at
Bad at
Things to know
Java Code
CUDA Code
Random Native Method
Example
Namemangling
Debug
Simple way
Sorting numbers
JIT compiler
JIT print
Benchmarks
Advanced
Big Data
Spark GPU
Machine Running Libraries
Challenges
Bad C code
GPUs in Java
Taught by
Devoxx
