Devnexus 2020 Large Scale Distributed AI on the JVM Mo Haghighi
About this talk
This talk explores the capabilities of the Java Virtual Machine (JVM) on ARM and X86 architectures, focusing on large-scale data mining using embedded GPUs such as Jetson Xavier and TX2. The speaker presents a benchmarking demonstration comparing CPUs and GPUs for processing time-series data, utilizing both Java and Python libraries. Mo Haghighi emphasizes the role of deep learning in IoT devices, showcasing a cluster of 125 Raspberry Pis integrated with over 500 GPU cores, and discusses the potential for these devices to perform complex reasoning rather than just processing simple sensory data.
More from this event
See all 57 talks →
Devnexus 2020 Keynote: Decoding the weather: How GRAF & New Computer Modeling can Change Everything
27:49
Devnexus 2020 Keynote: Dueling Banjos Code Liverance Burr Sutter, Ray Tsang
39:49
Devnexus 2020 Keynote Quantum Computing Jessica Pointing
40:49
Devnexus 2020 Afterlife Tales Ana Maria Mihalceanu
41:18