Edge AI Without GPU Acceleration: Premio x MemryX Multi-Stream Inference Demo
The Rugged Edge Media Hub · Thomas Riley
Audio is streamed directly from the publisher (podcast.wistia.com) as published in their RSS feed. Play Podcasts does not host this file. Rights-holders can request removal through the copyright & takedown page.
Show Notes
Explore how Premio and MemryX are advancing edge AI performance through a high-efficiency, GPU-free architecture.
In this Embedded World interview, Jordy Hsu (Premio) and Sam Fuller (MemryX) present a real-world edge AI deployment achieving approximately 1000 FPS multi-stream inference without traditional GPU acceleration. The solution combines Premio’s fanless BCO-3000 industrial computer with the MemryX MX3 AI accelerator to deliver scalable, low-power AI processing at the edge.
This approach highlights a shift toward specialized AI acceleration, enabling faster, more energy-efficient inference for applications such as machine vision, industrial automation, and real-time analytics.