14–15 Nov 2026
Bengaluru
Asia/Kolkata timezone

Design and Implementation of an Ubuntu-Powered RISC-V64 Edge AI Computing Platform with Heterogeneous Hardware Acceleration

Not scheduled
10m
Audi 1

Audi 1

Lightning Talk (5 mins) Artificial Intelligence & Machine Learning (AI/ML)

Speaker

Ravi Kiran

Description

The exponential rise of edge computing
necessitates robust, power-efficient, and open hardware
platforms capable of supporting AI inference, 3D & Video
processing, and general-purpose computing within
constrained environments. Whereas x86 and ARM
architectures dominate this space, they introduce
licensing, cost, and security barriers. In contrast, the
RISC-V64 architecture promises a truly open and
modular alternative, but its adoption in real-world,
application-ready edge systems remains limited due to
several critical challenges: lack of mature software stacks,
inadequate hardware acceleration support, and minimal
integration with mainstream desktop environments.
This paper introduces a RISC-V based Single
Board Computer (SBC) running a fully operational
Ubuntu 22.04 Linux distribution, architected specifically
for edge intelligence workloads. A key innovation is the
support for the GNOME-Wayland Window Manager,
fully optimized for RISC-V with hardware acceleration
via Mesa and OpenGL-ES & Vulkan-compatible drivers
and GPU offloading. We demonstrate seamless execution
of complex open-source desktop applications such as
Mozilla Firefox, LibreOffice, MPV Media Player (4K
playback with real-time hardware Video decoding), and
Geney & Jupyter Web IDE, all running natively on the
RISC-V64 Ubuntu 22.04 SBC. This paper demonstrates
the feasibility of RISC-V as a high-performance edge
computing architecture. The proposed system effectively
bridges the usability gap in RISC-V-based environments
but also demonstrates a scalable path for mass adoption
in open-source driven, hardware-accelerated edge
intelligence platforms.

Summary

Design and Implementation of an Ubuntu-Powered RISC-V64 Edge AI Computing Platform with Heterogeneous Hardware Acceleration

Speaker Bio

Senior Software Engineer & Team Lead at Exaleap Semiconductor Pvt Ltd.

Speaker LinkedIn / Socials https://www.linkedin.com/in/ravikiran-krishnappa
Difficulty Intermediate

Presentation materials

There are no materials yet.