You are viewing the 2025 edition of FOSDEM. Click here to view the 2026 edition
Using AI hardware accelerators for real-time DSP on embedded devices - NPU, TPU etc,
UB2.147 | Day 1 | 10:45 - 11:40 | Speakers: Sylvain AZARIAN
Using AI hardware accelerators for real-time DSP on embedded devices - NPU, TPU etc,
Abstract
While we presented last year what could be achieved using NVIDIA GPU, the hardware market has seen a lot of new comers following the need to have AI running locally on devices. Google has the Coral, Rockchip is adding a AI coprocessor on their ARM CPU, and now we have the Copilot PC coming with additional power focused on running Deep Learning models.
In this talk I will present how these accelerators can also be useful for SDR realtime processing, showing some results and presenting the first beta of a new open source library.
Attachments
Speakers
Sylvain AZARIAN
Links
External Links
Notice: The placeholder video image is licensed under CC BY-SA 4.0. The original image can be found hereChanges made to the image are: Cropped the image to a new ratio, part of the image was cut off.
