Running a Neural Network on the NPU of the i.MX 8M Plus
How to run TensorFlow Lite inference on the i.MX 8M Plus NPU with a fully open-source stack - Mesa, Gallium3D, the Teflon delegate and Etnaviv - reaching about 22x faster inference than the CPU.
Just published
Latest articles
Deep-dives and how-tos from the workbench.
Running a Neural Network on the NPU of the i.MX 8M Plus
ReadBlueZ Part 9: Understanding DBUS - Introspectable - (8)
ReadBlueZ Part 8: Understanding DBUS - PropertiesChanged - (7)
ReadBlueZ Part 7: Understanding DBUS - Get and GetAll properties - (6)
ReadUpstream activity
What we've just merged - and what's in review right now.
drivers: adc: add driver support for TI MSPM0 ADC module
drivers: spi: add driver support for TI MSPM0 SPI module
arch: add TriCore architecture and Infineon AURIX platform support
arch: add TriCore architecture and Infineon AURIX platform support
The Linumiz ecosystem
The engineering company and the platforms we build, beyond the blog.
Linumiz
Embedded software engineering & consulting for Linux and Zephyr RTOS - board bring-up, BSP, drivers, OTA, functional safety and device security.
Linumiz Automotive
From silicon to safety: a Zephyr RTOS automotive stack - multi-silicon bring-up, ISO 26262, automotive protocols (J1939, UDS) and CI/CD for ECUs.
labwire
Reserve, flash and monitor physical dev boards from anywhere - one command, no VPN. Hardware-in-the-loop testing, queued and fully audited.