NVIDIA Jetson Setup

If you need to set up Basestack, you must install the Nightly build variant of the distribution at: https://github.com/jhuapl-bio/Basestack/releases/latest

Note

Be sure to set up the Nano or Xavier (NX) properly before doing this. See more information here: https://developer.nvidia.com/embedded/downloads

Note

Depending on your distribution, you may already have docker installed it seems as NVIDIA is more fully supporting the Docker build toolkit.

To check, run docker –version. If so, skip that section of the supplemental software install process (see below) You will definitely need to be sure to run through the post-installation steps, though. See here: Docker Installation

Simply follow these 2 steps:

Download the arm64 AppImage from the above-mentioned releases page
OPTIONAL: Run the install script. This can be found here

Note

The install script described above will ask several questions for setting up your environment based on your needs. Make sure to select the r option for any question asking for arm64 or amd64

Also, try to follow steps in this link for minknow Or, try this link if you have troubles with CUDA and Guppy

Warning

The later versions of MinKNOW can lead to a failure to load the barcoding kits in the UI. If you experience this, follow this procedure Barcoding Kits Missing to remedy

Note

JetPack (Nviida Jetson only) does not ship with nvidia-smi as a way to monitor gpu usages. You can instead run sudo -H pip install -U jetson-stats to get the jtop command to monitor resources on your jetson device