slorado

Building for ARM64-based Jetson devices

We do not check the building for ARM64-based Jetson using CI/CD, so please note that the following instruction may not work as it is.

Install zlib development files needed for slow5lib:

 On Debian/Ubuntu : sudo apt-get install zlib1g-dev

As of 18/11/2024, no binary libtorch download is available for ARM64-based Jetson. So we will install pytorch that includes torchlib. First make a Python virtual environment:
```
 sudo apt install python3.8-venv
 python3 -m venv pytorch_venv
 source pytorch_venv/bin/activate
```
Update pip and install pytorch for your specific Nvidia Jetpack version. You can find this by running sudo apt-cache show nvidia-jetpack | grep "Version", or browse https://developer.download.nvidia.com/compute/redist/jp/ to find a suitable version of pytorch. We tested on a Jetson Xavier board with Jetpack 5.0 installed and the commands used were:
```
 pip3 install --upgrade pip
 pip3 install --no-cache  https://developer.download.nvidia.com/compute/redist/jp/v51/pytorch/torch-2.0.0a0+8aa34602.nv23.03-cp38-cp38-linux_aarch64.whl
```

Clone and build.

 git clone --recursive https://github.com/BonsonW/slorado.git
 cd slorado
 make -j cuda=1 cxx11_abi=1 LIBTORCH_DIR=/path/to/pytorch_venv/lib64/python3.8/site-packages/torch/

Advanced building options

You can optionally enable zstd support for built-in slow5lib when building slorado by invoking make zstd=1. This requires zstd 1.3 development libraries installed on your system (libzstd1-dev package for apt, libzstd-devel for yum/dnf).

This site is open source. Improve this page.