Working with Additional NVIDIA DriveOS Packages#
While the base NVIDIA DriveOS SDK provides the essential components for platform development, many developers require specialized SDKs and tools to build, test, and optimize their applications. These additional packages extend the functionality of your development environment to support specific use cases such as GPU computing, deep learning inference, autonomous vehicle software development, and system profiling.
These packages are particularly valuable if you need to accomplish the following tasks:
leverage NVIDIA’s hardware accelerators (GPU, DLA, PVA)
implement AI and machine learning workloads
develop with the NVIDIA DriveWorks SDK framework
profile and optimize application performance
By installing only the packages relevant to your project, you can maintain a streamlined development environment while ensuring you have all the necessary tools and libraries for your specific workflow.
Reference: Additional NVIDIA DriveOS Packages#
The following additional NVIDIA DriveOS packages are available for download:
Additional Packages |
|---|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Selecting Additional Packages by Development Goal#
Use the following sections to identify which packages you require based on your development goals.
CUDA Development#
To develop or test CUDA code or CUDA samples:
Required packages:
driveos-cuda-repo-ubuntu2404-*-local_*_amd64.deb(host)driveos-cuda-thor-nsr-repo-cross-aarch64-ubuntu2404-*-local_*_all.deb(cross)driveos-cuda-thor-nsr-tegra-repo-ubuntu2404-*-local_*arm64.deb(target)
Next steps: See the CUDA chapter.
TensorRT for AI Inference#
To work with TensorRT for AI inference:
Required packages:
nv-tensorrt-repo-ubuntu2404-cuda*-trt*-d7l-cross-ga-*_amd64.deb(cross)nv-tensorrt-repo-ubuntu2404-cuda*-trt*-d7l-target-ga-*_arm64.deb(target)cudnn-local-tegra-repo-ubuntu2404-*_arm64.deb(target)
Next steps: See the TensorRT chapter.
DriveWorks SDK Samples#
To develop with DriveWorks SDK samples:
Required packages:
driveworks-samples-linux-nsr-aarch64-*.tar.gzdriveworks-stm-samples-linux-nsr-aarch64-*.tar.gzdriveworks-data-all-*.tar.gz
Next steps: See the DriveWorks chapter.
LLM (Large Language Model) SDK#
To work with the LLM SDK:
Required packages:
driveos_llm_sdk-*.tar.gz
Next steps: See the DriveOS LLM SDK chapter.
PVA (Programmable Vision Accelerator)#
To use the Programmable Vision Accelerator:
Required packages:
pva-algos-lib-*-aarch64-linux-*.debpva-algos-lib-docs-html-*-algos.zip
Next steps: See the PVA chapter.
Application Profiling and Optimization#
To profile and optimize applications:
Required packages:
NsightSystems-linux-drive-*-nda-*.deb(host)NsightSystems-cli-linux-drive-*-nda-arm64-*.deb(target)
Next steps: See the Nsight Systems chapter.
Downloading Additional NVIDIA DriveOS Packages#
For detailed instructions on how to download and install these additional packages, refer to the NVIDIA DriveOS Installation Guide:
NVIDIA NVONLINE users (partners.nvidia.com): See Optional: Downloading and Installing Additional NVIDIA DriveOS Packages from Artifactory in the Installing NVIDIA DriveOS for NVIDIA NVONLINE Users section.
NVIDIA Developer users (members of the NVIDIA DRIVE® AGX SDK Developer Program): See Optional: Downloading and Installing Additional NVIDIA DriveOS Packages from NGC in the Installing NVIDIA DriveOS for NVIDIA Developer Users section.