JetPack 6 Key Features
| JetPack 6 Migration and Developer Guide |
|---|
|
| Introduction |
| Migrating From JetPack 5 to JetPack 6 |
| Installing JetPack 6 |
| JetPack 6 Key Features |
| Contact Us |
At its core, JetPack 6 utilizes a Linux kernel, finely tuned for high performance and efficiency to meet the demanding requirements of AI and deep learning workloads. Notably, it offers the flexibility to "bring your own kernel", allowing developers to customize the kernel version according to their specific needs. The Compute Stack in JetPack 6 includes the NVIDIA CUDA toolkit, cuDNN, TensorRT, and other libraries, and is now upgradable, ensuring that developers can easily stay up-to-date with the latest enhancements and features. This stack is integrated with an Ubuntu-based Linux distribution, ensuring a stable and familiar environment for development, enhanced with NVIDIA's customizations to support the unique capabilities of Jetson hardware. Additionally, JetPack 6 introduces support for new Linux distributions, providing more options and flexibility for developers.
- Bring Your Own Kernel section presents related information and high level steps about the Bring Your Own Kernel feature.
- Upgradable Compute Stack section presents the new feature: Upgradable Compute Stack, and shows high level steps on how to perform this upgrade.
- New Linux Distributions section introduces the new feature of JetPack 6 that enables new Linux distributions on JetPack and examples of some of these.
- JetPack Compute Stack section presents the new functionalities integrated in the latest Compute Stack within JetPack.