Skip to content

Jetson Discussions

This page contains paraphrased discussions that may be useful for future reference. Also note that these discussions are searchable from the search bar in the website.

Please note that all Jetson modules below refers to the Developer Kit instead of custom carrier boards.

For issues related to this page, please open a GitHub issue.

Jetson Orin Nano vs. Jetson Nano

Q: Is Jetson Orin Nano the same as Jetson Nano?

A: Jetson Orin Nano is not the same as Jetson Nano. Since Jetson Orin Nano is a newer model, the Jetpack version and associated packages have different versions compared to Jetson Nano.

Reference: Jetson Modules

2024-10-02. Jetson Orin Nano.

NVMe SSD vs. microSD Card

Q: Is it necessary to install an NVMe SSD on Jetson Orin Nano? Or is a microSD card sufficient?

A: While the optional NVMe SSD is not essential for initial experimentation, it is highly recommended for serious development. The built-in storage on Jetson Orin Nano is quite limited, and an NVMe SSD provides fast, large storage, which is particularly useful when working with containers, such as for Isaac ROS. If budget is a constraint, you can start with the internal eMMC storage and purchase an NVMe SSD later when needed. It is also worth pointing out that the external NVMe SSD is faster than the internal eMMC storage, microSD card, and USB. You can also install the operating system on the NVMe SSD.

Related:

2024-10-02. Jetson Orin Nano.

Q: What is the recommended NVMe SSD storage for Jetson Orin?

A: Personally, I recommend purchasing a 1TB (or 500GB if budget is a constraint) NVMe SSD. If you are located in Taiwan, you may want to purchase them from local retailers such as PCHome or MOMO for shipping within 24 hours. In addition, if you're purchasing for a university lab or company, you may need to enter your Business ID number (統一編號) during purchase. See the following videos for recommended SSDs suggested by JetsonHacks:

References:

2024-10-12. Jetson Orin Nano.

LLM/VLM on Jetson AGX Orin

Q: Suggested references for running LLM/VLM on Jetson AGX Orin?

A: I recommend using the following tutorials:

References:

2024-10-20. Jetson AGX Orin.

Power Supply for Jetson Orin Nano

Q: What is the recommended power supply for using Jetson Orin Nano on mobile ground robots?

A: If your mobile ground robot is large enough, you can simply use a large power bank. Personally I've tried a large enerpad AC40K with 40K mAh, and it seems to work well.

2024-10-23. Jetson Orin Nano.

(Open Issue) Fail to Control GPIO in Jetpack 6.x

Q: See this issue.

A: This is a known issue in Jetpack 6.0, 6.1, and 6.2. Please see the following for workarounds:

References:

2024-11-08. Jetson Orin Nano.

  • Jetpack 6.0 L4T 36.3.0
  • Jetpack 6.1 L4T 36.4.0
  • Jetpack 6.2 L4T 36.4.3

Memory Size of Jetson Orin Nano Super

Q: How much memory does the Jetson Orin Nano Super have?

A: The Jetson Orin Nano Super has 8GB of memory (i.e., software upgrade of Jetson Orin Nano 8GB). See the official response:

The Jetson Orin Nano Super has no physical hardware and package differences compared to the Jetson Orin Nano Developer Kit. The performance upgrade comes from software optimization. Existing Jetson Orin Nano Developer Kit users can experience this performance boost with just a software update.

-- Jetson FAQ

Reference:

2025-04-01. Jetson Orin Nano (Super).

Jetson Docker Base Images

Q: What's the difference between l4t-base, l4t-jetpack, and ubuntu.

A: See the following:

  • nvcr.io/nvidia/l4t-base is mainly used in JetPack 4.x, and I believe is no longer maintained anymore since the latest tag r36.2.0 (correspond to JetPack 6.0 Developer Preview) is from 2023.
  • nvcr.io/nvidia/l4t-jetpack is mainly used in JetPack 5.x, and I believe is still being maintained. The latest tag r36.4.0 corresponds to JetPack 6.1. (However, I'm not sure if r36.4.0 is fully compatible with JetPack 6.2, which corresponds to L4T R36.4.3.)
  • nvcr.io/nvidia/ubuntu is the official Ubuntu base image. The tag 22.04 can also be used for Jetson after JetPack 6.x.

I'll suggest using dusty-nv/jetson-containers to build custom images.

References:

2025-05-14.

JetPack and L4T Version Conversion

Q: How to convert JetPack version to L4T version?

A: See JetPack Archive. For example, JetPack 6.2 corresponds to L4T R36.4.3. You can also find the CUDA version by clicking the corresponding JetPack version.

If you forgot the JetPack version used to flash your Jetson device, find the L4T version by cat /etc/nv_tegra_release and then reverse look up the JetPack version.

Reference:

2025-05-14.

Environment Setup After Reflashing Jetson

Q: What is the recommended environment setup after reflashing Jetson?

A: Personally, I'll use Ansible to install all basic dependencies (AutoLogin, VNC, etc.) with a single command. See j3soon/ansible-playbooks for further details.

2025-05-14.

GPU Access Issue Before Logging In

Q: I'm facing an issue where the GPU is not accessible before logging in. How can I fix this?

A: Consider setting up automatic login by editing the /etc/gdm3/custom.conf file with the following:

AutomaticLoginEnable = true
AutomaticLogin = <YOUR_USERNAME>

Related:

2025-05-14.