JetPack 4.6 Production Release with L4T 32.6.1

We are pleased to announce JetPack 4.6, a production release supporting Jetson AGX Xavier series, Jetson Xavier NX, Jetson TX2 series, Jetson TX1, and Jetson Nano.

JetPack 4.6 includes support for Triton Inference Server, new versions of CUDA, cuDNN and TensorRT, VPI 1.1 with support for new computer vision algorithms and python bindings, L4T 32.6.1 with Over-The-Air update features, security features, and a new flashing tool to flash internal or external media connected to Jetson.

In addition to l4t-base container image, new CUDA runtime and TensorRT runtime container images are released on NVIDIA NGC, which include CUDA and TensorRT runtime components inside the container itself, as opposed to mounting those components from the host. These containers are built to containerize AI applications for deployment. Note that the L4T-base container continues to support existing containerized applications that expect it to mount CUDA and TensorRT components from the host.

Highlights of JetPack 4.6 are:

  • Support for new 20W mode on Jetson Xavier NX enabling better video encode and video decode performance and higher memory bandwidth. The included 10W and 15W nvpmodel configurations will perform exactly as did the 10W and 20W modes with previous JetPack releases. Any custom nvpmodel created with a previous release will require regeneration for use with JetPack 4.6. Please read release notes for details.
  • Image based Over-The-Air update tools for developing end-to-end OTA solution for Jetson products in the field. Supported on Jetson TX2 series, Jetson Xavier NX and Jetson AGX Xavier series. Download the OTA tools from the L4T page under the Tools section.
  • A/B Root File System redundancy to flash, maintain and update redundant root file systems. Enhances fault tolerance during OTA by falling back to the working root file system slot in case of a failure. Supported on Jetson TX2 series, Jetson Xavier NX and Jetson AGX Xavier series.
  • A new flashing tool to flash internal or external media connected to Jetson1. Supports Jetson TX2 series, Jetson Xavier NX and Jetson AGX Xavier. The new tool uses initial RAM disk for flashing and is up to1.5x faster when flashing compared to the previous method2.
  • Secure boot is enhanced3 for Jetson TX2 series to extend encryption support to kernel, kernel-dtb and initrd.
  • Disk encryption of external media supported to protect data at rest for Jetson AGX Xavier series, Jetson Xavier NX and Jetson TX2.
  • NVMe driver added to CBoot for Jetson Xavier NX and Jetson AGX Xavier series. Enables loading kernel, kernel-dtb and initrd from the root file system on NVMe.
  • Enhanced Jetson-IO tools to configure the camera header interface and dynamically add support for a camera using device tree overlays
  • Support for Scalable Video Coding (SVC) H.264 encoding
  • Support for YUV444 8, 10 bit encoding and decoding
  • Direct downloadable links to JetPack and L4T debian packages for Jetson

Visit JetPack 4.6 and L4T 32.6.1 page for more details.

JetPack 4.6 components:

  • L4T R32.6.1
  • CUDA 10.2
  • cuDNN 8.2.1
  • TensorRT 8.0.1
  • VisionWorks 1.6
  • OpenCV 4.1.1
  • Vulkan 1.2
  • VPI 1.1
  • Nsight Systems 2021.2
  • Nsight Graphics 2021.2
  • Nsight Compute 2019.3

Install JetPack 4.6 using SDK Manager or SD card image (for Jetson Nano 2GB Developer Kit, Jetson Nano Developer Kit and Jetson Xavier NX Developer Kit) or upgrade via Debian package management tool (refer to the instructions here)

Refer to Jetson Roadmap page for roadmap on Jetson hardware and software.

Please register to following webinars:

[1] Flashing from NFS is deprecated and replaced by the new flashing tool which uses initrd

Updated Jetson Xavier NX Thermal Design Guide and Jetson Xavier NX Datasheet are now available on Jetson Download Center Jetson Download Center | NVIDIA Developer

Triton Inference Server release supporting JetPack 4.6 is live now!

Visit server/jetson.md at main · triton-inference-server/server · GitHub

JetPack 4.6 containers are live now on NGC!
As mentioned in the announcement, we have release 2 new containers this time, a cuda runtime container and a tensorrt runtime container. These containers include CUDA and TensorRT runtime components inside the container itself, as opposed to mounting those components from the host.

L4T base container: NVIDIA NGC
CUDA Runtime Container: NVIDIA NGC
Tensorrt Runtime Container: NVIDIA NGC

The md5sum of NX sd image doesn’t match what is posted “6326244400c2f01aeece13a09a63d88d”. I got the following one:

c22ebfeca65055bda8327be074e34c58 jetson-nx-jp46-sd-card-image.zip

And in fact the Nano image doesn’t match either. Can Nvidia people please help verify? Thanks.

Which version of DeepStream can I use with JetPack 4.6 ?

Can I install DeepStream 5.1 on JetPack 4.6, by installing from NVIDIA apt server ?

Thanks.

Can you give more precise information on the 32.6.1 Multimedia API update? (which is not called Multimedia API anymore)

I can see libArgus now has 2 types of output streams. What is the advantage of using one over the other? Which one is comparable to the one from previous versions?

Can I download the multimedia api somewhere without installing a 4.6 Jetpack Jetson?

Thank you!
jb

@Colin this is fixed now. Can you recheck?

@ryanwong net DeepStream release DeepStream 6.0 will support JetPack 4.6.

It doesn’t seem to fix the issue. I still get:

$ md5sum jetson-nx-jp46-sd-card-image.zip
c22ebfeca65055bda8327be074e34c58 jetson-nx-jp46-sd-card-image.zip

The link I download image is https://developer.nvidia.com/embedded/l4t/r32_release_v6.1/jetson_xavier_nx/jetson-nx-jp46-sd-card-image.zip

Downloaded the jetson-nx-jp46-sd-card-image.zip

md5sum is incorrect

 md5sum jetson-nx-jp46-sd-card-image.zip 
1db4f5f1beb97049ad7b29cf6340a7eb  jetson-nx-jp46-sd-card-image.zip

Rather than stated

Jetson NX Developer Kit SD Card Image  4.62021/08/04
This SD card image works for the Jetson Xavier NX Developer Kit and is built with JetPack 4.6. Download and write it to your microSD card and use it to boot the developer kit.

md5sum: 6326244400c2f01aeece13a09a63d88d

blob

user@user-desktop:~/Downloads$ md5sum sd-blob.img 
325a8261caf0827be77f1dee77739c8a  sd-blob.img

Second attempt the same is the image compromised or do you even care at NVIDIA?

@desktop:~/Downloads$ md5sum jetson-nx-jp46-sd-card-image1.zip
1db4f5f1beb97049ad7b29cf6340a7eb  jetson-nx-jp46-sd-card-image1.zip

@Nick_579 we are looking into it. We will fix it soon and update this thread.

1 Like

@Nick_579 justy to explain the issue, the md5 checksum you have caluclated is what it supposed to be
1db4f5f1beb97049ad7b29cf6340a7eb

But the site is showing a wrong md5, it needs a refresh which we are working on

2 Likes

@Nick_579 its fixed now and the page shows the righ checksum.

1 Like

Thank You for clarifying. Just started to test with this version and have had the system crash here and there. Otherwise it feels a bit more snappy at least until the slowdown to freeze. It truly does not like to run at 6 cores 20W.

Hi nikoff90,

Would you please help to create a new topic for how to test and how to observe it can’t run at 6 core 20W?
We will support you there to fix the problem if any.

Thanks

@busch.johannes sorry for the late reply.

There are 2 types of OutputStreams, BufferStream and EGLStream

BufferStream

Application-managed Buffer Streams

In addition to outputting capture results to an EGLStream, clients now have the ability to output capture results directly to client-allocated image buffers.

Buffer - Created and owned by an OutputStream, these objects wrap application-managed buffer resources and are used to synchronize data access to the buffer resources between Argus and the client application. Buffers are released to their stream to be used by Argus for a capture request, and are acquired back when the capture is complete and the client is ready to consume the output.

EGLStream-linked OutputStream objects maintain a connection to an EGLStream as the producer endpoint. The EGLStream implementation is responsible for buffer allocation, management, and synchronization as frames are presented to the EGLStream from libargus and then consumed using an EGLStream consumer.

EGLStream is the earlier one already supported.

Please look into the documentation for more details.

https://docs.nvidia.com/jetson/l4t-multimedia/group__ArgusOutputStream.html

Hi there,

After the update, it seems like Jetpack is no longer recognizing my wifi card. Wlan0 has disappeared. Can i get some guidance on this issue? Thanks

Allen