NVIDIA Jetson Linux

Release Notes

Version 36.4 GA

RN_10698-r36.4 | September 2024

Table of Contents

1. About this Release 2

1.1. Login Credentials 3
1.2. What’s New 3
2. Known Issues 5
2.1. General System Usability 5
2.2. Flashing 7
2.3. Camera 8
2.4. Multimedia 9
2.5. Display 9
2.6. Compute Stack 10
2.7. DeepStream 10
3. Fixed Issues 11
4. Implementation Details 14
4.1. Camera 14
4.2. Device Registration 16
4.2.1. Device Tree Overlay 16
4.2.2. Using the Jetson IO Tool 18
4.3. UEFI 18

NVIDIA Jetson Linux Release Notes RN_10698-r36.4 |1

1. About this Release
The NVIDIA® Jetson™ Linux 36.4 General Availability (GA) release includes the Linux Kernel
5.15, the Ubuntu 22.04-based root file system, the UEFI-based Bootloader, NVIDIA drivers,
the necessary firmware, toolchain, and more. This release supports all Jetson Orin
production modules and Developer Kits.

Important: This GA release can be used for production purposes.

Platform and Release Information

Description Supported Version

Host machine Linux distribution for Ubuntu x64 20.04 or 22.04 (x64 distribution)
flashing software onto Jetson devices

Sample rootfs derived from the Ubuntu Ubuntu 22.04

operating system to run on Jetson
devices.

The supported Linux kernel version. 5.15 LTS

The supported ARM architecture. aarch64

NVIDIA Jetson Linux Release Notes RN_10698-r36.4 | 2

 Description Supported Version

Name of the configuration file used in

flashing. jetson-agx-orin-devkit-industrial.conf: Flashes the Jetson
Orin industrial module (P3701-0008), which is attached to a
Jetson Orin reference carrier board (P3737-0000).
Note: When you flash a configuration
file with the flash.sh script, specify the jetson-orin-nano-devkit.conf: Flashes one of the following
configuration's basename, for example, modules that is attached to a Jetson Orin Nano Carrier
the file name without the .conf suffix. board (P3768-0000):
Some of the products require flashing ● Jetson Orin Nano developer kit module with SD Card
through initrd instead of flash.sh (P3767-0005)
● Jetson Orin Nano 8GB module (P3767-0003)
For a complete description of ● Jetson Orin Nano 4GB module (P3767-0004)
supported platforms and configuration ● Jetson Orin NX 16GB module (P3767-0000)
● Jetson Orin NX 8GB module (P3767-0001)
names, see the Jetson Modules and
Configurations table in Environment
jetson-agx-orin-devkit.conf: Flashes one of the following
Variables.
modules that is attached to a Jetson AGX Orin Developer
Kit (P3730-0000) reference carrier board (P3737-0000):

● Jetson AGX Orin developer kit module (P3701-0000)

● Jetson AGX Orin 32GB module (P3701-0004)
● Jetson AGX Orin 64GB module (P3701-0005)

Board names, module names, and Refer to the Jetson FAQ for a detailed list of Jetson device
revision numbers information.

Toolchain for cross-compilation Bootlin GCC 11.3

https://developer.nvidia.com/embedded/jetson-linux

Release Tag jetson_36.4

1.1. Login Credentials

To create your login credentials, follow the system prompts at the first boot.

1.2. What’s New

Jetson Linux 36.4 is the latest production quality Jetson Linux release that supports NVIDIA
JetPack™ 6.1. Here are the highlights for the release:

● Upgraded compute stack with CUDA 12.6, TensorRT 10.3, cuDNN 9.3, VPI 3.2, DLA 3.1
and DLFW 24.0.

NVIDIA Jetson Linux Release Notes RN_10698-r36.4 | 3

 ● Camera
○ Enhanced the Argus library, reducing CPU utilization by up to 40%.
● Security
○ Added support for Firmware-based Trusted Platform Module (fTPM) on the
Orin platform. For a full list of security features, refer to Security in the Jetson
Linux Developer Guide.
● Multimedia
○ Added GStreamer nvunixfdsink and nvunixfdsrc IPC plugins for cross-process
use cases.
● Networking
○ Implemented the standard predictable naming scheme for Ethernet
interfaces.
● Support for the following Process Change Notifications (PCN):
○ PCN 211361
● Fixes for known security vulnerabilities.

Here is some additional information:

● Jetson Linux Sources are available on Git in addition to the Jetson Linux page (refer to
Working with Sources for more information).
For more information about the adaptation and bring up process for your custom carrier
boards, refer to Jetson Module Adaptation and Bringup for the Jetson AGX Orin, Orin
NX, and Orin Nano platforms.

● Refer to the Jetson Linux Developer Guide for Jetson Linux documentation and
Implementation Details for more information about implementation details that cover a
variety of topics.

NVIDIA Jetson Linux Release Notes RN_10698-r36.4 | 4

2. Known Issues
This section provides details about issues that were discovered during development and QA
but have not yet been resolved in this release.

2.1. General System Usability

The following general system usability-related issues are noted in this release.

Issue Description

4480028 With some USB cables that are shipped with Jetson AGX Orin, you might see the
following errors in the logs:
usb usb2-port2: Cannot enable. Maybe the USB cable is bad?
usb usb2-port2: config error

To stop the error messages, swap the USB cables.

The GPIO sysfs node has been deprecated, and /sys/class/gpio cannot be used
4845581 to control the GPIO state.

To use GPIO from the userspace, use the new GPIO character device API (for
example, libgpiod).

4201479 Although we support flashing Jetson-Linux BSP in multiple boot media like USB and
NVMe, the media should have the same version of BSP. An attempt to flash different
BSP versions to multiple boot media will lead to a system crash in the UEFI because
different overlays are flashed in the UEFI partition.

4868022 Suspend/resume sometimes fails on Jetson Orin NX and Jetson Orin Nano. Also,
these devices could fall back into a suspended state.
This is because the device tree has configured a "Suspend" button and mapped it to
GPIO G.02. This particular pin is not brought out to the module header on Jetson
Orin NX and Nano.

Solution: Delete the "suspend" gpio-key from the Jetson Orin NX/Nano device tree.
To delete this node from the prebuilt binaries:

fdtput tegra234-p3768-0000+p3767-0000-nv.dtb -r
/gpio-keys/key-suspend
fdtput tegra234-p3768-0000+p3767-0001-nv.dtb -r
/gpio-keys/key-suspend

NVIDIA Jetson Linux Release Notes RN_10698-r36.4 | 5

 Issue Description

fdtput tegra234-p3768-0000+p3767-0003-nv.dtb -r
/gpio-keys/key-suspend
fdtput tegra234-p3768-0000+p3767-0004-nv.dtb -r
/gpio-keys/key-suspend
fdtput tegra234-p3768-0000+p3767-0005-nv.dtb -r
/gpio-keys/key-suspend

If building a custom device tree from source, you should make a source change such
as:

diff --git
a/nv-platform/tegra234-p3768-0000+p3767-xxxx-nv-common.dtsi
b/nv-platform/tegra234-p3768-0000+p3767-xxxx-nv-common.dtsi
index f2c048a..3f68f82 100644
--- a/nv-platform/tegra234-p3768-0000+p3767-xxxx-nv-common.dtsi
+++ b/nv-platform/tegra234-p3768-0000+p3767-xxxx-nv-common.dtsi
@@ -429,3 +429,5 @@
status = "okay";
};
};
+
+/delete-node/ &{/gpio-keys/key-suspend};

4450559 On the Jetson AGX Orin series of products, enabling the Wake-On-Lan (WOL)
feature is not recommended when users connect to these products using SSH or
when the target is booting from NFS.

In both cases, users will lose connection with the target.

NVIDIA Jetson Linux Release Notes RN_10698-r36.4 | 6

2.2. Flashing
The following flashing-related issues are noted in this release.

Issue Description

4357750 On the Ubuntu 18.04 Linux host, the sudo ./apply_binaries.sh installation step
shows the following messages:

qemu: Unsupported syscall: 293

Reinstalling the qemu-user-static package on the Linux host sometimes

eliminates the messages. These messages are typically harmless and everything still
functions properly.

4229251 On some Linux hosts during flashing, the following message appears in dmesg logs,
followed by flashing failures:

Cannot enable. Maybe the USB cable is bad?

To work around the issue, try connecting to a different USB port on your host
machine. If this step does not help, change the USB cable or reboot the host
machine.

4695663 Initrd flash could fail when flashing with high speed USB port (>= 10Gbps); in that
case, the flashing will fail when trying to untar the file system like this:

tar: Read checkpoint 590000

tar: Read checkpoint 600000
Erased 67108864 bytes from address 0x00000000 in flash
Flash failure
Either the device cannot mount the NFS server on the host or a
flash command has failed. Debug log saved to /tmp/tmp.QwrriPM9V9.
You can access the target's terminal through "sshpass -p root ssh
root@fc00:1:1:0::2"

To fix this, you can try to put an USB hub between the host and the target or use a
lower speed USB port.

NVIDIA Jetson Linux Release Notes RN_10698-r36.4 | 7

2.3. Camera
The following camera-related issues are noted in this release.

Issue Description

3692128
The following issues were observed with the IMX318-CPHY sensor on Jetson AGX
Orin:

● The IMX318-CPHY sensor fails to stream when using the v4l2-ctl

command.
● Streaming using argus_camera and sample apps works, but it might
intermittently output corrupted frames.

4845798 For the HAWK Stereo camera [AR0234 sensor], the first captured image is dark
when using the argus_oneshot or argus_camera applications.

4807063 In the latest v4l2-utils, the VIDIOC_G/S_PARM test fails during the
v4l2-compliance test suite.

4264284 On Jetson Orin, captured images using argus_camera with DOL HDR sensors may
show a marginal increase in noise.

4685929 There may be issues while running argus_camera with the error "JPEG parameter
struct mismatch: library thinks size is 584, caller expects 728". This is due to a
mismatch in the JPEG library where a 3rd party JPEG encoding library was being
picked up.

Users can WAR this issue by PRELOADING the NVIDIA JPEG library before running
the argus_camera app:

"export LD_PRELOAD=/usr/lib/aarch64-linux-gnu/nvidia/libnvjpeg.so"

NVIDIA Jetson Linux Release Notes RN_10698-r36.4 | 8

2.4. Multimedia
The following issues are noted in this release related to multimedia.

Issue Description

4146738 The deinterlace feature is not supported in this release.

4419907 VK rendering is not supported for the AV1 MVC decoder feature.

2.5. Display
The following display issues are noted in this release.

Issue Description

4385047 Hotplugging the display using DP to Jetson AGX Orin after bootup might lead
to a corrupted screen.

4324714 The secondary display connected to Jetson AGX Orin and Jetson Orin NX/Nano
shows as connected in the xrandr utility, but gdm is not rendered on the
screen.

4618846 The CableCreation branded DP-to-HDMI converter does not work as expected
with Jetson AGX Orin. We recommend that you use a different DP-to-HDMI
converter.

4840276 Blank screen after SC7 on Orin AGX, if the Orin AGX is booted with multi user
target mode (ubuntu desktop is disabled and VT terminal is enabled),

NVIDIA Jetson Linux Release Notes RN_10698-r36.4 | 9

2.6. Compute Stack
The following issues related to Deep Learning are noted in this release.

Issue Description

4564075 To run VPI PVA algorithms in a Docker container, the same VPI version has to be installed
on the Docker host.

2.7. DeepStream
The following issues related to video streaming are noted in this release.

Issue Description

4325898 The pipeline gets stuck when using multifilesrc with nvv4l2decoder. DeepStream
developers use this pipeline to run decode and inferJPEG images.

NVIDIA Jetson Linux Release Notes RN_10698-r36.4 | 10

3. Fixed Issues
This section provides details about issues that were noted as issues in the previous
releases but fixed in the current release.

Issue Description

4385287 When only one IMX219 is connected to Jetson Orin NX 16GB/8GB, and a dual IMX219
overlay is applied, the preview and capture with argus_samples and argus_camera fails.

To work around this issue, connect the dual IMX219 cameras when applying dual overlay
support.

4327644 When an IMX219 or an IMX477 camera is connected to Jetson Orin NX and Jetson Orin
Nano devices through the CAM0 interface, to correctly complete a camera recording, users
need to explicitly specify sensor-id=1 instead of the default sensor-id=0 in the
gstreamer nvarguscamera element.

4389380 A half-preview image of IMX477 is observed when you run argus_camera for Orin Nano
4GB SKU4.

4389112 Multimedia sample compilation fails due to a missing header file, and the following
message is displayed:

NvJpegDecoder.cpp:36:10: fatal error: jpegint.h: No such file or

directory

4359668 Applying the 3840x2160@30/24Hz resolution on a display that is connected to Jetson Orin
NX makes the display go blank and the following error message is displayed:

No VSI InfoFrame exists on two video fields.

4349663 When building TensorRT engines for DLA, there is a known issue where the entire DLA
subgraphs that are listed in Layers Running on DLA (seen with TensorRT's verbose mode)
cannot be built and eventually falls back to the GPU with the following message:

{ForeignNode[...]} cannot be compiled by DLA, falling back to GPU.

This has been observed with the following ResNet-based models:

● PeopleNet v2.6
● TrafficCamNet from TAO.

In both cases, to fix this issue, change TensorRT's default DLA SRAM pool size from 1 MiB
to 0.5 MiB. When using trtexec to build the TensorRT engine, add the
--memPoolSize=dlaSRAM:0.5 argument. For other TensorRT applications that directly
call TensorRT APIs, refer to this code section in trtexec.

NVIDIA Jetson Linux Release Notes RN_10698-r36.4 | 11

 Issue Description

4391619 Resnet-10 batch_size=1 GPU-only inferencing is broken on Jetson AGX Orin.

Currently, syncpoint submission is updated to the end of cudaEventRecord() after all the
operations have finished, so that after syncpoint is complete, there should be no pending
work on the marker.

However, when the application thread invokes cudaEventElapsedTime, but the GPU is
still busy and has not yet completed its work, it will fail.

To work around this issue and avoid the application execution failure, invoke the
jetson_clocks utility to max out the SOC clocks and speed up the execution. This
workaround only improves the symptom and is not guaranteed to work every time.

4361621 DRC support in gst pipelines is a work in progress and will require some changes in the
video codec.

4297071 The latest version of GStreamer may have issues with the RTP stack, potentially causing
intermittent segmentation faults.

4617111 The TSEC CBB errors might be observed on Jetson AGX Orin during bootup, shutdown,
4620917 suspend, and resume when the target is connected to a display with HDCP 2.2 support.
This issue is the result of the missing synchronization between Tsec and DCE/display
engines. The HDCP can be disabled by following patch which prevents the CBB errors.

--- a/nv-soc/tegra234-soc-display-overlay.dtsi
+++ b/nv-soc/tegra234-soc-display-overlay.dtsi
@@ -176,7 +176,6 @@
"dpaux0_reset",
"dsi_core_reset",
"mipi_cal_reset";
- hdcp_enabled;
status = "disabled";
memory-region = <&fb0_reserved>;
nvidia,disp-sw-soc-chip-id = <0x2350>;

4185596 Waking up from Deep Sleep state (SC7) by USB events is not supported in the NVIDIA
JetPack 6.1 GA release for the Jetson Orin Series of products.

4494706 Previous releases of JetPack used the legacy Ethernet interface naming scheme (eth0,
eth1, etc.). Starting with this release, it uses the predictable naming scheme for Ethernet
interfaces. As part of this improvement, the parameter net.ifnames=0 has been removed
from the kernel command line.

4412365 During the Debian update process on Jetson AGX Orin with an external NVMe connected
and flashed using SDK Manager, you may see the following message:

FuPluginUefiCapsule cannot find default ESP: More than one available

ESP

NVIDIA Jetson Linux Release Notes RN_10698-r36.4 | 12

 Issue Description

WARNING: UEFI ESP partition not detected or configured

See https://github.com/fwupd/fwupd/wiki/PluginFlag:esp-not-found for
more information.

To recover from this issue:

1. Stop the Debian update.

2. Run the following commands:
$ sudo mkdir /boot/efi
$ sudo mount /dev/nvme0n1p10 /boot/efi
$ sudo vim /etc/fwupd/uefi_capsule.conf
change the OverrideESPMountPoint=/boot/efi
3. To trigger the OTA again, run the following command:
$ sudo apt dist-upgrade

4641609 ToneMap V3 is enabled for the HAWK AR0234 sensor by default on the ISAAC NOVA
platform.

4624028 Improved the stability of the argus_syncstereo app to avoid crashes in the nvargus
4599191 daemon during heavy CPU load.
4585557

4600459 Faded colors that appeared during camera preview with the HAWK [AR0234] sensor were
fixed, improving image quality (IQ) using the latest ISP tuning file.

4588292 Improved synchronization and stability in the argus_syncstereo app.

4460069 Fixed the detection of the IMX390 [GMSL] camera sensor.

4264284 IMX390 WDR mode is enabled in this release, and issues associated with line corruption
issue have been fixed

4262318 These v4l2 compliance tests previously failed or were invalid but have now been fixed in the
latest v4l2 utils:
1. ioctls
2. VIDIOC_REQBUFS/CREATE_BUFS/QUERYBUF

4247835 Line corruption issues associated with v4l2 have been fixed.

4697575 Memory leak issues while running the gstreamer app and during encode/decode use
cases have been fixed in the Argus library.

NVIDIA Jetson Linux Release Notes RN_10698-r36.4 | 13

4. Implementation Details
This section provides information about implementation details.

4.1. Camera
Here are the camera integration changes compared to previous Jetson Linux 35.x releases:

Note: We recommend that all camera drivers be packages like Loadable Kernel Module
(LKM) for JetPack 6 and later.

● Sensor kernel Drivers:

Driver source code for the supported sensors are located in the
<TOP>/kernel/nvidia-oot/drivers/media/i2c/ directory.
To obtain a complete understanding of the driver, examine this source file.
● LKM
The camera and sensor drivers are loadable modules.
The rebuilt sensor driver will be in the
<OUT>/l4t-generic-release-aarch64/nvidia/kernel-jammy-src/kernel-nvidia-oot/nvidia-
oot/drivers/media/i2c/<camera>.ko directory.
● On root file system, the loadable modules are located in the following directories:
○ Camera sensor driver:
/lib/modules/5.15.116-release-tegra/extra/drivers/media/i2c/.
○ Camera driver:
/lib/modules/5.15.116-release-tegra/extra/drivers/media/platform/tegra/camera.
○ RTCPU driver:
/lib/modules/5.15.116-release-tegra/extra/extra/drivers/platform/tegra/rtcpu/.
○ Nvhost VI driver:
/lib/modules/5.15.116-release-tegra/extra/drivers/video/tegra/host/vi/.
○ Nvhost CSI driver:
/lib/modules/5.15.116-release-tegra/extra/drivers/video/tegra/host/nvcsi/.
● Device Registration
After driver development is complete, you must add the new device information to the
system kernel device tree so it can be registered (instantiated) when the kernel boots.
To register your device, use one of the following methods:
○ Device-tree overlay
You need to create a device-tree overlay file to register the camera module. If your
camera module has on-board EEPROM and is programmed with a valid camera ID,

NVIDIA Jetson Linux Release Notes RN_10698-r36.4 | 14

 you can use that to apply the overlay for a specific camera module and update the
device-tree entries with proper information at runtime. Using DTB overlays with
EEPROM-ID allows a single system image to support multiple camera devices.

To change camera modules, power down the device, replace the camera module, and
then reboot. The new module works automatically.
○ Create and apply a DTB Overlay file
1. Add your .dtsi file to the camera configuration .dtsi file.
tegra234-p3737-camera-modules.dtsi is for Jetson AGX-orin.
2. Set the status of your device tree nodes to disabled.
imx274_cam0: imx274_a@1a {
status = "disabled";
};
3. Add the overlay information as fragments below to a new .dts file.
You can also refer to the camera DTB overlay files provided with the current
release:
<top>/hardware/nvidia/t23x/nv-public/overlay/camera-overlay-file.dts
4. Update the .dts file with proper overlay information and a compatible string
/{
overlay-name = "Jetson Camera Dual-IMX274";
jetson-header-name = "Jetson AGX CSI Connector";
compatible = "nvidia,p3737-0000+p3701-0000",
"nvidia,p3737-0000+p3701-0004", "nvidia,p3737-0000+p3701-0005",
"nvidia,p3737-0000+p3701-0008";

fragment@0 {
target-path = "/bus@0/i2c@3180000/tca9546@70/i2c@0/imx274_a@1a";
board_config {
ids = "LPRD-dual-imx274-002";
sw-modules = "kernel";
};
__overlay__ {
status = "okay";
};
};
fragment@1 {
… … …
};
};
5. Compile the .dts file to generate a .dtbo file.
6. Before you flash, move the .dtbo file to flash_folder/kernel/dtb/.

NVIDIA Jetson Linux Release Notes RN_10698-r36.4 | 15

 7. Add the following lines to the <board>.conf file, which is used to flash the device.
For Jetson AGX-orin board, the config file is p3737-0000-p3701-0000.conf.
OVERLAY_DTB_FILE="${OVERLAY_DTB_FILE},tegra234-p3737-camera-dual-imx274-over
lay.dtbo”.

4.2. Device Registration

After you complete the driver development, you must add the new device's information to
the system kernel device tree so it can be registered (instantiated) when the kernel boots.
The following sections describe ways to register a new device.

Before you begin, ensure that you obtain the kernel source files.

4.2.1. Device Tree Overlay

Because UEFI boot is enabled in this release, the plugin manager is no longer supported.
You must create a device tree overlay (DTB overlay or .dtbo) file to register the camera
module.

If your camera module has an on-board EEPROM, and is programmed with a valid camera
ID, at runtime, you can use the device tree overlay file to apply the overlay for a specific
camera module and update the device tree entries with proper information. Using a device
tree overlay with an EEPROM ID allows a system image to support multiple camera devices.
To select a different camera, power down the device, replace the camera module, and
reboot. The new module works automatically.

To create and apply a device tree overlay file:

1. Add the .dtsi file to the camera configuration .dtsi file.

2. Set the status of your device tree nodes to disabled.

imx185_cam0: imx185_a@1a {
status = "disabled";
};

3. Add the overlay information as fragments to a new .dts file.

<top>/hardware/nvidia/platform/t19x/common/kernel-dts/t19x-common-modules/tegra194
-camera-overlay-file.dts

You can also see the camera DTB overlay files that are provided with the current release
for examples.

NVIDIA Jetson Linux Release Notes RN_10698-r36.4 | 16

4. Update the .dts file with the correct overlay information and a compatible string.

/ {
overlay-name = "Jetson Camera Dual-IMX274";
jetson-header-name = "Jetson AGX Xavier CSI Connector";
compatible = "nvidia,p2822-0000+p2888-0001";

fragment@0 {
target= "<&imx185_cam0>";
board_config {
ids = "LPRD-dual-imx274-002" ;
sw-modules = "kernel";
};
__overlay__ {
status = "okay";
};
};
fragment@1 {
. . .
};

};

1. To generate a .dtbo file, compile the .dts file.

2. Before flashing, move the .dtbo file to flash_folder/kernel/dtb/.
3. Add the following line to the <board>.conf file, which is used to flash the device.

OVERLAY_DTB_FILE="${OVERLAY_DTB_FILE},tegra 194-camera-overlay-file.dtbo";

This line causes the following tasks to completed:

● If a specific camera board is found when the kernel boots, the override data is
applied to that camera board's tree nodes.
● The tree nodes are made available for the system to use.

NVIDIA Jetson Linux Release Notes RN_10698-r36.4 | 17

4.2.2. Using the Jetson IO Tool
If your camera module does not have an on-board EEPROM, you can use the same DTB
overlay file to statically configure the board for the attached camera.

1. After you attach the camera module, apply the camera module's DTB overlay using the
Jetson-IO tool, and reboot.

The new module will work immediately after Jetson Linux starts.

Note: You might have to delete the board_config{} node from the fragments in the DTB
overlay file.

2. After you compile the .dts file to generate a .dtbo file, move the .dtbo file to /boot on
the Jetson device, so that the Jetson-IO tool can recognize it.
3. Launch the Jetson-IO tool and configure the DTB overlay.

4.3. UEFI
For fixes that were made in the UEFI sources after the release, go to UEFI GitHub.

NVIDIA Jetson Linux Release Notes RN_10698-r36.4 | 18

 Notice

This document is provided for information purposes only and shall not be regarded as a warranty of a certain functionality, condition, or quality
of a product. NVIDIA Corporation (“NVIDIA”) makes no representations or warranties, expressed or implied, as to the accuracy or completeness
of the information contained in this document and assumes no responsibility for any errors contained herein. NVIDIA shall have no liability for
the consequences or use of such information or for any infringement of patents or other rights of third parties that may result from its use.
This document is not a commitment to develop, release, or deliver any Material (defined below), code, or functionality.

NVIDIA reserves the right to make corrections, modifications, enhancements, improvements, and any other changes to this document, at any
time without notice.

Customer should obtain the latest relevant information before placing orders and should verify that such information is current and complete.

NVIDIA products are sold subject to the NVIDIA standard terms and conditions of sale supplied at the time of order acknowledgement, unless
otherwise agreed in an individual sales agreement signed by authorized representatives of NVIDIA and customer (“Terms of Sale”). NVIDIA
hereby expressly objects to applying any customer general terms and conditions with regards to the purchase of the NVIDIA product
referenced in this document. No contractual obligations are formed either directly or indirectly by this document.

NVIDIA products are not designed, authorized, or warranted to be suitable for use in medical, military, aircraft, space, or life support equipment,
nor in applications where failure or malfunction of the NVIDIA product can reasonably be expected to result in personal injury, death, or
property or environmental damage. NVIDIA accepts no liability for inclusion and/or use of NVIDIA products in such equipment or applications
and therefore such inclusion and/or use is at customer’s own risk.

NVIDIA makes no representation or warranty that products based on this document will be suitable for any specified use. Testing of all
parameters of each product is not necessarily performed by NVIDIA. It is customer’s sole responsibility to evaluate and determine the
applicability of any information contained in this document, ensure the product is suitable and fit for the application planned by customer, and
perform the necessary testing for the application in order to avoid a default of the application or the product. Weaknesses in customer’s
product designs may affect the quality and reliability of the NVIDIA product and may result in additional or different conditions and/or
requirements beyond those contained in this document. NVIDIA accepts no liability related to any default, damage, costs, or problem which may
be based on or attributable to: (i) the use of the NVIDIA product in any manner that is contrary to this document or (ii) customer product
designs.

No license, either expressed or implied, is granted under any NVIDIA patent right, copyright, or other NVIDIA intellectual property right under
this document. Information published by NVIDIA regarding third-party products or services does not constitute a license from NVIDIA to use
such products or services or a warranty or endorsement thereof. Use of such information may require a license from a third party under the
patents or other intellectual property rights of the third party, or a license from NVIDIA under the patents or other intellectual property rights
of NVIDIA.

Reproduction of information in this document is permissible only if approved in advance by NVIDIA in writing, reproduced without alteration and
in full compliance with all applicable export laws and regulations, and accompanied by all associated conditions, limitations, and notices.

THIS DOCUMENT AND ALL NVIDIA DESIGN SPECIFICATIONS, REFERENCE BOARDS, FILES, DRAWINGS, DIAGNOSTICS, LISTS, AND OTHER
DOCUMENTS (TOGETHER AND SEPARATELY, “MATERIALS”) ARE BEING PROVIDED “AS IS.” NVIDIA MAKES NO WARRANTIES, EXPRESSED,
IMPLIED, STATUTORY, OR OTHERWISE WITH RESPECT TO THE MATERIALS, AND EXPRESSLY DISCLAIMS ALL IMPLIED WARRANTIES OF
NONINFRINGEMENT, MERCHANTABILITY, AND FITNESS FOR A PARTICULAR PURPOSE. TO THE EXTENT NOT PROHIBITED BY LAW, IN NO EVENT
WILL NVIDIA BE LIABLE FOR ANY DAMAGES, INCLUDING WITHOUT LIMITATION ANY DIRECT, INDIRECT, SPECIAL, INCIDENTAL, PUNITIVE, OR
CONSEQUENTIAL DAMAGES, HOWEVER CAUSED AND REGARDLESS OF THE THEORY OF LIABILITY, ARISING OUT OF ANY USE OF THIS
DOCUMENT, EVEN IF NVIDIA HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES. Notwithstanding any damages that customer
might incur for any reason whatsoever, NVIDIA’s aggregate and cumulative liability towards customer for the products described herein shall be
limited in accordance with the Terms of Sale for the product.

Trademarks
NVIDIA, the NVIDIA logo, are trademarks and/or registered trademarks of NVIDIA Corporation in the U.S. and other countries. Other company
and product names may be trademarks of the respective companies with which they are associated.

NVIDIA Corporation | 2788 San Tomas Expressway, Santa Clara, CA 95051

http://www.nvidia.com
 VESA DisplayPort
DisplayPort and DisplayPort Compliance Logo, DisplayPort Compliance Logo for Dual-mode Sources, and DisplayPort Compliance Logo for
Active Cables are trademarks owned by the Video Electronics Standards Association in the United States and other countries.

HDMI
HDMI, the HDMI logo, and High-Definition Multimedia Interface are trademarks or registered trademarks of HDMI Licensing LLC.

Arm
Arm, AMBA, and ARM Powered are registered trademarks of Arm Limited. Cortex, MPCore, and Mali are trademarks of Arm Limited. All other
brands or product names are the property of their respective holders. ʺArmʺ is used to represent ARM Holdings plc; its operating company
Arm Limited; and the regional subsidiaries Arm Inc.; Arm KK; Arm Korea Limited.; Arm Taiwan Limited; Arm France SAS; Arm Consulting
(Shanghai) Co. Ltd.; Arm Germany GmbH; Arm Embedded Technologies Pvt. Ltd.; Arm Norway, AS, and Arm Sweden AB.

OpenCL
OpenCL is a trademark of Apple Inc. used under license to the Khronos Group Inc.

Copyright

© 2024 NVIDIA Corporation & Affiliates. All rights reserved.

NVIDIA Corporation | 2788 San Tomas Expressway, Santa Clara, CA 95051

http://www.nvidia.com