Thermal Throttling Issue on Orin NX 16GB (JP 6.2 Super Mode)

Robotics & Edge Computing Jetson Systems Jetson Orin NX
1

Hi Team,

I’m using an Jetson Orin NX 16GB module running JetPack 6.2 in Super Mode. When both the CPU and GPU are stressed under 40W and MAXN Super Mode at full load, I see the following warning message:

“Hot surface, Do not touch system throttled”.

image
image569×727 74.6 KB
image
image589×727 73.5 KB

Could you please advise how to resolve this issue and achieve maximum performance?

Currently, the software thermal throttling limit appears to be 99°C.

  • Can this limit be increased to a higher value?

  • What would be the safe maximum temperature setting to maintain performance without risking hardware damage?

Thanks

Khushbu

2

*** Please note that this reply is generated by LLM automatically ***

Based on the provided context, I’ll address the issues and questions raised by Khushbu.

Issue: Warning message “Hot surface. Do not touch. System throttled.”

This warning message indicates that the system is experiencing high temperatures, and thermal throttling is engaged to prevent overheating. This is a safety mechanism to prevent damage to the hardware.

Cause:

The warning message is likely caused by the high power consumption of the CPU and GPU when both are stressed under full load. This exceeds the Thermal Design Power (TDP) budget, triggering thermal throttling.

Resolution:

To resolve this issue and achieve maximum performance, consider the following:

  1. Cooling System: Ensure that the cooling system is functioning correctly. Check if the heatsink is properly attached, and the fans are working as expected. Consider upgrading to a more efficient cooling solution if necessary.
  2. Thermal Settings: The software thermal throttling limit is set to prevent overheating. However, you can try adjusting the thermal settings to maintain performance without risking hardware damage.
  3. MAXN Mode: Since you’re experiencing thermal throttling, you might be running in MAXN mode. As mentioned in the context, MAXN mode is an experimental setting for adjusting clock settings and creating custom power profiles. It’s not recommended for prolonged heavy workloads. Consider switching to a different power mode or adjusting the MAXN settings to balance performance and power consumption.
  4. Power Estimator: Use the Power Estimator tool to estimate the power consumption of your workload and generate a custom power mode configuration file. This can help you optimize performance while staying within the TDP budget.
  5. Thermal Design Guide: Refer to the Jetson Orin NX Series and Jetson Orin Nano Series Modules Thermal Design Guide (TDG-11127-001_v1.5) for more information on thermal management and design considerations.

Increasing the software thermal throttling limit:

While it’s possible to adjust the thermal settings, increasing the software thermal throttling limit is not recommended. The current limit is set to prevent overheating and ensure safe operation. Exceeding this limit can lead to hardware damage.

Setting to maintain performance without risking hardware damage:

To maintain performance without risking hardware damage, consider the following:

  1. Monitor temperatures: Keep an eye on the system temperatures using tools like tmon or nvidia-smi.
  2. Adjust power modes: Experiment with different power modes, such as MAXN or custom power profiles, to find the optimal balance between performance and power consumption.
  3. Optimize workloads: Optimize your workloads to reduce power consumption and heat generation.
  4. Cooling system upgrades: Consider upgrading the cooling system to improve heat dissipation.

By following these guidelines, you should be able to achieve maximum performance while ensuring safe operation and preventing hardware damage.

*** Please give the thumbs up if you get answers you like. Or provide feedback to help us improve the answer. ***

5

Hi khushbu.rakholiya,

For the best performance, you need the respective cooling method on your board.
You may also need to create the custom power mode configuration according to your use case and what you run.

We don’t suggest to increase the threshold of temperature to prevent the damage to the module.