Orin AGX Jetpack 6 NVDEC limitation

Robotics & Edge Computing Jetson & Embedded Systems Jetson AGX Orin
1

quick follow up question on this thread

Orin AGX Jetpack 6 NVDEC limitation of <= 8 streams - Jetson & Embedded Systems / Jetson AGX Orin - NVIDIA Developer Forums

Does the patch provided on the above thread allows running hardware decoder with gstreamer for > 8 processes?
We are experiencing the issue on jetpack 6.2 where we run gstreamer decoders on more than 8 processes, and is this a regression in jetpack 6 compared to jetpack 5?

2

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

Based on the provided context, I’ll address the question and provide a comprehensive answer.

The question refers to a thread on the NVIDIA Developer Forums discussing a limitation of the NVDEC (NVIDIA Video Decoder Engine) on Jetson AGX Orin with Jetpack 6, where it’s limited to <= 8 streams. The thread provides a patch to potentially resolve this issue.

To answer the question, we need to understand the context of the NVDEC limitation and the patch provided. The NVDEC is a hardware-accelerated video decoder, and the limitation of <= 8 streams might be a restriction imposed by the hardware or the software stack.

The patch provided in the thread aims to allow running the hardware decoder with GStreamer for more than 8 processes. However, it’s essential to understand that the patch might not necessarily resolve the issue for all use cases.

Comparing Jetpack 6 to Jetpack 5, it’s possible that the regression occurred due to changes in the software stack or the NVDEC driver. Without further information, it’s difficult to pinpoint the exact cause of the regression.

To address the question directly:

  • The patch provided in the thread might allow running the hardware decoder with GStreamer for more than 8 processes, but it’s not guaranteed to resolve the issue for all use cases.
  • The regression in Jetpack 6 compared to Jetpack 5 might be due to changes in the software stack or the NVDEC driver, but further investigation is required to confirm this.

Additional information is needed to provide a more definitive answer:

  • More details about the patch and its implementation would help understand its effectiveness in resolving the issue.
  • Information about the specific use case and the error messages encountered when running more than 8 processes with GStreamer would provide more context.
  • Comparison of the NVDEC driver and software stack between Jetpack 5 and Jetpack 6 would help identify potential causes of the regression.

In summary, while the patch provided in the thread might offer a solution, it’s essential to carefully evaluate its effectiveness and consider the specific use case and error messages encountered. Further investigation is required to determine the root cause of the regression in Jetpack 6 compared to Jetpack 5.

4

Hi,
As suggested above, please apply the patch set if you would like to run multiple decoding processes.

6

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.