Memory not released on deepstream-app exit

Please provide complete information as applicable to your setup.

• Hardware Platform (Jetson / GPU) Xavier
• DeepStream Version 5.10
• JetPack Version (valid for Jetson only) 4.5.1
• TensorRT Version 7.1.3.0
• NVIDIA GPU Driver Version (valid for GPU only)
• Issue Type( questions, new requirements, bugs) bug

Hello,

I’m monitoring constant memory increase when restart the original deepstream-app with Yolo.

On every app exit there are about 20-30 MB unreleased (with num-sources=1).
The sequenxe is run the app with
deepstream-app -c deepstream_app_config_yoloV3.txt
press q to exit after 10 seconds.
Here are the copy/paste from top command:

// Initial
KiB Mem : 32686756 total, 25444008 free, 5742720 used, 1500028 buff/cache
KiB Swap: 16343360 total, 16343360 free, 0 used. 27514172 avail Mem

// After first quit
KiB Mem : 32686756 total, 25420296 free, 5766384 used, 1500076 buff/cache
KiB Swap: 16343360 total, 16343360 free, 0 used. 27514184 avail Mem

// After second quit
KiB Mem : 32686756 total, 25395632 free, 5791000 used, 1500124 buff/cache
KiB Swap: 16343360 total, 16343360 free, 0 used. 27513252 avail Mem

So, we have 20+ MB that are never released.

Here is my config file:
deepstream_app_config_yoloV3.txt (3.7 KB)

There aren’t any modification, it is untouched installed from sdkmanager.

The memory remains used. The only way to release it is with reboot the Xavier or with “sync; echo 3 > /proc/sys/vm/drop_caches”, but that seems not to be a solution.

How I may fix that?

Sorry for the late.
Will try to repro your issue, thanks for your patience.

similar issue with Memory leak in test5-app with smart-record

Is there any solution or workaround ?

Another 2 weeks and no answer!
Any update?
Please do not redirect me to issues which are not solved or closed due to timeout.

Thanks for the patience.
We will fix it in the future release.

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