Memory Leak Issue in cudaNvSci Sample

[haihua.wei]

Please provide the following info (tick the boxes after creating this topic):
Software Version
DRIVE OS 6.0.8.1
DRIVE OS 6.0.6
[*] DRIVE OS 6.0.5
DRIVE OS 6.0.4 (rev. 1)
DRIVE OS 6.0.4 SDK
other

Target Operating System
[*] Linux
QNX
other

Hardware Platform
DRIVE AGX Orin Developer Kit (940-63710-0010-300)
DRIVE AGX Orin Developer Kit (940-63710-0010-200)
DRIVE AGX Orin Developer Kit (940-63710-0010-100)
DRIVE AGX Orin Developer Kit (940-63710-0010-D00)
DRIVE AGX Orin Developer Kit (940-63710-0010-C00)
DRIVE AGX Orin Developer Kit (not sure its number)
[*] other

SDK Manager Version
1.9.3.10904
[*] other

Host Machine Version
native Ubuntu Linux 20.04 Host installed with SDK Manager
native Ubuntu Linux 20.04 Host installed with DRIVE OS Docker Containers
native Ubuntu Linux 18.04 Host installed with DRIVE OS Docker Containers
[*] other

Dear Nvidia Support Team,

I hope this message finds you well. I am currently working with the Nvidia CUDA interface and have encountered an issue while running the cudaNvSci sample from CUDA 11.4. Upon inspecting the sample using the valgrind tool, I identified a memory leak problem, and the error message is provided in the attached screenshot.

In an effort to incorporate this sample into our own application, we followed the same procedures, utilizing cudaExternalMemoryGetMappedMipmappedArray for mipmappedArray resource allocation. This is done to facilitate the mapping of an nvsci buffer to a CUDA array. Additionally, we employed cudaFreeMipmappedArray for mipmappedArray resource release. However, we are encountering the same memory leak problem in our application.

I would appreciate your assistance in identifying potential causes for this issue and providing guidance on how to address it.

Thank you very much for your help!

Best regards

[VickNV]

Could you please provide information about the hardware platform you’re using? Additionally, we recommend upgrading to DRIVE OS 6.0.8.1, as it may include bug fixes.

Furthermore, to investigate the memory leak issue, could you share the complete command and output of running valgrind? This will help us analyze the problem more comprehensively.

[haihua.wei]

Our system is based on Baidu’s ACU, using the OrinX chipset. For memory leak analysis, we ran the following Valgrind command:

/opt/data/zyp/valgrind/bin/valgrind --tool=memcheck --leak-check=full --log-file=valgrind.log.txt ./cudanvsci ./

Regarding the output results, the memory leak reported corresponds to the screenshot I provided earlier. Thanks for your ongoing support.

[VickNV]

This forum is specifically designated for addressing inquiries related to the P3710 devkit. Considering that your system is built on Baidu’s ACU and utilizes the OrinX chipset, I presume you may have another dedicated support channel. It is advisable to consistently report issues through that channel for more efficient assistance.

[haihua.wei]

Our devkit version is p3710-10-s05, and we are running this CUDA sample in the Driver OS version 6.0.5 environment. As illustrated in the attached image, we are able to replicate the same issue.

[VickNV]

There is no update from you for a period, assuming this is not an issue any more.
Hence we are closing this topic. If need further support, please open a new one.
Thanks

Thank you for sharing the details. To troubleshoot the issue more effectively, could you please try running the CUDA sample on the latest release, 6.0.8.1? Let us know if the issue is also reproducible.