cpu memory leak when change batch with setDeviceMemory

[qq332982511]

Description

cpu memory leak when change batch with setDeviceMemory

Environment

TensorRT Version: 7.2.2.3
NVIDIA GPU: T4
NVIDIA Driver Version: 470.57.02
CUDA Version: 11.4
CUDNN Version: 8.2.1
Operating System: Ubuntu 18.04.5
Python Version (if applicable):
Tensorflow Version (if applicable):
PyTorch Version (if applicable):
Baremetal or Container (if so, version):

Relevant Files

Steps To Reproduce

instance context with createExecutionContextWithoutDeviceMemory
call setDeviceMemory to set pointer when batch changed
asan and tcmalloc show that setDeviceMemory -> cudnnCreatePoolingDescriptor lead to memory leak

[zerollzeng]

Can you provide a reproduce code? cc @jhalakp-nvidia for viz

[qq332982511]

Can you provide a reproduce code? cc @jhalakp-nvidia for viz
Reproduce code in here
https://github.com/qq332982511/TensorRTIssue/tree/main/pooling_memleak

[zerollzeng]

I notice you are using TRT 7.2, can you do a test with the latest TRT 8.4? if there are still memory leaks I will take a look at it.

Thanks!

[qq332982511]

I notice you are using TRT 7.2, can you do a test with the latest TRT 8.4? if there are still memory leaks I will take a look at it.

Thanks!

no leak on TensorRT-8.4.3.1

[zerollzeng]

Okay can we close this?