Resnet50 on NVIDIA device "condition: mEngineCounter.use_count() == 1 Destroying a runtime before destroying deserialized engines"

[mahmoodn]

Any idea about the following runtime error for resnet50? Apparently, the workload is running and I can see the result. So, I am not sure how much important is that.

[2024-07-01 07:52:15,243 __init__.py:46 INFO] Running command: ./build/bin/harness_default --logfile_outdir="/work/build/logs/2024.07.01-07.51.45/rtx3080_ryzen3700x_TRT/resnet50/Offline" 
--logfile_prefix="mlperf_log_" --performance_sample_count=2048 --gpu_res2res3_loop_count=1 
--gpu_copy_streams=2 --gpu_inference_streams=1 --run_infer_on_copy_streams=false --gpu_batch_size=1 
--map_path="data_maps/imagenet/val_map.txt" --tensor_path="build/preprocessed_data/imagenet/ResNet50/int8_linear"
 --use_graphs=false --gpu_engines="./build/engines/rtx3080_ryzen3700x/resnet50/Offline/resnet50-Offline-gpu-b1-int8.lwis_k_99_MaxP.plan" 
--mlperf_conf_path="build/loadgen-configs/rtx3080_ryzen3700x_TRT/resnet50/Offline/mlperf.conf" 
--user_conf_path="build/loadgen-configs/rtx3080_ryzen3700x_TRT/resnet50/Offline/user.conf" 
--max_dlas=0 --scenario Offline --model resnet50
[2024-07-01 07:52:15,243 __init__.py:53 INFO] Overriding Environment
benchmark : Benchmark.ResNet50
...
...
[I] [TRT] [MemUsageChange] TensorRT-managed allocation in engine deserialization: CPU +0, GPU +19, now: CPU 0, GPU 19 (MiB)
[I] Device:0.GPU: [0] ./build/engines/rtx3080_ryzen3700x/resnet50/Offline/resnet50-Offline-gpu-b1-int8.lwis_k_99_MaxP.plan has been successfully loaded.
[E] [TRT] 3: [runtime.cpp::~Runtime::401] Error Code 3: API Usage Error (Parameter check failed at: runtime/rt/runtime.cpp::~Runtime::401, condition: mEngineCounter.use_count() == 1 Destroying a runtime before destroying deserialized engines created by the runtime leads to undefined behavior.)
[I] [TRT] [MemUsageChange] Init cuBLAS/cuBLASLt: CPU +0, GPU +10, now: CPU 58, GPU 666 (MiB)
[I] [TRT] [MemUsageChange] Init cuDNN: CPU +0, GPU +8, now: CPU 58, GPU 674 (MiB)

...
...
No warnings encountered during test.

No errors encountered during test.
Finished running actual test.
Device Device:0.GPU processed:

  660000 batches of size 1
  Memcpy Calls: 0
  PerSampleCudaMemcpy Calls: 0
  BatchedCudaMemcpy Calls: 660000
&&&& PASSED Default_Harness # ./build/bin/harness_default
[2024-07-01 07:57:12,498 run_harness.py:165 INFO] Result: result_samples_per_second: 2266.11, Result is INVALID