please refer to [DS5.0GA_Jetson_GPU_Plugin] Measure of the FPS of pipeline in DeepStream SDK FAQ - #9 by mchi to measure the fps.
What your platform ? NANO or Xavier?
Could you measure the trtexec performance of your trtengine by referring to GitHub - NVIDIA-AI-IOT/deepstream_tao_apps: Sample apps to demonstrate how to deploy models trained with TAO on DeepStream ?