infrence time tensorrt px2

allenchen8210 · December 29, 2018, 9:28am

hi

we have two machines:

1.GPU: 1070ti and CPU: i7-8700k
2.PX2

The same model inference on host is three times faster than px2.

One is 17ms, the other is almost 60ms.

Both iGPU and dGPU on px2 are same speed.

Why are there such a big gap?

AastaLLL · January 2, 2019, 4:17am

Hi,

May I know which model/sample you are using for testing?
To benchmark the performance, it’s recommended to use our built-in sample for testing.
Ex. /usr/src/tensorrt/samples/sample_googlenet

Thanks.

allenchen8210 · January 3, 2019, 5:19am

We are using SSD300 model(TensorRT).

AastaLLL · January 11, 2019, 7:11am

Hi,

Thanks for your feedback.

Have you executed other GPU related app at the same time?
Ex. hyperionlauncher

If yes, please kill the app and try it again.
Thanks.

Topic		Replies	Views
FP16 doesn't bring improvement to inference TensorRT	0	912	May 29, 2019
Detectron2 model converted into Tensorrt taking large time TensorRT	1	554	September 20, 2022
Detection tesnorRT takes seconds to run on TX2 Jetson TX2 tensorrt , jetson-inference	8	652	October 18, 2021
Why my inference time is so long when using trtexec - FP16? Jetson TX2 jetson-inference	4	1959	October 18, 2021
TensorRT poor inference performance on Ampere TensorRT	1	458	February 25, 2021
TensorRT inference is slower than tensorflow model TensorRT	1	954	June 28, 2019
Inference Time between Tesla K80 and GTX 1080 in Tensorflow TensorRT	1	1419	January 3, 2019
Jetson Xavier NX slower than Jetson TX2 at pytorch inferences Jetson Xavier NX benchmarks	4	555	June 29, 2023
There is a difference in inference speed in TensorRT 8 TensorRT tensorrt	4	506	October 28, 2021
Tensorrt is slower than pytorch TensorRT	2	2231	September 15, 2021

infrence time tensorrt px2

Related topics