DeepStream 6.1.1 not utilizing GPU after upgrading from 6.1 (docker image)

Please provide complete information as applicable to your setup.

• Hardware Platform (Jetson / GPU)
dGPU (Titan V)
• DeepStream Version
6.1.1 (using deepstream:6.1.1-triton docker image)
• NVIDIA GPU Driver Version (valid for GPU only)
515.76
• Issue Type( questions, new requirements, bugs)
I am developing an application for which, after upgrading my docker base image from 6.1 to 6.1.1, it appears to run an order of magnitude slower and is not utilizing the GPU. No other changes were made. Just wondering if there was a configuration change that could cause this? I can attempt to reproduce in a sample app if there is nothing obvious.

DeepStream 6.1.1 can work with DeepStream 6.1 configuration file. Can you elaborate more details or provide the steps to reproduce the problem?

Hi, it looks like this might actually be just a difference in logging or a transient condition on initialization. In 6.1 the reported FPS started higher than the source frame rate, then went down to the nominal frame rate, whereas in 6.1.1 the frame rate starts low and after a while catches up to the nominal frame rate. Here is an example using 6.1.1:

**PERF:  0.00 (0.00)    0.00 (0.00)
** INFO: <bus_callback:194>: Pipeline ready

** INFO: <bus_callback:180>: Pipeline running

**PERF:  1.33 (1.07)    1.33 (1.07)
**PERF:  1.16 (1.14)    1.16 (1.14)
**PERF:  1.45 (1.24)    1.25 (1.16)
**PERF:  8.36 (3.47)    8.31 (3.41)
**PERF:  10.25 (4.89)   10.25 (4.84)
**PERF:  10.05 (5.78)   10.05 (5.74)
**PERF:  9.95 (6.40)    9.95 (6.37)
**PERF:  10.02 (6.87)   10.02 (6.84)
**PERF:  9.98 (7.22)    9.98 (7.20)

There also seems to be more jitter in the reported frame rate compared to 6.1 (used to be solid 10.0 most of the time) but not enough to cause concern.

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.