| ROS Resources: Documentation | Support | Discussion Forum | Index | Service Status | ros @ Robotics Stack Exchange |
I have successfully built autoware.ai from source with CUDA enabled. The rosbag demo also runs properly.
My question is: will the node like lidar_euclidean_cluster_detect automatically utilize gpu so long as it compiled with cuda? Is there any way to check whether the node is using gpu at run-time? Thanks!
You can run nvidia-smi in your terminal and see what processes are using your gpu. Combine this with watch and you can get pseudo real-time resource usage. watch -n 1 nvidia-smi
To confirm if your docker is compatible with GPU, run nvidia-docker run nvidia/cuda:10.0-base nvidia-smi.
Do this check only if you are using a docker to run your application.
Asked: 2021-01-23 18:03:37 -0500
Seen: 374 times
Last updated: Jan 24 '21
ROS Answers is licensed under Creative Commons Attribution 3.0 Content on this site is licensed under a Creative Commons Attribution Share Alike 3.0 license.
