When running deep learning programs, sometimes the program is forced to terminate, but the GPU resources occupied by the program are still not released. After being trapped for a long time, it is thought that the GPU has been occupied by others. As a result, the GPU resources are leaked.
You can use this command to view the usage of GPU in Linux system
nvidia-smiThe result is as shown in the figure
At this time, you can manually kill the process that occupies the GPU to release the GPU resources
kill -9 49461If the screen command is used, the program running in the background stops and occupies the GPU, you can also close all screen windows to release the GPU
killall screenOf course, it’s OK to kill the process directly
Read More:
- Through PID (process identification) to find the port (port) occupied applications, to solve the problem of port occupied
- Oracle stops a job
- Android error: ADB port is occupied（ adb.exe ,start-server’ failed — run manually if necessary)
- Solution to the problem that listen TCP 0.0.0.0:3306: bind: address already in use port is occupied in Linux centos7
- Appium connecting to the ADB 5037 port of nocturnal simulator is occupied by itself
- [OpenGL] cannot start this program because the computer is missing glut32.dll. Try to re install the program to fix this problem. …
- Kafka opens JMX port and reports that the error port is occupied
- Idea startup error: Port occupied/address already exists
- Python program uses OS. System () method to call exe program, resulting in no response of main program process
- Eclipse port occupied（ java.net.BindException : address already in use: bind) solution
- This program cannot be started because vcruntime140 is missing from your computer_ 1.dll。 Try to install the program again to solve the problem.
- After the new video card rtx3060 arrives, configure tensorflow and run “TF. Test. Is”_ gpu_ The solution of “available ()” output false
- After the vs2013 + OpenGL environment is set up, there is an error in running the first program
- Solution to stray’\357′ in program when gcc is compiled
- RuntimeError: CUDA out of memory. Tried to allocate 600.00 MiB (GPU 0； 23.69 GiB total capacity)
- Error 1500. Another program is in progress. You must complete another installation before continuing with this installation.
- Angular: Program ng failed to run No application is associated
- Vscode running C program error luanch:program does not exsist
- Ant Design ‘cross env’ is not an internal or external command, nor is it an error reporting problem for a runnable program
- GPU hardware acceleration related problems, solve flash screen