Please provide the following info (tick the boxes after creating this topic):
Software Version
DRIVE OS 6.0.8.1
Target Operating System
Linux
Hardware Platform
DRIVE AGX Orin Developer Kit (940-63710-0010-C00)
SDK Manager Version
other
Host Machine Version
native Ubuntu Linux 20.04 Host installed with SDK Manager
Issue Description
I wanted to use the NvRM APIs to test the power consumption on the drive orin GPU. I discovered that under the nvrm/gpu/ folder in SDK, there’s a clock tool (nvrm_gpuclk) and a temperature tool (nvrm_gpu_temperature). I modified the temperature tool to test power since they have similar API structures.
The program runs well, but the NvRmGpuDeviceListPowerSensors API returns with no sensors, i.e. numSensors=0
When I used the similar original nvrm_gpu_temperature tool, the NvRmGpuDeviceListTemperatureSensors API returned numSensors=0 too.
Does it mean my drive orin doesn’t contain a power sensor or a temperature sensor? Or do I need to make some modifications in the source code in order to make it work?
Error String
- no power sensor available
Logs
Provide logs in text box instead of image
Please paste the complete application log here. If there are multiple logs, please use multiple text box
nvidia@tegra-ubuntu:~$ ./nvrm_gpu_power
Usage: nvrm_gpu_power [-list] [-i=interval] [-r=iterations]
check GPU power
options
-l --list list available devices
-i --interval interval between output in ms, default is 500ms
-r --iterate number of iterations, -1 for infinite
default is set to 1
Note: mandatory arguments to long arguments are mandatory for short
arguments as well.
examples
nvrm_gpu_power --list list all available GPUs and
get corresponding
nvrm_gpu_power -i=1000 -r=20 display the power
for 20 times with
1000 ms interval
nvrm_gpu_power -i=5000 -r=-1 display the power
for infinite times with
5000 ms interval
nvrm_gpu_power -i -r=1000 display the power
for 1000 times with
500(default) ms interval
nvidia@tegra-ubuntu:~$ ./nvrm_gpu_power -l
GPU list:
gpu-index state probe name
2000 attached nvgpu:/dev/nvgpu/igpu0
nvidia@tegra-ubuntu:~$ ./nvrm_gpu_power -i=1000 -r=-1
gpu index not set
try ‘nvrm_gpu_power’ for more information
nvidia@tegra-ubuntu:~$ ./nvrm_gpu_power 2000 -i=1000 -r=-1
- no power sensor available
nvidia@tegra-ubuntu:~$ ./nvrm_gpu_temperature
Usage: nvrm_gpu_temperature [-list] [-cfk] [-i=interval] [-r=iterations]
check GPU temperature
options
-l --list list available devices
-c --celsius check temperature in Celsius(default)
-f --fahrenheit check temperature in Fahrenheit
-k --kelvin check temperature in Kelvin
-i --interval interval between output in ms, default is 500ms
-r --iterate number of iterations, -1 for infinite
default is set to 1
Note: mandatory arguments to long arguments are mandatory for short
arguments as well.
examples
nvrm_gpu_temperature --list list all available GPUs and
get corresponding
nvrm_gpu_temperature -c -f display the temperature in
Celsius and Fahrenheit
nvrm_gpu_temperature -ckf display the temperature in
Celsius, Fahrenheit, Kelvin
nvrm_gpu_temperature -c -i=1000 -r=20 display the temperature in
Celsius for 20 times with
1000 ms interval
nvrm_gpu_temperature -ck -i=5000 -r=-1 display the temperature in
Celsius and Kelvin for
infinite times with
5000 ms interval
nvrm_gpu_temperature -c -f -i -r=1000 display the temperature in
Celsius and Fahrenheit for
1000 times with
500(default) ms interval
nvidia@tegra-ubuntu:~$ ./nvrm_gpu_temperature 2000 -c -i=1000 -r=-1
- no temperature sensor available
nvidia@tegra-ubuntu:~$ ./nvrm_gpu_temperature 2000 -l
GPU list:
gpu-index state probe name
2000 attached nvgpu:/dev/nvgpu/igpu0
nvidia@tegra-ubuntu:~$ ./nvrm_gpu_temperature 2000 -c -i=1000 -r=-1
- no temperature sensor available