The purpose for this inquiry is to understand how to set throttling on TX2 and how to monitor throttling on TX2 when CPU/GPU temperature is high. TX2’s Thermal Design Guide talks about this topic, however, I would like to test it on my own.
Could the combination of sudo jetson_clocks --restore command and a host of ‘.conf’ files be helpful in setting the system’s clock frequencies (CPU, GPU, and EMC) to a pre-defined saved configuration.
Playing around with sudo jetson_clocks --show command showed the following results:
~/[TX2] sudo jetson_clocks --show
SOC family:tegra186 Machine:quill
Online CPUs: 0,3-5
CPU Cluster Switching: Disabled
cpu0: Online=1 Governor=schedutil MinFreq=2035200 MaxFreq=2035200 CurrentFreq=2035200 IdleStates: C1=0 c7=0
cpu1: Online=0 Governor=schedutil MinFreq=345600 MaxFreq=2035200 CurrentFreq=2035200 IdleStates: C1=0 c6=0 c7=0
cpu2: Online=0 Governor=schedutil MinFreq=345600 MaxFreq=2035200 CurrentFreq=2035200 IdleStates: C1=0 c6=0 c7=0
cpu3: Online=1 Governor=schedutil MinFreq=2035200 MaxFreq=2035200 CurrentFreq=2035200 IdleStates: C1=0 c7=0
cpu4: Online=1 Governor=schedutil MinFreq=2035200 MaxFreq=2035200 CurrentFreq=2035200 IdleStates: C1=0 c7=0
cpu5: Online=1 Governor=schedutil MinFreq=2035200 MaxFreq=2035200 CurrentFreq=2035200 IdleStates: C1=0 c7=0
GPU MinFreq=1122000000 MaxFreq=1122000000 CurrentFreq=1122000000
EMC MinFreq=40800000 MaxFreq=1600000000 CurrentFreq=1600000000 FreqOverride=1
Fan: speed=0
NV Power Mode: MAXP_CORE_ARM
When throttling happened, dmesg shall have some logs printed.
In the meantime, you could refer to the tegrastats and see what is the current temperature.
Thank you for feedback. Attempting to run Cuda code on TX2 by hello world of cuda in a ‘.cu’ file. Commands ‘lspci | grep -i nvidia’, ‘nvcc --version’ and ‘nvidia-smi’ fail. In addition the ‘usr/local’ directory does not include any ‘cuda’ directory. What is the manual procedure to set up CUDA on TX2 (Ubuntu 18.04)? I am following the native compilation step from Jetson/Installing CUDA - eLinux.org!!!
Need help finding the *.deb file for CUDA 10.0 Toolkit/L4T (releases R32.4)???
Was not able to download file “cuda-repo-cross-aarch64-ubuntu1804-10-2-local_10.2.460-1_all.deb”. I had to build an Ubuntu 18.04 machine and download the Jetsonpack onto it. CDed into directory ‘Linux_for_Tegra’. Ran a find command looking for ‘cuda-repo*.deb’ file. Found the the file and transferred it to TX2 using usb.
sudo dpkg -i cuda-repo-cross-aarch64-ubuntu1804-10-2-local_10.2.460-1_all.deb
sudo apt-get update
sudo apt-key add /var/cuda-repo-cross-aarch64-ubuntu1804-10-2-local/7fa2af80.pub
sudo apt --fix-broken install
sudo apt-get install cuda-toolkit-10-2
echo "### Add 32-bit CUDA library & binary paths: " >> ~/.bashrc
echo “export PATH=/usr/local/cuda-10.2/bin:$PATH” >> ~/.bashrc
echo “export LD_LIBRARY_PATH=/usr/local/cuda-10.2/lib:$LD_LIBRARY_PATH” >> ~/.bashrc
This step allowed me to run the set/monitor command lines for the the GPU cores.
Hi,
If the question here is not same as original topic, please file a new one.