Thread cpu error on tx2

https://scontent.fhan3-3.fna.fbcdn.net/v/t1.15752-9/53792520_2488668704499568_3873340892106457088_n.png?_nc_cat=100&_nc_oc=AQkxP4sR-BW94caCiyrU6xL9aajTFj4aikAakPTV4XyaDtENdo2aiB-dYeCRuZF4aMo&_nc_ht=scontent.fhan3-3.fna&oh=1ccf02d9fcfeb0633a00f6e690fd2e10&oe=5D0CF7A4
I am using tx2 and i encountered errors all the programs are running on cpu1. CPUs 2, 3, 4 are frozen. please help

How to reproduce this issue?

help me. I want to operate on all 4 CPUs

You will need to give information on how to reproduce any error. The command producing the load would be nice to see. Any input used would also be good to know about.

The problem that I encountered is that there is only 1 core of the cpu working and the remaining 3 cores freeze.

Unless the program was written to work on more than one core then only one core will run. That isn’t a bug. Thus, the command used must be known to find out if the software you are using was intended to run on more cores.

One other consideration if it is performance you are after is to set a high performance mode instead of an energy saving mode before you run your program:

sudo nvpmodel -m 0
sudo ~ubuntu/jetson_clocks.sh

Even on the largest super computer made only one core will be used unless the software is designed for all cores…in which case you would require a faster core instead of more cores.

thank you. I have tried multithreading, but the result is still running on one cpu core

Please see if OpenMP API helps with your use cases. You may refer to https://devtalk.nvidia.com/default/topic/1048446/jetson-tx2/-openmp-performance-bug-in-denver-cores-/ .

Is the program actually written to use more cores? If not, then it isn’t possible for the system to use multiple cores…and it would also be a bad idea since it would remove the benefits of cache.