yade-users team mailing list archive

[gaoxuesong <question682916@xxxxxxxxxxxxxxxxxxxxx>]

New question #682916 on Yade:
https://answers.launchpad.net/yade/+question/682916

Hi, I have some questions about the cores and threads in parallel mode. 

1.  What's the difference between the --cores and -j ? In the documentation, they almost have the same defination. 
    --cores  Set number of OpenMP threads (as –threads) and in addition set affinity of threads to the cores given. 
    -j 	Number of OpenMP threads to run 

2. What can we get from the parameter of --performance? 
     I have tried this parameter, it shows the time consuming of different engine parts for different number of particles. So it will help to improve the program?  

3. In the section of Controlling parallel computation, there are to two ways to  set the cores or threads for the job. One is the environmental variable OMP_NUMBER_THREADS, and the other is the -j  X option. Also it says the cores number can be found in the /proc/cpuinfo. i have looked up this file, it has 8 processors and each has 1 cores. In fact, i set 8 cores to the ubuntu system in the VM. Does it mean my processors are single core processors? 

4.  How to set the right cores for a job? 
    In the graphical controller created when yades stars, there is a item named iter. It shows the current iter and a velocity. Does this velocity is the calculation speed, iter per second? If so, is it feasible to try the number of cores used by this velocity? 

Thanks, 
Xuesong 

-- 
You received this question notification because your team yade-users is
an answer contact for Yade.