[Pw_forum] Why my test-calculation is so slowly? SGI4700 for Si take about 1 hours

[Lo_wan_2005XW]

Dear axel

Thank you very much for your help!

Just following your suggestion, I export OMP_NUM_THREADS=1, the calculation finish within several seconds!

Thank you very much for your help! 

The machine I used is SGI-Altex-4700. After I download the code, I just do 
1) configure
2) make all

I donot know if I still other set or configurations to optimize my calculation in SGI-Altex-4700?

Thanks

Xiangang Wan
Department of Physics
Nanjing University
P.R. China

> Date: Mon, 10 Nov 2008 21:55:58 -0500
> From: akohlmey at cmm.chem.upenn.edu
> To: lo_wan_2005 at hotmail.com
> Subject: Re: [Pw_forum] Why my test-calculation is so slowly? SGI4700 for Si take about 1 hours
> 
> On Tue, 11 Nov 2008, Lo_wan_2005XW wrote:
> 
> XW> 
> XW> Dear PWscf users
> 
> dear xw,
> 
> please don't forget to specify your full name and affiliation.
> 
> you are a victim of not overly smart sysadmins and the strange
> choices of intel. you are most likely linking to MKL and as of
> version 10.0 it automatically multi-threads across _all_ available
> cpu 'cores'. on an SGI altix this has disasterous consequences
> for the performance. you have to set OMP_NUM_THREADS=1 and
> it will be _much_ less of a problem (check out the mailing list
> archive for multiple occurances of this "feature".
> 
> please not that the first test examples don't overly scale
> as well (for testing, you may be better off with the contents
> of the tests directory using only 2-4 processors.
> 
> cheers,
>    axel.
> 
> p.s.: if your sysadmins are not totally ignorant, please
> ask them to at export OMP_NUM_THREADS=1 to the default
> environment, and they'll avoid a lot of problems with
> jobs not running overly efficient.
> 
> XW> 
> XW> I just download the code and install it in our SGI-4700 without any warnning or error.
> XW> 
> XW> After that, I try to do the example-calculation (example01).
> XW> 
> XW> My machine is SGI-Altex-4700, so I change the the following line in the file of run_example 
> XW> 
> XW>  
> XW> 
> XW> $PW_COMMAND < si.scf.$diago.in > si.scf.$diago.out
> XW> 
> XW>  
> XW> 
> XW> to 
> XW> 
> XW>  
> XW> 
> XW> $PARA_PREFIX $PW_COMMAND < si.scf.$diago.in > si.scf.$diago.out
> XW> 
> XW>  
> XW> 
> XW> I set 
> XW> 
> XW>  
> XW> 
> XW> PARA_PREFIX="mpirun -np 12"
> XW> 
> XW>  
> XW> 
> XW> After that, I type ./run_example
> XW> 
> XW> I can see the following in the screen:
> XW> 
> XW>  
> XW> 
> XW> sgi4700:~/Quantum-Espresso/espresso-4.0.3/examples/example01> ./run_example 
> XW>  
> XW> /disk2/Quantum-Espresso/espresso-4.0.3/examples/example01 : starting
> XW>  
> XW> This example shows how to use pw.x to calculate the total energy and
> XW> the band structure of four simple systems: Si, Al, Cu, Ni.
> XW>  
> XW>   executables directory: /disk2/Quantum-Espresso/espresso-4.0.3/bin
> XW>   pseudo directory:      /disk2/Quantum-Espresso/espresso-4.0.3/pseudo
> XW>   temporary directory:   /disk2/tmp
> XW>   checking that needed directories and files exist... done
> XW>  
> XW>   running pw.x as: mpirun -np 12 /disk2/Quantum-Espresso/espresso-4.0.3/bin/pw.x -npool 4
> XW>   running bands.x as: mpirun -np 12 /disk2/Quantum-Espresso/espresso-4.0.3/bin/bands.x -npool 4
> XW>  
> XW>   cleaning /disk2/xgwan/tmp... done
> XW>   running the scf calculation for Si...
> XW>  
> XW> 
> XW> 
> XW>  
> XW> 
> XW>  
> XW> 
> XW> By using "top", I can find there are 12 pw.x.  My machine take about 2 hours to finish this simple calculation. It is quite strange! 
> XW> I compare the my output and the data saved in referece-directory, it shows that my result, like total energy, is right comparing with the referenced data.
> XW> 
> XW> Attached pleased find my output data.
> XW> 
> XW> I will appreciate if someone can kindly tell me the possible reason for it. 
> XW> 
> XW> Thanks
> XW> 
> XW> XW
> XW> 
> XW> _________________________________________________________________
> XW> MSN ÖÐÎÄÍø£¬×îÐÂʱÉÐÉú»î×ÊѶ£¬°×Áì¾Û¼¯ÃÅ»§¡£
> XW> http://cn.msn.com
> 
> -- 
> =======================================================================
> Axel Kohlmeyer   akohlmey at cmm.chem.upenn.edu   http://www.cmm.upenn.edu
>    Center for Molecular Modeling   --   University of Pennsylvania
> Department of Chemistry, 231 S.34th Street, Philadelphia, PA 19104-6323
> tel: 1-215-898-1582,  fax: 1-215-573-6233,  office-tel: 1-215-898-5425
> =======================================================================
> If you make something idiot-proof, the universe creates a better idiot.

_________________________________________________________________
MSN资讯快递，帮助你第一时间了解最新资讯！
http://im.live.cn/newsexpress
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20081111/30f78794/attachment.html>