The ABINIT Users Mailing List ( CLOSED )

["D. R. Hamann" <drh@bell-labs.com>]

Dear Hu,

I don't know about the MPI error, but the elastic tensor error may be
caused by a recently discovered bug in Src_6response/ewald4.f.  One
line should be added following line 96.  This and a few nearby lines
are attached.  This is an initialization bug, and only showed up on a
few platforms.

Don Hamann

kxyhu@yahoo.com.cn wrote:

Hello,

I use a sgi3800 ,and ABINIT 4.3.2 . There has a MPI erro, which is beyond the MPI_GROUP_MAX ,what is the meaning?  
When I run the Test_v4/t75.in using abinis,there are zero divided by zero in the output file 
 Rigid-atom elastic tensor , in cartesian coordinates,  
    j1       j2             matrix element              
 dir pert dir pert     real part    imaginary part      
                                                        
  1    5   1    5    0.0047961362    0.0000000000       
  1    5   2    5    0.0021679259    0.0000000000       
  1    5   3    5    0.0021679259    0.0000000000       
  1    5   1    6   -0.0000000001    0.0000000000       
  1    5   2    6             NaN    0.0000000000       
  1    5   3    6             NaN    0.0000000000       
                                                        
  2    5   1    5    0.0021679218    0.0000000000       
  2    5   2    5    0.0047961402    0.0000000000       
  2    5   3    5    0.0021679259    0.0000000000       
  2    5   1    6   -0.0000000001    0.0000000000       
  2    5   2    6    0.0000000000    0.0000000000       
  2    5   3    6    0.0000000000    0.0000000000       
                                                        
  3    5   1    5    0.0021679218    0.0000000000       
  3    5   2    5    0.0021679259    0.0000000000       
  3    5   3    5    0.0047961401    0.0000000000       
  3    5   1    6             NaN    0.0000000000       
  3    5   2    6   -0.0994762006    0.0000000000       
  3    5   3    6             NaN    0.0000000000       
                                                        
  1    6   1    5    0.0000000000    0.0000000000       
  1    6   2    5    0.0000000000    0.0000000000       
  1    6   3    5             NaN    0.0000000000       
  1    6   1    6             NaN    0.0000000000       
  1    6   2    6    0.0000000000    0.0000000000       
  1    6   3    6    0.0000000000    0.0000000000       
                                                        
  2    6   1    5             NaN    0.0000000000       
  2    6   2    5    0.0000000000    0.0000000000       
  2    6   3    5   -0.0994762005    0.0000000000       
  2    6   1    6    0.0000000000    0.0000000000       
  2    6   2    6    0.0030446917    0.0000000000       
  2    6   3    6    0.0000000000    0.0000000000       
                                                        
  3    6   1    5             NaN    0.0000000000       
  3    6   2    5    0.0000000000    0.0000000000       
  3    6   3    5             NaN    0.0000000000       
  3    6   1    6    0.0000000000    0.0000000000       
  3    6   2    6    0.0000000000    0.0000000000       
  3    6   3    6    0.0030446917    0.0000000000       

What is the reasion?

Thank you
hu

  

-- 
D. R. Hamann				Phone: 908-582-4454
Director, Theoretical Materials    	Fax:   908-582-4702
  Physics Research (retired)		email: drh@physics.bell-labs.com
Bell Laboratories 
Lucent Technologies
700 Mountain Ave, Room 1D-371
Murray Hill, NJ 07974-0636 USA

! Loop over 1st strain index, upper triangle only
  do is1=1,is2

   ka=idx(2*is1-1);kb=idx(2*is1)
   d2rm(:,:,is1,is2)=0.d0
   d2gm(:,:,is1,is2)=0.d0  !***BUG FIX***
   do jj = 1,3
    if(ka==kg) d2rm(:,jj,is1,is2)=d2rm(:,jj,is1,is2)&
&               +rprimd(kb,:)*rprimd(kd,jj)+rprimd(kd,:)*rprimd(kb,jj)