Which ifort actually works

Questions regarding the compilation of VASP on various platforms: hardware, compilers and libraries, etc.


Moderators: Global Moderator, Moderator

Post Reply
Message
Author
janbor
Newbie
Newbie
Posts: 8
Joined: Wed Jul 25, 2007 8:38 am
License Nr.: 734

Which ifort actually works

#1 Post by janbor » Thu Jul 26, 2007 9:25 am

I have attempted to compile VASP with Intels ifort 10.0.023 and 9.1.05. They
both fail due to "obsolescent features in Fortran 95". My question is this; how old a version of ifort do I actually need ?

Jan

fortcom: Error: fft3dlib.f90, line 1623: Sharing of a DO termination statement by more than one DO statement is an obsolescent feature in Fortran 95. Use an END DO or CONTINUE statement for each DO statement. [20]
20 CONTINUE
---^
fortcom: Error: fft3dlib.f90, line 1700: The computed GOTO statement is an obsolescent feature in Fortran 95.
GOTO (10,50,90,130,170,210,250),IGO
------^
Last edited by janbor on Thu Jul 26, 2007 9:25 am, edited 1 time in total.

admin
Administrator
Administrator
Posts: 2921
Joined: Tue Aug 03, 2004 8:18 am
License Nr.: 458

Which ifort actually works

#2 Post by admin » Thu Jul 26, 2007 1:47 pm

this error is due to the assigned GOTO, which is not standard for fortran 90 (but for fortran 77)
for (probably all) releases of the intel compiler later than 9.1, please skip the -e95 compiler option for the compilation of fft3dlib in the makefile. This option causes an abortion (rather than just a warning) if fortran 77 statements (which are not standard in fortran 90) are found in the code (as assigned goto-s are.
Last edited by admin on Thu Jul 26, 2007 1:47 pm, edited 1 time in total.

janbor
Newbie
Newbie
Posts: 8
Joined: Wed Jul 25, 2007 8:38 am
License Nr.: 734

Which ifort actually works

#3 Post by janbor » Mon Jul 30, 2007 8:20 am

Thanks. Removing the -e95 option did indeed solve the problem.
Now, however, I have what I believe is another installation problem:
when I run the tutor case "diamant" I get this:
---------------------------------------------------------
vasp.4.6.31 08Feb07 complex
POSCAR found : 1 types and 2 ions
LDA part: xc-table for Ceperly-Alder, standard interpolation
POSCAR, INCAR and KPOINTS ok, starting setup
WARNING: wrap around errors must be expected
FFT: planning ... 1
reading WAVECAR
entering main loop
N E dE d eps ncg rms rms(c)
DAV: 1 -0.758421968191E+01 -0.75842E+01 -0.30607E+03 176 0.581E+02
DAV: 2 -0.205812044133E+02 -0.12997E+02 -0.12609E+02 208 0.107E+02
DAV: 3 -0.206095736249E+02 -0.28369E-01 -0.28369E-01 192 0.580E+00
DAV: 4 -0.206095868597E+02 -0.13235E-04 -0.13235E-04 200 0.961E-02
DAV: 5 -0.206095868631E+02 -0.33222E-08 -0.33221E-08 144 0.112E-03BRMIX: very serious problems
the old and the new charge density differ
old charge density: 8.01760 new 8.03836
0.449E+00
--------------------------

From other posts here I realize this may be a problem with recent
ifort compilers. Someone suggested to change all -O4 to -O3.
I find no -O4 in my makefiles.
I have also tried changing all settings: ISMEAR, SIGMA, NEDOS,
PREC.... nothing works. Any suggestions?

Jan
Last edited by janbor on Mon Jul 30, 2007 8:20 am, edited 1 time in total.

janbor
Newbie
Newbie
Posts: 8
Joined: Wed Jul 25, 2007 8:38 am
License Nr.: 734

Which ifort actually works

#4 Post by janbor » Tue Jul 31, 2007 8:06 am

I solved the problem by changing the -O3 option, in Intels ifort
makefile, to -O1

Jan
Last edited by janbor on Tue Jul 31, 2007 8:06 am, edited 1 time in total.

jialy_25
Newbie
Newbie
Posts: 8
Joined: Fri Apr 21, 2006 9:00 am

Which ifort actually works

#5 Post by jialy_25 » Tue Aug 07, 2007 8:44 am

Hi, janbor. I would like to know which ifc and atlas library are you using. I am trying to compile vasp with ifc 10.0.23 and mkl 9.1.021. However, i can't compile succesfully with the error as below:

rm -f vasp
ifort -ipo -o vasp -L/usr/local/ifc/lib/ -lsvml main.o base.o mpi.o smart_allocate.o xml.o constant.o jacobi.o main_mpi.o scala.o asa.o .......(skip)..................elpol.o setlocalpp.o aedens.o fftw3d.o fft3dlib.o /usr/lib/libfftw3.a -L../vasp.4.lib -ldmy ../vasp.4.lib/linpack_double.o -lmkl_lapack /usr/local/intel/mkl/lib/32 -lmkl_p4 -lguide -lpthread
backend signals

fortcom: Severe: **Internal compiler error: internal abort** Please report this error along with the circumstances in which it occurred in a Software Problem Report. Note: File and line given may not be explicit cause of this error.

ifort: error #10014: problem during multi-file optimization compilation (code 3)
make: *** [vasp] Error 3


Does anyone have idea what cause this error? thanks.
Last edited by jialy_25 on Tue Aug 07, 2007 8:44 am, edited 1 time in total.

janbor
Newbie
Newbie
Posts: 8
Joined: Wed Jul 25, 2007 8:38 am
License Nr.: 734

Which ifort actually works

#6 Post by janbor » Mon Aug 13, 2007 10:27 am

I use the same ifort and mkl as you. Not using Atlas - only mkl.
From what I can see your error does not suggest the compiler has a problem with the VASP code - it has a problem with itself. Erroneous installation of ifort/mkl/icc or unsupported hardware would be my guess.

Jan
Last edited by janbor on Mon Aug 13, 2007 10:27 am, edited 1 time in total.

jialy_25
Newbie
Newbie
Posts: 8
Joined: Fri Apr 21, 2006 9:00 am

Which ifort actually works

#7 Post by jialy_25 » Mon Aug 20, 2007 1:43 am

Thanks. I got it solved by using "-lmkl" instead of "-lmkl_p4" for blas library linking, so that the appropriate .so file will be dispatched based on cpuid check.
Last edited by jialy_25 on Mon Aug 20, 2007 1:43 am, edited 1 time in total.

Post Reply