For AMD machine, you can try the free version of PathScale, which shares the libraries with open64, but has better compability. Further, according to my experience, even for AMD machine, ifort+MKL provide higher efficiency than PathSclae+AMCL, this may be mainly caused by the missing of FFTW (only FFT) in AMCL. Of course, I have not tested every possibility, hope you can find better compiler and math-libraries combination in AMD platform.
-- GAO Zhe CMC Lab, Materials Science & Engineering Department, Seoul National University, South Korea At 2012-06-05 17:11:03,Arturo <artginer at bifi.es> wrote: Now it gives me another error: mpif90 -O3 -ftpp -D__FFTW -D__MPI -D__PARA -I../include -I../../iotk/src -I../../Modules -I. -c cglib.f90 real(DP) :: zmat( MAXVAL(descla(:)%nrlx), nudx, nspin ), fmat( MAXVAL(descla(:)%nrlx), nudx, nspin ), fdiag( nx ) ^ openf95-870 openf90: ERROR CALCMT, File = cglib.f90, Line = 25, Column = 25 The intrinsic call "MAXVAL" is not valid in a specification expression. real(kind=DP) z0( MAXVAL(descla(:)%nrlx), nudx, nspin ) ^ openf95-870 openf90: ERROR ROTATE, File = cglib.f90, Line = 102, Column = 28 The intrinsic call "MAXVAL" is not valid in a specification expression. We have a machine with AMD interlagos processor and the vendor recommend us to use this compiler Open64 from AMD. Thanks for your help!! Arturo El 05/06/12 10:49, Lorenzo Paulatto escribi?: On Tue, Jun 5, 2012 at 10:32 AM, Arturo <artginer at bifi.es> wrote: real(dp), parameter :: fac = (1.d0/2.d0) * 1.d0/(3.d0*pi**2)**(1.d0/3.d0) ^ openf95-206 openf90: ERROR DO_RDG, File = elf.f90, Line = 203, Column = 70 The exponent in a constant initialization expression must be type integer. How can I solve it? Open the file with a text editor and replace the right-hand-side expression with its result: 0.161620459673995d0 But, where did you find this compiler? And, I'm just curious, why did you choose it? Are you sure it is reliable? bests -- Lorenzo Paulatto IdR @ IMPMC/CNRS & Universit? Paris 6 phone: +33 (0)1 44275 084 / skype: paulatz www: http://www-int.impmc.upmc.fr/~paulatto/ mail: 23-24/4?16 Bo?te courrier 115, 4 place Jussieu 75252 Paris C?dex 05 _______________________________________________ Pw_forum mailing list Pw_forum at pwscf.orghttp://www.democritos.it/mailman/listinfo/pw_forum -- Arturo Giner Gracia HPC research group System Administrator Instituto de Biocomputaci?n y F?sica de Sistemas Complejos (BIFI) Universidad de Zaragoza e-mail: artginer at bifi.es phone: (+34) 976762992 -------------- next part -------------- An HTML attachment was scrubbed... URL: http://www.democritos.it/pipermail/pw_forum/attachments/20120605/f315d752/attachment-0001.htm