Hello Michel, I compiled SIESTA on a cluster of linux machines with pgf90 and mpi; In fact, I get some errors at the compilation, typically coming from atom.f . What I did is after the first compilation, is to compile only the atom.f WITHOUT optimisation option, and to recompile then siesta (so it passed the atom.f ...). It has worked in my case. It has been mentioned in one of the makefiles proposed. Hope this help in you case.
Regards, Frédéric -----Message d'origine----- De : Siesta, Self-Consistent DFT LCAO program, http://www.uam.es/siesta [mailto:[EMAIL PROTECTED] De la part de Michel Gelize Envoyé : jeudi 4 novembre 2004 16:16 À : SIESTA-L@LISTSERV.UAM.ES Objet : [SIESTA-L] siesta compilation with pgf90 Hi all, I just downloaded SIESTA and am having trouble compiling on a LINUX system with Redhat 7.1 and pgf90 compiler. The problem seems to be related to the code defining some variables as arrays, and later on as variables that are not arrays. Specifically, I get compile errors in atom.f from lines with iz, charge, and basistypes. Also, in basis_specs.f from nzeta. These are all defined in basis_types.f as arrays but used in atom.f and basis_specs.f as if they are not arrays. Thanks for your help. Best regards Michel Gelize