Lucas, It could also depend on using ifort 10.0 or 10.1, I guess. With ifort 10.1 I compiled siesta with no hassle, with -O2 and some other options. It's a shame that intel is so predictable but in a bad sense...
Marcos On Sat, Mar 13, 2010 at 5:56 PM, Lucas Fernandez Seivane <[email protected]> wrote: > Yikes, diferent compilers give different bugs. I had not managed to > get siesta 3 beta to work properly and precisely with ifort 10/11 > except with -g -O0... (I did some tweaking in the Makefile to compile > with that flags the pseudopotential/basis parts). If anybody is > suffering the same I can look it up (it may take some time to find the > old files). > > 2010/3/13 Marcos Veríssimo Alves <[email protected]>: >>> Thank's at all for information. >>> Do you think that the problem is connected only with >>> bugs in compiler not with potential bugs in siesta?. >>> I mean that maybe intel fortran sees some bugs in siesta (that's why during >>> compilation some warnings apears) >>> which cannot see g95 compiler? >>> >> >> No, the bugs are not in siesta. They are rather in the way ifort reads >> the code and in the way it handles it. >> >> Marcos >> > > > > -- > Lucas Fernández Seivane > Ph. D. Candidate > Universidad de Oviedo - CINN >
