problem in interface
in [Fortran]
|
Prev: What exactly does the SAVE statement do ?
Next: "static" explanation. What exactly does it mean?
From: mecej4 on 26 May 2010 10:33 rudra wrote: > There is no improvement/change when I compiled everything without > openmp That's to be expected. Given buggy serial code, adding OpenMP on top of it is like putting a rose garland around the neck of a starved horse and expecting it to win the race. -- mecej4
From: rudra on 26 May 2010 11:26 On May 26, 7:28 pm, mecej4 <mecej4.nyets...(a)operamail.com> wrote: > > Three people have told you that your subroutine calls have the arguments > mismatched in number and type. So far you have ignored these posts. > Not really!! I have managed to cleanup the argument list, and that I was about to post. But then I also found Les's post and make it what he has said. which now looks like as: 1 implicit none 2 3 ! ***********************************************************************! 4 interface 5 subroutine lm_a(oidxdn,opnu,opold,oqnu,oqold, & 6 clabl_1,lmx,&! w_oidn,w_opp,w_oqnu,w_oqold, & 7 d_ialpha,d_ifmt3d,d_itrans,d_jbasdn,d_ldn,d_lmaxw, & 8 d_ltmax,d_mdn,d_mmixat,d_mmixpq,d_nbas,d_nclass,d_ndimin, & 9 d_ngen,d_nit,d_nitat,d_niter,d_nkdmx,d_nkxyz,d_nl,d_nopts, & 10 d_norder,d_npts,d_nrxc,d_nrxyz,d_nsp, & 11 d_as,d_beta,d_dele,d_deltr,d_efermi,d_kap2,d_kfit, & 12 d_emax,d_emaxc,d_emin,d_eminc,d_eps,d_facvol,d_gamma,d_ommax1, & 13 d_ommax2,d_range,d_rmaxes,d_rmaxs,d_rmaxs2,d_rmines,d_rms2, & 14 d_rmsdel,d_tolef,d_toleto,d_tolews,d_vmtz,d_wc,d_width,switch, & 15 asnit) 16 17 integer::d_ialpha,d_ifmt3d,d_itrans,d_jbasdn,d_ldn,d_lmaxw, & 18 d_ltmax(3),d_mdn,d_mmixat,d_mmixpq,d_nbas,d_nclass,d_ndimin, & 19 d_ngen,d_nit,d_nitat,d_niter,d_nkdmx,d_nkxyz(3),d_nl,d_nopts, & 20 d_norder,d_npts,d_nrxc,d_nrxyz(3),d_nsp,oidxdn,opnu,opold,oqnu, & 21 oqold,lmx(200),lmxrho(200),asnit!,oidxdn,opnu,opold,oqnu,oqold 22 double precision::d_as,d_beta,d_dele,d_deltr,d_efermi,d_kap2, & 23 d_emax,d_emaxc,d_emin,d_eminc,d_eps,d_facvol,d_gamma,d_ommax1(3)& 24 ,d_ommax2(3),d_range,d_rmaxes,d_rmaxs,d_rmaxs2,d_rmines,d_rms2, & 25 d_rmsdel,d_tolef,d_toleto,d_tolews,d_vmtz,d_wc,d_width,d_kfit 26 logical:: switch(31) 27 character(4)::clabl_1 28 end subroutine lm_a 29 30 subroutine lm_end(oidxdn,opnu,opold,oqnu,oqold, & 31 clabl_1,lmx,&! w_oidn,w_opp,w_oqnu,w_oqold, & 32 ialpha,ifmt3d,itrans,jbasdn,ldn,lmaxw, & 33 ltmax, mdn,mmixat,mmixpq, nbas,d_nclass, ndimin,ngen,nit, & 34 nitat,niter,nkdmx,nkxyz,d_nl,nopts,norder,npts,nrxc,nrxyz,d_nsp,& 35 as, beta,dele,deltr, efermi,kap2,kfit,emax,emaxc,emin, & 36 eminc,eps,facvol,gamma,ommax1,ommax2,range,rmaxes, & 37 rmaxs,rmaxs2,rmines,rms2,rmsdel,tolef,toleto,tolews,vmtz,wc, & 38 width,switch) 39 40 integer::d_ialpha,d_ifmt3d,d_itrans,d_jbasdn,d_ldn,d_lmaxw, & 41 ltmax(3),d_mdn,d_mmixat,d_mmixpq,d_nbas,d_nclass,d_ndimin, & 42 d_ngen,d_nit,d_nitat,d_niter,nkdmx,nkxyz(3),d_nl,d_nopts, & 43 d_norder,d_npts,d_nrxc,nrxyz(3),d_nsp,oidxdn,opnu,opold,oqnu, & 44 oqold,lmx(200),lmxrho(200) 45 character(4)::clabl_1 46 double precision::as,beta,dele,deltr,efermi,kap2, & 47 emax,emaxc,emin,eminc,eps,facvol,gamma,ommax1(3) & 48 ,ommax2(3),range,rmaxes,rmaxs,rmaxs2,rmines,rms2, & 49 rmsdel,tolef,toleto,tolews,vmtz,wc,width,kfit 50 logical switch(31) 51 end subroutine lm_end 52 53 subroutine lm_b(oidxdn_b,opnu_b,opold_b,oqnu_b,oqold_b, & 54 clabl_1_b,lmx_b,d_ialpha_b,d_ifmt3d_b, & 55 d_itrans_b,d_jbasdn_b,d_ldn_b,d_lmaxw_b,d_ltmax_b,d_mdn_b, & 56 d_mmixat_b,d_mmixpq_b,d_nbas_b,d_nclass_b,d_ndimin_b,d_ngen_b, & 57 d_nit_b,d_nitat_b,d_niter_b,d_nkdmx_b,d_nkxyz_b,d_nl_b, & 58 d_nopts_b,d_norder_b,d_npts_b,d_nrxc_b,d_nrxyz_b,d_nsp_b,d_as_b,& 59 d_beta_b,d_dele_b,d_deltr_b,d_efermi_b,d_kap2_b,d_kfit_b, & 60 d_emax_b,d_emaxc_b,d_emin_b,d_eminc_b,d_eps_b,d_facvol_b, & 61 d_gamma_b,d_ommax1_b,d_ommax2_b,d_range_b,d_rmaxes_b,d_rmaxs_b, & 62 d_rmaxs2_b,d_rmines_b,d_rms2_b,d_rmsdel_b,d_tolef_b,d_toleto_b, & 63 d_tolews_b,d_vmtz_b,d_wc_b,d_width_b,switch,asnit) 64 65 integer::d_ialpha_b,d_ifmt3d_b,d_itrans_b,d_jbasdn_b,d_ldn_b, & 66 d_lmaxw_b,d_ltmax_b(3),d_mdn_b,d_mmixat_b,d_mmixpq_b,d_nbas_b, & 67 d_nclass_b,d_ndimin_b,d_ngen_b,d_nit_b,d_nitat_b,d_niter_b, & 68 d_nkdmx_b,d_nkxyz_b(3),d_nl_b,d_nopts_b,d_norder_b,d_npts_b, & 69 d_nrxc_b,d_nrxyz_b(3),d_nsp_b,oidxdn_b,opnu_b,opold_b,oqnu_b, & 70 oqold_b,lmx_b(200),asnit 71 72 character(4)::clabl_1_b 73 logical::switch(31) 74 double precision::d_as_b,d_beta_b,d_dele_b,d_deltr_b,d_efermi_b,& 75 d_kap2_b, d_emax_b,d_emaxc_b,d_emin_b,d_eminc_b,d_eps_b, & 76 d_facvol_b,d_gamma_b,d_ommax1_b(3),d_ommax2_b(3),d_range_b, & 77 d_rmaxes_b,d_rmaxs_b,d_rmaxs2_b,d_rmines_b,d_rms2_b,d_rmsdel_b, & 78 d_tolef_b,d_toleto_b,d_tolews_b,d_vmtz_b,d_wc_b,d_width_b, & 79 d_kfit_b 80 end subroutine lm_b 81 82 end interface 83 84 ! LMTO variables 85 integer::d_ialpha,d_ifmt3d,d_itrans,d_jbasdn,d_ldn,d_lmaxw, & 86 d_ltmax(3),d_mdn,d_mmixat,d_mmixpq,d_nbas,d_nclass,d_ndimin, & 87 d_ngen,d_nit,d_nitat,d_niter,d_nkdmx,d_nkxyz(3),d_nl,d_nopts, & 88 d_norder,d_npts,d_nrxc,d_nrxyz(3),d_nsp,oidxdn,opnu,opold,oqnu, & 89 oqold,lmx(200),lmxrho(200),nit 90 character(4)::clabl_1 91 logical::switch(31) 92 93 double precision::d_as,d_beta,d_dele,d_deltr,d_efermi,d_kap2, & 94 d_emax,d_emaxc,d_emin,d_eminc,d_eps,d_facvol,d_gamma,d_ommax1(3)& 95 ,d_ommax2(3),d_range,d_rmaxes,d_rmaxs,d_rmaxs2,d_rmines,d_rms2, & 96 d_rmsdel,d_tolef,d_toleto,d_tolews,d_vmtz,d_wc,d_width,d_kfit 97 98 integer::d_ialpha_b,d_ifmt3d_b,d_itrans_b,d_jbasdn_b,d_ldn_b, & 99 d_lmaxw_b,d_ltmax_b(3),d_mdn_b,d_mmixat_b,d_mmixpq_b,d_nbas_b, & 100 d_nclass_b,d_ndimin_b,d_ngen_b,d_nit_b,d_nitat_b,d_niter_b, & 101 d_nkdmx_b,d_nkxyz_b(3),d_nl_b,d_nopts_b,d_norder_b,d_npts_b, & 102 d_nrxc_b,d_nrxyz_b(3),d_nsp_b,oidxdn_b,opnu_b,opold_b,oqnu_b, & 103 oqold_b,lmx_b(200) 104 105 double precision::d_as_b,d_beta_b,d_dele_b,d_deltr_b,d_efermi_b,& 106 d_kap2_b, d_emax_b,d_emaxc_b,d_emin_b,d_eminc_b,d_eps_b, & 107 d_facvol_b,d_gamma_b,d_ommax1_b(3),d_ommax2_b(3),d_range_b, & 108 d_rmaxes_b,d_rmaxs_b,d_rmaxs2_b,d_rmines_b,d_rms2_b,d_rmsdel_b, & 109 d_tolef_b,d_toleto_b,d_tolews_b,d_vmtz_b,d_wc_b,d_width_b, & 110 d_kfit_b 111 character(4)::clabl_1_b 112 113 114 call lm_a(oidxdn,opnu,opold,oqnu,oqold,clabl_1,lmx, & 115 d_ialpha,d_ifmt3d,d_itrans,d_jbasdn,d_ldn, & 116 d_lmaxw,d_ltmax,d_mdn,d_mmixat,d_mmixpq,d_nbas,d_nclass, & 117 d_ndimin,d_ngen,d_nit,d_nitat,d_niter,d_nkdmx,d_nkxyz,d_nl, & 118 d_nopts,d_norder,d_npts,d_nrxc,d_nrxyz,d_nsp,d_as,d_beta,d_dele,& 119 d_deltr,d_efermi,d_kap2,d_kfit,d_emax,d_emaxc,d_emin,d_eminc, & 120 d_eps,d_facvol,d_gamma,d_ommax1,d_ommax2,d_range,d_rmaxes, & 121 d_rmaxs,d_rmaxs2,d_rmines,d_rms2,d_rmsdel,d_tolef,d_toleto, & 122 d_tolews,d_vmtz,d_wc,d_width,switch,nit) 123 124 125 call lm_end(oidxdn,opnu,opold,oqnu,oqold,clabl_1,lmx, & 126 d_ialpha,d_ifmt3d,d_itrans,d_jbasdn,d_ldn, & 127 d_lmaxw,d_ltmax,d_mdn,d_mmixat,d_mmixpq,d_nbas,d_nclass, & 128 d_ndimin,d_ngen,d_nit,d_nitat,d_niter,d_nkdmx,d_nkxyz,d_nl, & 129 d_nopts,d_norder,d_npts,d_nrxc,d_nrxyz,d_nsp,d_as,d_beta,d_dele,& 130 d_deltr,d_efermi,d_kap2,d_kfit,d_emax,d_emaxc,d_emin,d_eminc, & 131 d_eps,d_facvol,d_gamma,d_ommax1,d_ommax2,d_range,d_rmaxes, & 132 d_rmaxs,d_rmaxs2,d_rmines,d_rms2,d_rmsdel,d_tolef,d_toleto, & 133 d_tolews,d_vmtz,d_wc,d_width,switch) 134 135 end $ ifort trylmto.f90 -L/home/Rudra/LMTO-ASR-DISTRO_NEW/LMTO-A -llm_a $ so, compilation is done! offcourse some runtime error is still there and thats why i asked for idea...if you read my last but 1 post, what I really asked is to give some better idea of splitting the code.
From: dpb on 26 May 2010 13:25 rudra wrote: .... > ... what I really asked is to give some better idea of splitting the > code. To amplify my response somewhat from earlier on the subject... That would require stepping back and looking at the overall objective of the project as a start, followed by an evaluation of what the pieces are at present and then a plan to get to the object end. Once one has that overall grasp, then one can begin to look at the structure of the existing codes and see what makes sense from their present structure (or, sadly, lack thereof is what one finds far too often in legacy code). At that point some decisions can be made as to how to proceed; sometimes, even, the logical decision is to throw the baby out w/ the bath water and start again if the code base is already truly horrible and the objective is valuable (and particularly if it is long term rather than simply a one-time hack). But, assuming the answer is to try to meld some pieces together, it surely should be possible after six months to have enough of an idea of how the code works and what it does that there is some logic in making data structures of related parts that can be placed in call lists as composites, other data that may be so widely interspersed as to be likely candidates for module-level variables, and such general levels of abstraction. Then, of course, there's the idea of factorization of the code itself into compilation units that do a more restricted piece of the puzzle so they _don't_ have to know the whole world to operate. That isn't a five-minute task by any means, granted. But, you've spent six-months and have apparently nothing to show so another month or so that might lead to some results afterwards wouldn't seem to bad an investment. Of course, there's also the expedient (assuming this is real work and not either coursework or just personal entertainment) of hiring some "real" help either to do the project or at consult on it. (And, no, I personally really, really don't want to be that person; I am happily retired (from technical consulting anyway) and am fully intending on staying that way (unless, of course, somebody makes an offer I just can't refuse... :) ). Anyway, while in theory it's possible you can get through the jumbled tangled mess you have posted and come out the other side, I personally think a rehash and reboot would likely lead to the promised land at least as quickly and would have a higher value on the other end. --
From: Richard Maine on 26 May 2010 14:40 dpb <none(a)non.net> wrote: > That isn't a five-minute task by any means, granted. But, you've spent > six-months and have apparently nothing to show so another month or so > that might lead to some results afterwards wouldn't seem to bad an > investment. Unless the 6 months has all been spent attempting to do the kinds of spot fixes that so many of the questions seem to indicate, without much looking at the bigger picture, in which case another month of the same is likely to give the same results. > Of course, there's also the expedient (assuming this is real work and > not either coursework or just personal entertainment) of hiring some > "real" help either to do the project or at consult on it. > > (And, no, I personally really, really don't want to be that person; I am > happily retired (from technical consulting anyway) and am fully > intending on staying that way (unless, of course, somebody makes an > offer I just can't refuse... :) ). Very much the same. And for those who might think such professional consulting is cheap. Well, I've never actually established rates, as, like dpb, I'd rather stay retired. But I was making a bit over $500/day plus benefits at a regular, full-time job when I retired. As a general rule, consulting usually runs several times the daily rate of a full-time job, and does tend to be billed in daily increments, even for part of a day, and comes with no guarantee of results. Something that just matched the "general rule" would not count as an offer I could not refuse. No, I'm not holding my breath. -- Richard Maine | Good judgment comes from experience; email: last name at domain . net | experience comes from bad judgment. domain: summertriangle | -- Mark Twain
From: Louis Krupp on 26 May 2010 15:26
On 5/26/2010 8:16 AM, rudra wrote: > There is no improvement/change when I compiled everything without > openmp > And exactly what happened? What error message(s) did you get? Have you made sure you're recompiling absolutely everything? That's always a potential problem with makefiles: Unless the makefile itself is a dependency, you have to a "make clean" (or the manual equivalent) if you change anything in the makefile. Louis |