On 6/1/06, David van der Spoel <[EMAIL PROTECTED]> wrote:
> Trying withou the "-ipo" flag now. It still yelds the same weird warning
> (warning #147: declaration is incompatible with "double invsqrt(double)"
> (declared at line 565 of "/usr/local/intel/include/math.h")
rename all gromacs invsqrt calls to gmx_invsqrt
Just a stupid question. No, unfortunatelly I still didn't found a magic flag (trying to be in touch with the intel premier support still). Anyway, I was trying to grep every invsqrt in gromacs directories, and after verbosing out strings like "invsqrta", "_invsqrt", "invsqrtdata", "vecinvsqrt", "invsqrtm", "invsqrtfracttab" and "invsqrtexptab", and also some comments stuff... Well, the command line below:
grep -Hni 'invsqrt' -r * | grep -v 'Arquivo' | grep -v '#' | grep -v 'icc' | grep -v 'gcc' | grep -v 'pgi' | grep -v 'intel' | grep -vi '.bak' | grep -v 'html' | grep -vi 'invsqrta' | grep -v '_invsqrt' | grep -v 'invsqrtdata' | grep -v 'vecinvsqrt' | grep -v 'invsqrtm' | grep -v 'Software' | grep -v 'fast' | grep -v 'invsqrtfracttab' | grep -v 'invsqrtexptab' | grep -v 'version'
I finally arrived to 57 entries to invsqrt. Just before start, and since I'm not good at C (F90 addicted), are those 57 entries in only 12 files all that I should have to modify (I expected little more)? Just in case, I'm attaching the output from that load of "piped greps".
I'll be working on it tonight, and hope tomorrow I'll have more news on it. I'll, if those are the only lines of code to change.
Despite of that, I've made a trial without the --disable-x86-64-sse flag for config. The compilation fails at another point now:
*************************************************************************************************************************************
/bin/sh ../../../../libtool --mode=compile /usr/bin/gcc -O3 -ffast-math -funroll-all-loops -fpeel-loops -c -o nb_kernel010_x86_64_sse.lo nb_kernel010_x86_64_sse.s
libtool: compile: unable to infer tagged configuration
libtool: compile: specify a tag with `--tag'
make[5]: ** [nb_kernel010_x86_64_sse.lo] Erro 1
*************************************************************************************************************************************
I found some information on this kind of error from libtool on google (something about compiler changing in different parts of compilation process), but I'm not finding a way around this. Does anybody has any idea on this, please?
Thanks a lot for everything in advance.
Jones
include/vec.h:54: real invsqrt(real x)
include/vec.h:146:static inline real invsqrt(real x)
include/vec.h:466: cos=ip*invsqrt(ipa*ipb); /* 7 */
include/vec.h:674: linv=invsqrt(norm2(src));
src/mdlib/csettle.c:256: axlng = invsqrt(xaksxd * xaksxd + yaksxd * yaksxd +
zaksxd * zaksxd);
src/mdlib/csettle.c:257: aylng = invsqrt(xaksyd * xaksyd + yaksyd * yaksyd +
zaksyd * zaksyd);
src/mdlib/csettle.c:258: azlng = invsqrt(xakszd * xakszd + yakszd * yakszd +
zakszd * zakszd);
src/mdlib/csettle.c:296: cosphi = tmp*invsqrt(tmp);
src/mdlib/csettle.c:305: cospsi = tmp2*invsqrt(tmp2);
src/mdlib/csettle.c:323: sinthe = (alpa * gama - beta *
tmp2*invsqrt(tmp2)) / al2be2;
src/mdlib/csettle.c:328: costhe = tmp2*invsqrt(tmp2);
src/mdlib/vsite.c:307: c=b*invsqrt(iprod(temp,temp));
src/mdlib/vsite.c:328: invdij = invsqrt(iprod(xij,xij));
src/mdlib/vsite.c:334: b1 = b*invsqrt(iprod(xp,xp));
src/mdlib/vsite.c:380: d=c*invsqrt(iprod(temp,temp));
src/mdlib/vsite.c:631: invl=invsqrt(iprod(xix,xix));
src/mdlib/vsite.c:703: invdij = invsqrt(iprod(xij,xij));
src/mdlib/vsite.c:710: invdp = invsqrt(iprod(xperp,xperp));
src/mdlib/vsite.c:860: invl=invsqrt(iprod(xix,xix));
src/mdlib/clincs.c:199: rlen = invsqrt(tmp0*tmp0+tmp1*tmp1+tmp2*tmp2);
src/mdlib/clincs.c:287: mvb = blc[b]*(len - u0*invsqrt(u0));
src/mdlib/clincs.c:364: li->blc[i] = invsqrt(invmass[a1] + invmass[a2]);
src/mdlib/clincs.c:586: len = r2*invsqrt(r2);
src/tools/gmx_nmeig.c:90:
mass_fac=invsqrt(top->atoms.atom[i].m*top->atoms.atom[k].m);
src/tools/gmx_nmeig.c:112: mass_fac =
invsqrt(top->atoms.atom[j].m);
src/tools/gmx_nmeig.c:156:
mass_fac=invsqrt(top->atoms.atom[iatom].m*top->atoms.atom[katom].m);
src/tools/gmx_nmeig.c:174: mass_fac =
invsqrt(top->atoms.atom[j].m);
src/tools/calcpot.c:123: rinv1O = invsqrt(rsqO);
src/gmxlib/orires.c:302: invr = invsqrt(r2);
src/gmxlib/orires.c:460: invr = invsqrt(r2);
src/gmxlib/disre.c:160: rt_1 = invsqrt(rt2);
src/gmxlib/disre.c:328: weight_rt_1 = invsqrt(rt2);
src/gmxlib/ewald_util.c:200: rinv = invsqrt(dr2);
src/gmxlib/ewald_util.c:287: rinv = invsqrt(dr2);
src/gmxlib/nonbonded/nb_free_energy.c:159: rinv =
invsqrt(rsq);
src/gmxlib/shift_util.c:237: R_1 = invsqrt(r2);
src/gmxlib/shift_util.c:239: R = invsqrt(R_2);
src/gmxlib/shift_util.c:361: dr_1 = invsqrt(dr2);
src/gmxlib/bondfree.c:191: dr = dr2*invsqrt(dr2);
/* 10 */
src/gmxlib/bondfree.c:200: fbond = -two*be*temp*cbomtemp*invsqrt(dr2);
/* 9 */
src/gmxlib/bondfree.c:250: dr = dr2*invsqrt(dr2);
/* 10 */
src/gmxlib/bondfree.c:381: dr = dr2*invsqrt(dr2); /* 10
*/
src/gmxlib/bondfree.c:394: fbond *= invsqrt(dr2); /* 6
*/
src/gmxlib/bondfree.c:438: dr = dr2*invsqrt(dr2); /* 10
*/
src/gmxlib/bondfree.c:446: fbond *= invsqrt(dr2); /* 6
*/
src/gmxlib/bondfree.c:522: r_nW = invsqrt(iprod(nW,nW));
src/gmxlib/bondfree.c:526: r_OD = invsqrt(iprod(dOD,dOD));
src/gmxlib/bondfree.c:600: r12_1 = invsqrt(r12sq); /* 5 */
src/gmxlib/bondfree.c:721: cik=st*invsqrt(nrkj2*nrij2); /* 12
*/
src/gmxlib/bondfree.c:781: dr = dr2*invsqrt(dr2); /* 10
*/
src/gmxlib/bondfree.c:804: cik=st*invsqrt(nrkj2*nrij2); /* 12
*/
src/gmxlib/bondfree.c:833: fbond *= invsqrt(dr2); /* 6
*/
src/gmxlib/bondfree.c:907: cik=st*invsqrt(nrkj2*nrij2); /* 12
*/
src/gmxlib/bondfree.c:972: nrkj = nrkj2*invsqrt(nrkj2); /* 10 */
src/gmxlib/bondfree.c:1278: c = st*invsqrt(nrij2*nrkl2); /* 11
*/
src/gmxlib/bondfree.c:1568: rij_1 = invsqrt(iprod(r_ij,r_ij));
src/gmxlib/bondfree.c:1569: rkj_1 = invsqrt(iprod(r_kj,r_kj));
_______________________________________________ gmx-users mailing list gmx-users@gromacs.org http://www.gromacs.org/mailman/listinfo/gmx-users Please don't post (un)subscribe requests to the list. Use the www interface or send it to [EMAIL PROTECTED] Can't post? Read http://www.gromacs.org/mailing_lists/users.php
