AMBER Archive (2008)

Subject: RE: AMBER: AMBER10 Compilation (SGI Altix)

From: Ross Walker (ross_at_rosswalker.co.uk)
Date: Mon Sep 15 2008 - 21:15:54 CDT

Hi Anthony,

This is not good... :-(. Can you try compiling in serial with the ifort
target and see what happens there. It is really strange that you see all the
NaN's etc, this normally means some kind of memory corruption, the sort of
thing you'd see calling a 32 bit library from a 64 bit routine for example,
although one would normally expect some kind of segfault here. The compiler
shouldn't let you do this though.

Thus more likely there is a routine, probably in the dihedral code, that is
being misvectorized by the compiler. My first suggestion would be to try
upgrading to the version 10 compilers if you can and see if this fixes it.
If you can't do that or it is difficult, then try editing config_amber.h and
changing the FFLAGS and FOPTFLAGS lines so they don't do any optimization
(-O0) and remove any -axW etc flags, then make clean and recompile. If this
works then it is a case of trying to track down which routine is causing the
compiler problem, my first guess would be ephi.f and/or extra_pts.f in which
case you could edit the Makefile so these are compiled with FFLAGS instead
of FOPTFLAGS (and recreate config_amber.h by re-running configure_amber) and
see if this helps, again make clean before building.

Let us know how things go - hopefully just upgrading the compilers will fix
the problem.

All the best
Ross

> -----Original Message-----
> From: owner-amber_at_scripps.edu [mailto:owner-amber_at_scripps.edu] On Behalf
> Of Anthony Cruz
> Sent: Monday, September 15, 2008 11:22 AM
> To: amber_at_scripps.edu
> Subject: AMBER: AMBER10 Compilation (SGI Altix)
>
> Hi Users:
> Today I try to compile AMBER10 on a SGI Altix (uname -a : Linux gelato
> 2.6.5-7.283-sn2 #1 SMP Wed Nov 29 16:55:53 UTC 2006 ia64 ia64 ia64
> GNU/Linux)
> To configure I use sgi_altix option (icc v 9.0 , ifort v9.0 , MKL v9.0).
> It
> seems to compile without error but all the test fail. Also some of the
> test
> from amber tools fail. I check the netcdf (make check) and pass all test.
> Here is the output for the test.serial (TEST_FAILURES.diff):
>
> possible FAILURE: check mdout.dmp.dif
> /tmp/amber/amber10/test/dmp
> 67,68c67,68
> < 1 -4.2572E+1 1.2493E+1 5.6294E+1 P8 6
> < BOND = 0.7921 ANGLE = 2.0379 DIHED =
> 0.7121
> ---
> > 1 -4.2573E+1 1.2492E+1 5.6291E+1 P8 6
> > BOND = 0.7921 ANGLE = 2.0379 DIHED =
> 0.7120
> 74,75c74,75
> < 2 -4.2582E+1 1.2460E+1 5.6128E+1 P8 6
> < BOND = 0.7874 ANGLE = 2.0325 DIHED =
> 0.7121
> ---
> > 2 -4.2582E+1 1.2459E+1 5.6125E+1 P8 6
> > BOND = 0.7874 ANGLE = 2.0325 DIHED =
> 0.7120
> 81,82c81,82
> < 3 -4.2593E+1 1.2420E+1 5.5930E+1 P8 6
> < BOND = 0.7817 ANGLE = 2.0260 DIHED =
> 0.7120
> ---
> > 3 -4.2593E+1 1.2420E+1 5.5927E+1 P8 6
> > BOND = 0.7817 ANGLE = 2.0260 DIHED =
> 0.7119
> 88,89c88,89
> < 4 -4.2606E+1 1.2373E+1 5.5692E+1 P8 6
> < BOND = 0.7750 ANGLE = 2.0182 DIHED =
> 0.7119
> ---
> > 4 -4.2606E+1 1.2373E+1 5.5689E+1 P8 6
> > BOND = 0.7750 ANGLE = 2.0182 DIHED =
> 0.7118
> 95,96c95,96
> < 5 -4.2622E+1 1.2317E+1 5.5407E+1 P8 6
> < BOND = 0.7669 ANGLE = 2.0089 DIHED =
> 0.7117
> ---
> > 5 -4.2622E+1 1.2316E+1 5.5404E+1 P8 6
> > BOND = 0.7669 ANGLE = 2.0089 DIHED =
> 0.7116
> 104,105c104,105
> < 5 -4.2622E+1 1.2317E+1 5.5407E+1 P8 6
> < BOND = 0.7669 ANGLE = 2.0089 DIHED =
> 0.7117
> ---
> > 5 -4.2622E+1 1.2316E+1 5.5404E+1 P8 6
> > BOND = 0.7669 ANGLE = 2.0089 DIHED =
> 0.7116
> ---------------------------------------
> possible FAILURE: check mdout.adenine.dif
> /tmp/amber/amber10/test/adenine
> 69,70c69,70
> < 1 1.4956 4.7935 2.7454E+1 N3 25
> < BOND = 0.4545 ANGLE = 5.3215 DIHED =
> 17.5374
> ---
> > 1 1.5174 4.7855 2.7453E+1 N3 25
> > BOND = 0.4545 ANGLE = 5.3215 DIHED =
> 17.5592
> 76,77c76,77
> < 2 1.2908 3.4147 1.7746E+1 N3 25
> < BOND = 0.4346 ANGLE = 5.1757 DIHED =
> 17.5358
> ---
> > 2 NaN NaN 0. H5T 1
> > BOND = 0.4346 ANGLE = 5.1751 DIHED = NaN
> 79c79
> < 1-4 VDW = 7.1923 1-4 EEL = 6.1795 RESTRAINT =
> 0.
> ---
> > 1-4 VDW = 7.1922 1-4 EEL = 6.1795 RESTRAINT =
> 0.
> 83,137c83,137
> < 3 1.1266 2.1977 9.9540 N3 25
> < BOND = 0.4585 ANGLE = 5.0362 DIHED =
> 17.5348
> < VDWAALS = -3.6387 EEL = -24.8259 HBOND =
> 0.
> < 1-4 VDW = 7.1691 1-4 EEL = 6.1467 RESTRAINT =
> 0.
> < EPOLAR = -6.7540
> < Dipole convergence: rms = 0.258E-6 iters = 3.00
> < NSTEP ENERGY RMS GMAX NAME
> NUMBER
> < 4 1.0010 1.6429 7.4065 EP23 26
> < BOND = 0.4942 ANGLE = 4.9350 DIHED =
> 17.5342
> < VDWAALS = -3.6421 EEL = -24.7961 HBOND =
> 0.
> < 1-4 VDW = 7.1421 1-4 EEL = 6.0892 RESTRAINT =
> 0.
> < EPOLAR = -6.7554
> < Dipole convergence: rms = 0.659E-6 iters = 3.00
> < NSTEP ENERGY RMS GMAX NAME
> NUMBER
> < 5 9.2259E-1 3.1051 1.3689E+1 EP23 26
> < BOND = 0.4804 ANGLE = 4.9120 DIHED =
> 17.5322
> < VDWAALS = -3.6456 EEL = -24.7132 HBOND =
> 0.
> < 1-4 VDW = 7.1263 1-4 EEL = 5.9910 RESTRAINT =
> 0.
> < EPOLAR = -6.7605
> < Dipole convergence: rms = 0.766E-6 iters = 3.00
> < NSTEP ENERGY RMS GMAX NAME
> NUMBER
> < 6 1.1652 8.9082 4.6701E+1 N3 25
> < BOND = 0.4929 ANGLE = 5.1589 DIHED =
> 17.5309
> < VDWAALS = -3.6507 EEL = -24.7333 HBOND =
> 0.
> < 1-4 VDW = 7.1146 1-4 EEL = 6.0155 RESTRAINT =
> 0.
> < EPOLAR = -6.7637
> < Dipole convergence: rms = 0.269E-6 iters = 4.00
> < NSTEP ENERGY RMS GMAX NAME
> NUMBER
> < 7 8.6292E-1 2.7405 1.4587E+1 N3 25
> < BOND = 0.4660 ANGLE = 4.8797 DIHED =
> 17.5307
> < VDWAALS = -3.6498 EEL = -24.7002 HBOND =
> 0.
> < 1-4 VDW = 7.1191 1-4 EEL = 5.9808 RESTRAINT =
> 0.
> < EPOLAR = -6.7633
> < Dipole convergence: rms = 0.365E-6 iters = 4.00
> < NSTEP ENERGY RMS GMAX NAME
> NUMBER
> < 8 8.9665E-1 4.6266 2.1383E+1 N3 25
> < BOND = 0.4635 ANGLE = 4.9140 DIHED =
> 17.5299
> < VDWAALS = -3.6499 EEL = -24.6488 HBOND =
> 0.
> < 1-4 VDW = 7.1229 1-4 EEL = 5.9289 RESTRAINT =
> 0.
> < EPOLAR = -6.7640
> < Dipole convergence: rms = 0.348E-6 iters = 3.00
> < NSTEP ENERGY RMS GMAX NAME
> NUMBER
> < 9 8.0872E-1 1.2342 5.5340 EP23 26
> < BOND = 0.4542 ANGLE = 4.8390 DIHED =
> 17.5297
> < VDWAALS = -3.6511 EEL = -24.6601 HBOND =
> 0.
> < 1-4 VDW = 7.1201 1-4 EEL = 5.9415 RESTRAINT =
> 0.
> < EPOLAR = -6.7647
> < Dipole convergence: rms = 0.180E-6 iters = 4.00
> < NSTEP ENERGY RMS GMAX NAME
> NUMBER
> < 10 7.9896E-1 2.3141 1.2573E+1 N3 25
> < BOND = 0.4486 ANGLE = 4.8418 DIHED =
> 17.5287
> < VDWAALS = -3.6535 EEL = -24.6493 HBOND =
> 0.
> < 1-4 VDW = 7.1163 1-4 EEL = 5.9331 RESTRAINT =
> 0.
> < EPOLAR = -6.7666
> < Dipole convergence: rms = 0.436E-6 iters = 3.00
> ---
> > 3 NaN NaN 0. H5T 1
> > BOND = NaN ANGLE = NaN DIHED = NaN
> > VDWAALS = 0. EEL = 0. HBOND = 0.
> > 1-4 VDW = NaN 1-4 EEL = NaN RESTRAINT =
> 0.
> > EPOLAR = NaN
> > Dipole convergence: rms = NaN iters = 21.00
> > NSTEP ENERGY RMS GMAX NAME
> NUMBER
> > 4 NaN NaN 0. H5T 1
> > BOND = NaN ANGLE = NaN DIHED = NaN
> > VDWAALS = 0. EEL = 0. HBOND = 0.
> > 1-4 VDW = NaN 1-4 EEL = NaN RESTRAINT =
> 0.
> > EPOLAR = NaN
> > Dipole convergence: rms = NaN iters = 21.00
> > NSTEP ENERGY RMS GMAX NAME
> NUMBER
> > 5 NaN NaN 0. H5T 1
> > BOND = NaN ANGLE = NaN DIHED = NaN
> > VDWAALS = 0. EEL = 0. HBOND = 0.
> > 1-4 VDW = NaN 1-4 EEL = NaN RESTRAINT =
> 0.
> > EPOLAR = NaN
> > Dipole convergence: rms = NaN iters = 21.00
> > NSTEP ENERGY RMS GMAX NAME
> NUMBER
> > 6 NaN NaN 0. H5T 1
> > BOND = NaN ANGLE = NaN DIHED = NaN
> > VDWAALS = 0. EEL = 0. HBOND = 0.
> > 1-4 VDW = NaN 1-4 EEL = NaN RESTRAINT =
> 0.
> > EPOLAR = NaN
> > Dipole convergence: rms = NaN iters = 21.00
> > NSTEP ENERGY RMS GMAX NAME
> NUMBER
> > 7 NaN NaN 0. H5T 1
> > BOND = NaN ANGLE = NaN DIHED = NaN
> > VDWAALS = 0. EEL = 0. HBOND = 0.
> > 1-4 VDW = NaN 1-4 EEL = NaN RESTRAINT =
> 0.
> > EPOLAR = NaN
> > Dipole convergence: rms = NaN iters = 21.00
> > NSTEP ENERGY RMS GMAX NAME
> NUMBER
> > 8 NaN NaN 0. H5T 1
> > BOND = NaN ANGLE = NaN DIHED = NaN
> > VDWAALS = 0. EEL = 0. HBOND = 0.
> > 1-4 VDW = NaN 1-4 EEL = NaN RESTRAINT =
> 0.
> > EPOLAR = NaN
> > Dipole convergence: rms = NaN iters = 21.00
> > NSTEP ENERGY RMS GMAX NAME
> NUMBER
> > 9 NaN NaN 0. H5T 1
> > BOND = NaN ANGLE = NaN DIHED = NaN
> > VDWAALS = 0. EEL = 0. HBOND = 0.
> > 1-4 VDW = NaN 1-4 EEL = NaN RESTRAINT =
> 0.
> > EPOLAR = NaN
> > Dipole convergence: rms = NaN iters = 21.00
> > NSTEP ENERGY RMS GMAX NAME
> NUMBER
> > 10 NaN NaN 0. H5T 1
> > BOND = NaN ANGLE = NaN DIHED = NaN
> > VDWAALS = 0. EEL = 0. HBOND = 0.
> > 1-4 VDW = NaN 1-4 EEL = NaN RESTRAINT =
> 0.
> > EPOLAR = NaN
> > Dipole convergence: rms = NaN iters = 21.00
> 141,146c141,146
> < 10 7.9896E-1 2.3141 1.2573E+1 N3 25
> < BOND = 0.4486 ANGLE = 4.8418 DIHED =
> 17.5287
> < VDWAALS = -3.6535 EEL = -24.6493 HBOND =
> 0.
> < 1-4 VDW = 7.1163 1-4 EEL = 5.9331 RESTRAINT =
> 0.
> < EPOLAR = -6.7666
> < Dipole convergence: rms = 0.436E-6 iters = 3.00
> ---
> > 10 NaN NaN 0. H5T 1
> > BOND = NaN ANGLE = NaN DIHED = NaN
> > VDWAALS = 0. EEL = 0. HBOND = 0.
> > 1-4 VDW = NaN 1-4 EEL = NaN RESTRAINT =
> 0.
> > EPOLAR = NaN
> > Dipole convergence: rms = NaN iters = 21.00
> ---------------------------------------
>
> What could be the problem?
>
> Cheers,
>
> Anthony
> -----------------------------------------------------------------------
> The AMBER Mail Reflector
> To post, send mail to amber_at_scripps.edu
> To unsubscribe, send "unsubscribe amber" (in the *body* of the email)
> to majordomo_at_scripps.edu

-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" (in the *body* of the email)
      to majordomo_at_scripps.edu