Hello -
 
This is an issue that we've seen with the 64-bit machines - the issue is in regard to MALOC's BLAS libraries.  Some discussion can be found here:
 
http://cholla.wustl.edu/pipermail/apbs-users/2005-April/000410.html
 
And installation instructions (for amd64) can be found here:
 
http://apbs.sourceforge.net/doc/user-guide/x146.html#AEN396

Although you said you were using APBS 0.3.2, this problem will still persist in APBS 0.4.0 as we're using the same version of MALOC;  that being said, the Opteron binaries (x86_64 RPMs) were compiled with the PGI BLAS libraries and thus don't have this problem.
 
Todd
 
On 12/28/05, <glc02@mails.tsinghua.edu.cn> wrote:
Dear APBS users:

    I currently installed APBS-0.3.2 on an AMD64 machine with source code. It seems no problem during installation. However, when I tried the Born ion example, the programme broke down with segmentation fault warning. I have searched the mailing list, there is a mail about this question and without answer. Could anyone help me? thanks a lot!:)
    Some messages in stdout and io.mc are listed below

    stdout:
    ........................
    Preparing to run 2 PBE calculations.
----------------------------------------
CALCULATION #1 (solv): MULTIGRID
Setting up problem...
Vpbe_ctor:  Using max ion radius (0 A) for exclusion function
Debye length:  0 A
Current memory usage:  153.422 MB total, 153.422 MB high water
Using cubic spline charge discretization.
Grid dimensions: 97 x 97 x 97
Grid spacings: 0.330 x 0.330 x 0.330
Grid lengths: 31.680 x 31.680 x 31.680
Grid center: (0.000, 0.000, 0.000)
Multigrid levels: 4
Molecule ID: 1
Linearized traditional PBE
Multiple Debye-Huckel sphere boundary conditions
0 ion species (0.000 M ionic strength):
Solute dielectric: 1.000
Solvent dielectric: 78.540
Using "molecular" surface definition; no smoothing
Solvent probe radius: 1.400 A
Temperature:  298.150 K
Surface tension:  0.105 kJ/mol/A^2
Electrostatic energies will be calculated
Potential to be written to potential-8.00.dx
Solving PDE (see io.mc* for details)...
Segmentation fault
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
   io.mc:

...........
Setting up PBE object...
Vpbe_ctor2: Started constructing Vacc object with 32 x 32 x 32 hash table
Vacc_ctor2:  Using 32 x 32 x 32 hash table
Vacc_ctor2:  Using 1.9 max radius
Vacc_ctor2:  Using 200-point probe sphere
Vacc_ctor2:  Grid lengths = (28.116, 28.116, 28.116)
Vacc_ctor2:  Grid lower corner = (- 14.058, -14.058, -14.058)
Vacc_ctor2:  Have 13824 atom entries
Vpbe_ctor2: Done constructing Vacc object...
Setting up PDE object...
Vpmp_ctor2:  Using meth = 0, mgcoar = 2, mgsolv = 0
Setting PDE center to local center...
Vpmg_ctor2: PMG chose nx = 97, ny = 97, nz = 97, nlev = 4
Vpmg_fillco:  filling in source term.
fillcoCharge:  Calling fillcoChargeSpline2...
Vpmg_fillco:  filling in source term.
Vpmg_fillco:  marking ion and solvent accessibility.
fillcoCoef:  Calling fillcoCoefMol...
Vpmg_fillco:  filling coefficient arrays
Vpmg_fillco:  done filling coefficient arrays
Vpmg_fillco:  filling boundary arrays
Vpmg_fillco:  done filling boundary arrays
Vnm_tstop: stopping timer 27 (Setup timer).  CPU TIME = 5.600000e-01
Vnm_tstart: starting timer 28 (Solver timer)..
Vnm_tstart: starting timer 30 (MGDRIV2: fine pro)..
Vnm_tstop: stopping timer 30 (MGDRIV2: fine pro).  CPU TIME = 1.500000e-01
Vnm_tstart: starting timer 30 (MGDRIV2: coarse pro)..
Vnm_tstop: stopping timer 30 (MGDRIV2: coarse pro).  CPU TIME = 3.400000e-01
Vnm_tstart: starting timer 30 (MGDRIV2: solve)..
Vnm_tstop: stopping timer 40 (MG iteration).  CPU TIME = 1.060000e+00
PMG: iteration =  0
PMG: relative residual =  1.000000e+00
PMG: contraction number =  1.000000e+00







-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
_______________________________________________
apbs-users mailing list
apbs-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/apbs-users