Large numbers of bqs problem
Clicked A Few Times
6:05:54 PM PDT - Mon, Jun 24th 2013
Hello,

I am attempting to run a calculation with 8229 embedded point charges. I am aware that I cannot do this in the geometry block and so I am attempting to use the bq block. I encounter errors, however, whether I attempt to list the charges within the .nw file or use a separate bq_charges file. Listing the point charges in the file the calculation ends like this:

Quote:

------------------------------------------------------------------------

bq_input: too many centers        0

------------------------------------------------------------------------

------------------------------------------------------------------------

 current input line : 

 7016:     0.371548333333333    9.704969363722501   24.880367686617504    7.311721834400000

------------------------------------------------------------------------

------------------------------------------------------------------------

This error has not yet been assigned to a category

------------------------------------------------------------------------


And using the bq_charges file the calculation ends like this:

Quote:

Summary of "ao basis" -> "" (cartesian)

------------------------------------------------------------------------------

      Tag                 Description            Shells   Functions and Types

---------------- ------------------------------  ------  ---------------------

*                        aug-cc-pVDZ              all atoms except bq


MA_verify_allocator_stuff: starting scan ...
stack block 'qbq scratch', handle unknown, address 0xdd7e8580:
       current right signature 2837656253 != proper right signature 1431655765
stack block 'cbq scra <garbage characters> handle unknown, address 0xdd7f60a0:
       current checksum 18410048152472985750 != stored checksum 4600337559855963576
0:Segmentation Violation error, status=: 11
(rank:0 hostname:gellmann.scs.uiuc.edu pid:9218):ARMCI DASSERT fail. signaltrap.c:SigSegvHandler():301 cond:0
 0: ARMCI aborting 0 (0).


The text of the latter input file is as follows:

Quote:
start bq_file
title "bq_file"
set bq:max_nbq 10000
echo
memory 6400 mb
geometry units angstrom nocenter noautoz
symmetry c1

O   -1.005597595777500   -0.578717835982500   -1.279109106900000

H   -0.168275665377499   -0.146901774682500   -1.590554632600000

H   -1.032021380477500   -0.549630633582500   -0.287324590500000
end
basis 
* library aug-cc-pVDZ except bq
end
bq 
load bq_charges
end
scf
thresh 1.0e-10

tol2e 0.1e-10

maxiter 400

noprint "final vectors analysis" "mulliken"

singlet

rhf
end
task scf energy
geometry units angstrom nocenter noautoz
symmetry c1

O   -1.005597595777500   -0.578717835982500   -1.279109106900000

H   -0.168275665377499   -0.146901774682500   -1.590554632600000

H   -1.032021380477500   -0.549630633582500   -0.287324590500000
end
esp
recalculate
end
task esp


Does anyone know what the problem may be?

Forum Vet
1:39:00 PM PDT - Tue, Jun 25th 2013
I don't know which version you are running, but the latest NWChem 6.3 should be able to handle up to 25000 charges. Older versions may have been set up to handle fewer charges, but you and modify this if you want to. In src/bq/bq_input.F find the max_nbq variable and increase it, recompile and relink and you should be good to go.

Clicked A Few Times
3:55:51 PM PDT - Wed, Jun 26th 2013
I was using 6.0. This was the problem, thanks.


Forum >> NWChem's corner >> NWChem functionality