[gmx-users] NMA error

[Fabrizio Mancinelli]

Hello all,
excuse me if I re-post the same question, but I didn't got any answer so maybe 
the post has been overlooked.
In performing a NMA in vacuum of a large (~600 residue) protein
(that means:
ns_type: simple
pbc=no
rlist=0
coulombtype=cut-off
vdwtype=cut-off
rcoulomb=o
rvdw=0)
I got the following error while grompping:

Non-cutoff electrostatics used, forcing full Hessian format.

and,  of course, even with 2 Gb memory, the simulation won't start:

-------------------------------------------------------
Program mdrun_d, VERSION 3.3
Source code file: smalloc.c, line: 113

Fatal error:
calloc for full_matrix (nelem=280629504, elsize=8, file minimize.c, line 1866)
-------------------------------------------------------

"I'm Your Worst Nightmare" (Creep)
: Cannot allocate memory

I'm wondering, is there any mean to overcome the problem (but trivially 
expanding memory or NMAing on a subset, let's say CA's)? Does not the 3.3 
version use sparse matrix representation?
Maybe I should use cut-off's, but does this work in vacuum?
Thanks again for attention.