Topic Options
#8142 - 09/11/05 01:17 PM Error: MALLOC returned ZERO
chemer Offline
Forum Member

Registered: 04/01/05
Posts: 8
Loc: MO
I am using AFM module to pull a ligand out a protein with constant force,
I run two calculations with different pulling direction,
one run is ok, another stopped with the following error after 156800 steps. I compiled the programm with option xxlarge. My system has 51720 atoms. Anybody has experienced the similar problem?

DYNA> 156800 463.60000-111232.59691 45965.10895-157197.81714 298.15858

SELECTED IMAGES ATOMS BEING CENTERED ABOUT 0.000000 0.000000 0.000000
UPDECI: Nonbond update at step 156829
VEHEAP> Expanding heap size by50921472 words.
VEHEAP> Expanding heap size by75988992 words.
VEHEAP> Expanding heap size by******** words.
VEHEAP> Error: MALLOC returned ZERO.

***** LEVEL -4 WARNING FROM *****
***** COULD NOT SATISFY STORAGE REQUEST.
******************************************

Top
#8143 - 09/12/05 12:30 PM Re: Error: MALLOC returned ZERO [Re: chemer]
rmv Online   content

Forum Member

Registered: 09/17/03
Posts: 8373
Loc: 39 03 48 N, 77 06 54 W
You may wish to try XLARGE instead, as the memory allocated to static arrays for XXLARGE may be interfering with your need for dynamic memory for the calculation.

Another option is to count the number of VEHEAP> statements in the output, and see if you can reduce them via NBSCAL and IMSCAL (see nbonds.doc). Another way to limit VEHEAP> statements is to pre-allocate a large memory block via GETHEAP (test.doc). (Each VEHEAP> statement represents a memory allocation; the problem is that unused memory blocks from earlier allocations count against the program total.)

You could also try reducing CUTNB (keep CTOFNB and CTONNB the same), and increasing INBFRQ (if not -1; likewise for CUTIM/IMGFRQ).
_________________________
Rick Venable
computational chemist


Top

Moderator:  John Legato, lennart