[Octopus-users] inherent memory limits?

[Emily]

Hello,

I'm trying to run a simulation that takes a lot of memory, on a computer
that has even more memory, and it's failing. 

Specifically, I'm getting an error:

**************************** FATAL ERROR *****************************
*** Fatal Error (description follows)
*--------------------------------------------------------------------
* Failed to allocate     122581 Kb in file 'td_rti.F90' line   287
*--------------------------------------------------------------------
* Stack:
**********************************************************************

But the node I'm running on has 32 Gb of memory, and I get this error
after the code has buffered just over 6 Gb of memory, and used 2.25 Gb,
at the beginning of the time dependent calculation.

I'm wondering if there is some inherent limit in the code that would
cause this error even if I have plenty of memory available.  

My pbs submission script contains the lines:

#PBS -l nodes=1:emt64,pmem=32000mb,vmem=32000mb -W
x=NACCESSPOLICY:SINGLEJOB
#PBS -q hugemem
#PBS -j oe
limit stacksize unlimited

which I have reason to believe (but not certainty) are supposed to allow
me to use the 32 Gb.

Can anybody suggest reasons I can't access more memory?

Oh, and:
 Running octopus, version 2.1.0
              build time - Tue Dec  4 14:31:54 EST 2007
                            svn revision:

                           Compiler: ifort
Compiler flags: -u -zero -fpp1 -nbs -pc80 -pad -align -unroll -O3 -ip
-tpp7 -xW

Thanks,
Emily