pressure profile calculations

From: upayan baul (margheritan22_at_gmail.com)
Date: Sat Nov 17 2012 - 04:55:59 CST

Hi,

I have been trying to compute pressure profile for a mixed bilayer
(POPE+POPG) with 128 lipids per layer.

After equilibrating for 15ns I computed 'kinetic, bonded and nonbonded'
contributions over a 5ns run and subsequently tried to compute the
'reciprocal space' contribution from the dcd output for the same- it had
5000 frames. I am attaching the configuration files for both.

The run for reciprocal space contribution did not run for parallel runs and
gave error message:

Info: Coordinate file ./pressure-dir-15-20ns.dcd opened for reading.
Info: Reading timestep from file.
Info: Updating unit cell from timestep.
Info: Reading from binary file PO-PE_PG-run3.restart.vel
Info: REMOVING COM VELOCITY -9.27188e-08 9.60535e-09 4.82838e-08
TCL: Setting parameter firstTimestep to 2500
TCL: Original numsteps 0 will be ignored.
TCL: Running for 0 steps
FATAL ERROR: ComputeMgr::computeEwaldObject in recvData is NULL!
[41] Stack Traceback:
  [41:0] _Z8NAMD_diePKc+0x72 [0x5438d2]
  [41:1]
_ZN18CkIndex_ComputeMgr42_call_recvComputeEwaldData_ComputeEwaldMsgEPvP10ComputeMgr+0x24
 [0x5b9cd4]
  [41:2] _Z15_processHandlerPvP11CkCoreState+0x59f [0xb0c3cf]
  [41:3] CsdScheduler+0x41a [0xc0b82a]
  [41:4] _ZN7BackEnd4initEiPPc+0x30e [0x54e34e]
  [41:5] main+0x5f [0x54792f]
  [41:6] __libc_start_main+0xf4 [0x2b0648c41184]
  [41:7] _ZNSt8ios_base4InitD1Ev+0x49 [0x505939]

The error did not appear in a serial run but the results obtained did not
seem aggreable.

Could you please advise me on circumventing the problem with the parallel
computation? Also, could you please let me know if computing the reciprocal
lattice contribution over 5000 frames would be enough for a mixed bilayer
having 128 lipids per layer?

Regards,
Upayan


This archive was generated by hypermail 2.1.6 : Tue Dec 31 2013 - 23:22:45 CST