Molcas Forum

PIjush Karak

Member

Registered: 2022-04-18

Posts: 16

memory issue in NAC calculation between T2 and T1 of a large molecule

Dear OpenMolcas Developer,

I want to compute the NAC between the T2 and the T1 states of a molecule with 105 atoms and  462 electrons.  The molecule consists of H,C,N and Au atoms.

I am using the pymolcas command for running the calculation.

After the GATEWAY part the calculation gets killed at the SEWARD part and computed only the 10% 2-electron integral. I am using 1 core and 1 node. I think it is a memory problem due to large molecular system. Here I attached the input and the SEWARD output section.

INPUT

&GATEWAY
  coord=tsdp.xyz
  basis = DEF2-SVP
  group = NOSYM

&SEWARD
&SCF

&RASSCF
  Nactel= 4 0 0
  Inactive= 229
  Ras2= 4
  SPIN=3
  CiRoot= 3 3 1
  Rlxroot=2

&GRID_IT
All

&ALASKA
 NAC = 2 1

output section of SEWARD

             Nuclear Potential Energy           7301.07091809 au

--- Stop Module: gateway at Sat Jun 21 12:54:20 2025 /rc=_RC_ALL_IS_WELL_ ---
*** files: xmldump
    saved to directory /scratch/swapanc.uc/1177722
--- Module gateway spent 1 second ---

--- Start Module: seward at Sat Jun 21 12:54:20 2025 ---

()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()

                                              &SEWARD

                                   only a single process is used
                       available to each process: 2.0 GB of memory, 1 thread?
                                            pid: 246839
()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()

               SEWARD will generate:
                  Multipole Moment integrals up to order  2
                  Kinetic Energy integrals
                  Nuclear Attraction integrals (point charge)
                  One-Electron Hamiltonian integrals modified with ECP contributions
                  Velocity integrals
                  Pseudo Potential integrals
                  Orbital angular momentum around ( -0.0701  -0.0011  -0.0000 )
                  Velocity quadrupole around ( -0.0701  -0.0011  -0.0000 )
                  Two-Electron Repulsion integrals

                   Integrals are discarded if absolute value <: 1.00E-14
                   Integral cutoff threshold is set to       <: 1.00E-16

            Nuclear Potential Energy           7301.07091809 au


      Basis set specifications :
      Symmetry species         a
      Basis functions         1047

 ###############################################################################
[ process      0]: xquit (rc =    161): _IO_ERROR_WRITE_
 ###############################################################################

Program aborted. Backtrace:
 ###                                                                         ###
 ###                                                                         ###
 ###    Location: AixWr                                                      ###
 ###    File: TEMP01                                                         ###
 ###                                                                         ###
 ###                                                                         ###
 ###    Premature abort while writing buffer to disk: Disk full?             ###
 ###                                                                         ###
 ###                                                                         ###
 ###############################################################################
 ###############################################################################
#0  0x636c6a in ???
#1  0x522a52 in ???
#2  0x6813ea in ???
#3  0x5ed452 in ???
#4  0x5ec965 in ???
#5  0x5ec806 in ???
#6  0x4930da in ???
#7  0x5eb504 in ???
#8  0x48ef0a in ???
#9  0x4839d0 in ???
#10  0x47ed91 in ???
#11  0x47bce2 in ???
#12  0x410358 in ???
#13  0x4056e2 in ???
#14  0x404654 in ???
#15  0x2b399a50a554 in ???
#16  0x404690 in ???
#17  0xffffffffffffffff in ???
--- Stop Module: seward at Sat Jun 21 13:18:31 2025 /rc=-6 ---
*** files: xmldump
    saved to directory /scratch/swapanc.uc/1177722
--- Module seward spent 24 minutes 10 seconds ---

.########################.
.# Non-zero return code #.
.########################.

    Timing: Wall=1451.48 User=1357.93 System=91.30

Can you suggest me the best possible core, node and memory setup to complete the calculation successfully ?

Looking forward to your reply.

with regards,
Pijush

Ignacio

Administrator

From: Uppsala

Registered: 2015-11-03

Posts: 1,124

Re: memory issue in NAC calculation between T2 and T1 of a large molecule

Premature abort while writing buffer to disk: Disk full?

Do you have enough disk space? Please use RICD in &GATEWAY for much improved performance.

PIjush Karak

Member

Registered: 2022-04-18

Posts: 16

Re: memory issue in NAC calculation between T2 and T1 of a large molecule

Dear Ignacio,

Thank you for your mail and help.

Yes, I will use the RICD in &GATEWAY for much better performance as you mentioned. 

Can you give some idea of how about the disk space required for this ?

with regards,
Pijush

Ignacio

Administrator

From: Uppsala

Registered: 2015-11-03

Posts: 1,124

Re: memory issue in NAC calculation between T2 and T1 of a large molecule

With RICD it will be much less.