Linker IT Software
Google
Web www.oraxcel.com
menubar-top-links menubar-top-rechts
Home Help Search Login
Welcome, Guest. Please Login.
SQL*XL: Database to Excel bridge litLIB: Excel power functions pack ExcelLock: Locking and securing your valuable Excel spreadsheets encOffice: Protect your Excel file easy and safe encOffice: Protect your Excel file easy and safe
Pages: 1 2 
MOLCAS problems & bugs (Read 18956 times)
Gerrit-Jan Linker
YaBB Administrator
*****




Posts: 75
MOLCAS problems & bugs
23.09.08 at 20:46:11
 
MOLCAS bugs
 
In this topic I would like to collect some problems I encounter with MOLCAS. Perhaps you encouter the same problems and can it be of help to you.
 
Back to top
 
« Last Edit: 19.10.08 at 10:47:03 by Gerrit-Jan Linker »  

Gerrit-Jan Linker
Linker IT Software
Email WWW Gerrit-Jan Linker   IP Logged
Gerrit-Jan Linker
YaBB Administrator
*****




Posts: 75
SCF: Segmentation dump
Reply #1 - 23.09.08 at 20:47:50
 
SCF: Segmentation dump
Quote:

--- Stop Module: scf at Mon Sep 22 22:02:46 2008 /rc=11 ---
*** file /scratch/project.5479/project.ScfOrb saved as /home/linker/project/c0gf/project.ScfOrb.autosaved
Non-zero return code - check program input/output

    Segmentation dump

 
Remarks:

  • Assigning less memory can fix the problem.
    I have seen that this error occurred when allocating 2047 MB of memory (maximum for a 32bits MOLCAS) on a 4GB machine but that it did work ok with 750 MB of memory on a 1GB machine.
  • It was suggested that segmentation dump errors may be due to specifying the wrong compiler optimizer options.
Back to top
 
« Last Edit: 28.09.08 at 10:28:28 by Gerrit-Jan Linker »  

Gerrit-Jan Linker
Linker IT Software
Email WWW Gerrit-Jan Linker   IP Logged
Gerrit-Jan Linker
YaBB Administrator
*****




Posts: 75
SCF: rc=11
Reply #2 - 23.09.08 at 20:54:58
 
SCF: rc=11
 
Quote:
Detected guessorb starting orbitals


--- Stop Module: scf at Tue Sep 23 08:27:46 2008 /rc=11 ---
*** file /scratch/edottf.24916/edottf.ScfOrb saved as /home/linker/edottf/c0go.a/edottf.ScfOrb.autosaved
Non-zero return code - check program input/output
"edottf.output" 459L, 24087C
       
 
Workaround for this problem is to use the RASSCF program and to make the input such that only the SCF solution is found. For example if you have 20 electrons you can set INACT to 19, put the RAS2 space to 1 and allow 2 electrons to become active. In such a calculation only 1 CSF can be made and that CSF is the HF solution.
 
This problem is due to memory constraints.
Give it more memory and it runs ok.  
I received this error running SCF with 750 MB of memory on a 1GB machine. Running it with 2047MB on a 4GB machine works fine.
 
NODIIS
Keyword NODIIS seems to require more memory than using the DIIS convergence excelerator. Without NODIIS my job runs fine but with NODIIS it crashes with this message.
Back to top
 
« Last Edit: 20.11.08 at 07:41:53 by Gerrit-Jan Linker »  

Gerrit-Jan Linker
Linker IT Software
Email WWW Gerrit-Jan Linker   IP Logged
Gerrit-Jan Linker
YaBB Administrator
*****




Posts: 75
RASSCF: Non-zero rc in Cho_X_init
Reply #3 - 05.10.08 at 10:46:21
 
RASSCF: Non-zero rc in Cho_X_init
 
Quote:
Make Quasi-Newton update
RASSCF. Non-zero rc in Cho_X_init.
--- Stop Module: rasscf at Sun Oct 5 00:01:48 CEST 2008 /rc= -2 ---
Non-zero return code - check program input/output
--- Stop Module: auto at Sun Oct 5 00:01:48 CEST 2008 /rc=254 ---

 
Problem occurred after a successful SEWARD + RASSCF. I copied all files and invoked molcas again copying in all these files. I did only the RASSCF step again resulting in the error.
Back to top
 
« Last Edit: 09.10.08 at 08:20:57 by Gerrit-Jan Linker »  

Gerrit-Jan Linker
Linker IT Software
Email WWW Gerrit-Jan Linker   IP Logged
Gerrit-Jan Linker
YaBB Administrator
*****




Posts: 75
NBAS does not match
Reply #4 - 09.10.08 at 08:22:12
 
LoProp: NBAS does not match
 
Quote:
 No dynamic properties will be computed.

 Expansion centers of the domains are for the
  atomic domains: the atom center
  bond domains  : the center of the bond

    Location: rdVec_  
    File: INPORB  

   NBAS does not match  

--- Stop Module: loprop at Thu Oct  9 01:22:36 2008 /rc= 97 ---

    Internal error or incompatibility in MOLCAS

 
This error occurred when using the 6-31G* basisset.
The same input file changed to use the STO-3G basisset did work correctly!
Back to top
 
« Last Edit: 09.10.08 at 10:52:04 by Gerrit-Jan Linker »  

Gerrit-Jan Linker
Linker IT Software
Email WWW Gerrit-Jan Linker   IP Logged
Gerrit-Jan Linker
YaBB Administrator
*****




Posts: 75
Re: MOLCAS bugs
Reply #5 - 09.10.08 at 22:53:26
 
RASSCF: Premature abort while reading buffer from disk
 
Quote:
 I/O statistics
 --------------

 Unit  Name    Xtent pages MBytes     pages MBytes
  (MBytes)     in    in  out  out
 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
 11   RUNFILE   0.03 0.01 0.03 0.00 0.00
 13   TRAINT    0.00 0.00 0.00 0.00 0.00
 15   JOBIPH    4.51 0.00 0.00 0.03 0.48
 27   TEMP02    0.00 0.00 0.00 0.00 0.00
 37   TEMP01    0.00 0.00 0.00 0.00 0.00
 40   ORDINT 1499.77 5.86    1499.77 0.00 0.00
 77   ONEINT    0.63 0.03 1.04 0.00 0.00
 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
 Total   5.90    1500.84 0.03 0.48
 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

    Location: AixRd    
    File: ORDINT  
 
    Premature abort while reading buffer from disk:  
End of file reached
 
 
--- Stop Module: rasscf at Thu Oct  9 22:33:32 2008 /rc= 132 ---

  Execution terminated due to problems with reading data from the disk

Non-zero return code - check program input/output

 
This problem arose when copying the job files while a RASSCF was in progress. Is it not possible to copy integral files when they are read at the same time?
 
Edited:
Perhaps this copying of integral files was not the problem as I had precisely the same error when experiencing a full disk. I think this error is due to lack of free disk space. The disk is full.

 
Back to top
 
« Last Edit: 02.04.09 at 11:38:15 by Gerrit-Jan Linker »  

Gerrit-Jan Linker
Linker IT Software
Email WWW Gerrit-Jan Linker   IP Logged
Gerrit-Jan Linker
YaBB Administrator
*****




Posts: 75
RASSCF: MA failed to release the memory block
Reply #6 - 09.10.08 at 23:14:50
 
RASSCF: MA failed to release the memory block for further use
 
Quote:
 Natural spin orbitals have been written to file  SPDORB1 for root number 1
MA failed to release the memory block for further use.

 Memory allocation map
<<snip>>

The parameters of the last call to GetMem were:
Name  =FI
FldNam=SMAT
VarTyp=REAL
Key   =FREE
Length= 28441
iPos  =    10732067
--- Stop Module: rasscf at Thu Oct  9 23:07:44 2008 /rc= 143 ---

 
This error occurred after doing 200 iterations and not obtaining convergence. 750MB of memory was allowed. Is most memory needed by RASSCF at the end of the run?
 
The same error occurred allowing 2047MB of memory usage.
 
Back to top
 
« Last Edit: 13.10.08 at 13:40:26 by Gerrit-Jan Linker »  

Gerrit-Jan Linker
Linker IT Software
Email WWW Gerrit-Jan Linker   IP Logged
Gerrit-Jan Linker
YaBB Administrator
*****




Posts: 75
Alaska: rc= 65
Reply #7 - 13.10.08 at 14:58:16
 
Alaska: rc= 65
 
Quote:
 Alaska requests the Numerical_Gradient module to be executed!

--- Stop Module: alaska at Fri Oct 10 00:21:45 2008 /rc= 65 ---


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

   MOLCAS executing module NUMERICAL_GRADIENT with 750 MB of memory    
   at 00:21:46 Fri Oct 10 2008    
()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()




 Original coordinates    
 mat. size =     3x   20
  10.01554846 11.90527459 12.66116504  5.48020576  8.12582234  8.50376756 11.90527459 11.14938414  6.61404144  5.66917837
   9.07068540  8.69274017 10.77143891  9.25965801 11.52732936  8.50376756  6.80301405  8.69274017 12.47219242  7.36993189
   9.63760324 -2.45664396 -5.48020576 10.01554846 -5.29123315 -5.66917837  1.32280829  6.04712360  1.70075351  6.61404144
   2.83458919  4.91328792 -0.75589045  8.50376756 -3.77945225 -0.56691784  8.69274017 -4.34637009 -2.64561657 -2.07869874
  -6.80301405  6.42506882  5.10226054 -5.29123315  6.23609621  3.02356180 -0.37794522 -4.15739747  1.32280829 -2.26767135
  -0.56691784 -2.07869874  1.70075351 -5.48020576  4.72431531  2.45664396 -4.53534270  4.53534270  2.45664396  4.34637009

Root to use:  1
Original CASSCF   Energy=-2046.4497184241    

Effective number of dispacements are  120

Here we go ...

 
It is not very clear whether Alaska has actually failed or not. It gives return code 65 but MOLCAS seems to continue regardless. What is odd is that every iteraction gives very minimal output now:
 
 
Quote:
()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()

   MOLCAS executing module SEWARD with 750 MB of memory  
   at 07:29:46 Sun Oct 12 2008    
()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()





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

   MOLCAS executing module RASSCF with 750 MB of memory    
   at 12:35:11 Sun Oct 12 2008    
()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()



*** Detected Cholesky or RI/DF calculation: THRE reset to Decom. Thr. =
 1.E-06

The MO-coefficients are taken from the file INPORB
title: RASSCF average (pseudo-natural) orbitals  
Total energies have been shifted with     -2000.00 au

WARNING: large rotation in molecular orbital  62 of symmetry 1 MO space 2  weight is    0.393629
WARNING: large rotation in molecular orbital  70 of symmetry 1 MO space 2  weight is    0.344704
WARNING: large rotation in molecular orbital  71 of symmetry 1 MO space 2  weight is    0.336654
Convergence after 67 iterations
root number  1 E-EVAC = -2046.45003650 a.u.
 Natural spin orbitals have been written to file  SPDORB1 for root number 1




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

   MOLCAS executing module SEWARD with 750 MB of memory    
   at 11:35:54 Mon Oct 13 2008    
()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()

Back to top
 
« Last Edit: 13.10.08 at 15:00:41 by Gerrit-Jan Linker »  

Gerrit-Jan Linker
Linker IT Software
Email WWW Gerrit-Jan Linker   IP Logged
Gerrit-Jan Linker
YaBB Administrator
*****




Posts: 75
Seward:  Disk Full? rc=131
Reply #8 - 19.10.08 at 10:52:58
 
Seward:  Disk Full? rc=131
 
Quote:

***    Location: AixWr ***
***    File: /scratch/edottf.24843/edottf.OrdInt;   ***
*** ***
*** ***
***    Premature abort while writing buffer to disk: Disk full? ***
*** ***
*** ***
*******************************************************************************
*******************************************************************************
--- Stop Module: seward at Sat Oct 18 09:40:52 2008 /rc= 131 ---

  Execution terminated due to problems with writting to the disk

Non-zero return code - check program input/output

 
Cholesky could be used to decrease the size of the integral files.
Back to top
 
« Last Edit: 19.10.08 at 10:54:57 by Gerrit-Jan Linker »  

Gerrit-Jan Linker
Linker IT Software
Email WWW Gerrit-Jan Linker   IP Logged
Gerrit-Jan Linker
YaBB Administrator
*****




Posts: 75
SCF: CHOSCF_DRV. Non-zero return code
Reply #9 - 01.11.08 at 09:43:20
 
SCF: CHOSCF_DRV. Non-zero return code
 
Quote:
Detected guessorb starting orbitals


SCF iterations: Energy and convergence statistics

Iter Tot. SCF One-electron Two-electron Energy Max Dij or Max Fij DNorm TNorm AccCon Time
Energy Energy Energy Change Delta Norm
in Sec.
1 -2024.07330406 -5475.16095460 2100.45949823 0.00E+00 0.23E+00* 0.31E+02* 0.51E+02 0.84E+03 NoneDa 1939.
CHOSCF_DRV. Non-zero return code.
real 1999.36
user 1952.74
sys 15.22
--- Stop Module: scf at Fri Oct 31 23:48:51 2008 /rc= 101 (Unknown) ---
--- Stop Module: scf at Fri Oct 31 23:48:51 2008 /rc= 101 (Unknown) ---
Back to top
 
 

Gerrit-Jan Linker
Linker IT Software
Email WWW Gerrit-Jan Linker   IP Logged
Gerrit-Jan Linker
YaBB Administrator
*****




Posts: 75
RASSCF: Command terminated by signal 6
Reply #10 - 03.11.08 at 14:24:49
 
RASSCF: Command terminated by signal 6
 
Quote:
Iter CI   SX   CI RASSCF Energy    max BLB   max BLB  max ROT   Level Ln srch  Step   QN    Time(min)
   iter iter root energy change    element    value    param    shift minimum  type update     CPU
Command terminated by signal 6
real 2292.41
user 3493.62
sys 356.68
--- Stop Module: rasscf at Sun Nov  2 23:00:06 2008 /rc=6 (Unknown) ---
--- Stop Module: rasscf at Sun Nov  2 23:00:06 2008 /rc=6 (Unknown) ---

Back to top
 
« Last Edit: 03.11.08 at 19:47:07 by Gerrit-Jan Linker »  

Gerrit-Jan Linker
Linker IT Software
Email WWW Gerrit-Jan Linker   IP Logged
Gerrit-Jan Linker
YaBB Administrator
*****




Posts: 75
SCF: severe (174): SIGSEGV, segmentation fault
Reply #11 - 23.11.08 at 11:34:12
 
SCF: severe (174): SIGSEGV, segmentation fault

 
This severe error happened during a MOLCAS SCF run with NODIIS specified. First this crashed with a memory error rc=11. Giving it more memory (20GB instead of 750MB) it crashed with this error.
 
Quote:
Orbital file label: * UHF orbitals + arbitrary occupations


UHF SCF  iterations: Energy and convergence statistics

Iter     Tot. SCF One-electron     Two-electron   Energy   Max Dij or   Max Fij   DNorm TNorm     AccCon    Time
    Energy    Energy    Energy Change   Delta Norm
 in Sec.
  1  -2046.18084200  -5502.73718487   2104.05743798  0.00E+00   0.80E-02*  0.67E-02*   0.18E+02   0.84E+03   NoneDa  373.
  2  -2046.18159210  -5502.91083403   2104.23033705 -0.75E-03*  0.55E-02*  0.68E-02*   0.19E+02   0.84E+03   Damp    670.
...<<snip>>...
 16  -2046.18239218  -5502.91295857   2104.23166150 -0.87E-06   0.39E-03*  0.65E-04    0.18E+02   0.84E+03   Damp    731.
forrtl: severe (174): SIGSEGV, segmentation fault occurred
Image  PC    Routine Line  Source
libmkl_mc.so 00002AAAAC096FBA  Unknown   Unknown  Unknown
Command exited with non-zero status 174
real 11322.38
user 11185.75
sys 123.46
--- Stop Module: scf at Sat Nov 22 13:54:13 2008 /rc=174 (Unknown) ---
--- Stop Module: scf at Sat Nov 22 13:54:13 2008 /rc=174 (Unknown) ---
Back to top
 
« Last Edit: 23.11.08 at 11:42:17 by Gerrit-Jan Linker »  

Gerrit-Jan Linker
Linker IT Software
Email WWW Gerrit-Jan Linker   IP Logged
Gerrit-Jan Linker
YaBB Administrator
*****




Posts: 75
RASSCF: MA error: MA_free_heap
Reply #12 - 29.11.08 at 20:55:29
 
RASSCF: MA error: MA_free_heap
 
 
Quote:
 Natural spin orbitals have been written to file  SPDORB1 for root number 1

Failed to generate input file to MOLDEN
No mixing allowed of spherical and cartesian d, f, g-functions
MA error: MA_free_heap: memhandle 122 (name: 'INFRED') not in heap
MA failed to release the memory block for further use.

 Memory allocation map

 
RASSCF failed at the end of the run after printing the Dipole moments.
Back to top
 
« Last Edit: 29.11.08 at 20:56:35 by Gerrit-Jan Linker »  

Gerrit-Jan Linker
Linker IT Software
Email WWW Gerrit-Jan Linker   IP Logged
Gerrit-Jan Linker
YaBB Administrator
*****




Posts: 75
RASSCF: Not yet defined error reading INPORB
Reply #13 - 30.11.08 at 12:23:45
 
RASSCF: Not yet defined error reading INPORB
 
Quote:
Make Quasi-Newton update

Location: rdVec_
File: INPORB

Error during reading INPORB
not defined yet    

real 1.78
user 0.14
sys 0.07
--- Stop Module: rasscf at Sat Nov 29 22:24:54 2008 /rc= 97 ---
Back to top
 
« Last Edit: 30.11.08 at 12:26:35 by Gerrit-Jan Linker »  

Gerrit-Jan Linker
Linker IT Software
Email WWW Gerrit-Jan Linker   IP Logged
Gerrit-Jan Linker
YaBB Administrator
*****




Posts: 75
MCLR: Not enough CORE space
Reply #14 - 27.03.09 at 20:46:42
 
MCLR: Not enough CORE space
 
Quote:

NOT ENOUGH CORE SPACE FOR SORTING IN TRA2
TOTAL SORTING SPACE IS 653519473
STEP1: AVAILABLE IS 339874920 NEEDED IS 224785
STEP2: " 339874920 " 224785
STEP3: " 0 " 2994163
STEP4: " 339874920 " 446890
real 0.45
user 0.13
sys 0.03
--- Stop Module: mclr at Thu Mar 26 16:34:44 2009 /rc= _INTERNAL_ERROR_ ---
--- Stop Module: mclr at Thu Mar 26 16:34:44 2009 /rc= _INTERNAL_ERROR_ ---
Non-zero return code - check program input/output

 
MCLR was started with 5000MB of memory and unlimited disk (practical limit 1TB).
 
A memory problem is assumed.
Back to top
 
 

Gerrit-Jan Linker
Linker IT Software
Email WWW Gerrit-Jan Linker   IP Logged
Pages: 1 2