Very high time effort for writing .op2 file

Jens_22368

Hi!

 

I try to calculate the resulting spcforces for a given displacement field on a hugh model with about 2.000.000 nodes. The displacements are applied to the model using spcs for each node and each component. This usually should be easy to calculate since the stiffness matrix does not even have to be inverted. The problem is that after the solver is finished, writing the .op2 file tooks a big amount of time... 

Time spent in section: SOLVER;  CPU=    3916.96;  WALL=    704.95

Time spent in section :     OP2;  CPU=   39045.11;  WALL=  39026.59

 

Is this a common problem? For which reasons could writing the .op2 file that slow?

 

 

In the following you can find the .stat file of the job:

 

 

 Program execution started at Wed Jun 19 12:16:49 EDT 2019
 Command line arguments:
 mono_wo_RBE3_and_2_only_one_subcase.fem
 -ram=92000    same as: -len
 -nproc=8
  
  
 Options read from config files:
  DOS_DRIVE_R = /.
  env(OS_RAM)        6000

TIME SPENT IN MODULE :    READ;  CPU=     133.56;  WALL=    133.51

START RUNNING MODULE :  PREPR1 ...

TIME SPENT IN MODULE :  PREPR1;  CPU=       4.13;  WALL=      4.18

START RUNNING MODULE :  DESIGN ...

TIME SPENT IN MODULE :  DESIGN;  CPU=       6.48;  WALL=      6.47

START RUNNING MODULE :  PREPRO ...

   Start running section :  PRPDDM ...

   Time spent in section :  PRPDDM;  CPU=       0.01;  WALL=      0.01

 
***** datamanager call insert irlist for array size 542810504 using 32-bit
 
***** datamanager call insert scr4_ for array size 516096848 using 32-bit

   Start running section :  PRPSL4 ...

   Time spent in section :  PRPSL4;  CPU=     167.88;  WALL=    167.87

Request for memory resize from       9200 to      54973 MB
Resize ok with new memory.

TIME SPENT IN MODULE :  PREPRO;  CPU=     314.15;  WALL=    310.87

ITERATION   0

START RUNNING MODULE : CHECKEL ...

TIME SPENT IN MODULE : CHECKEL;  CPU=       6.01;  WALL=      3.71

START RUNNING MODULE :  SOLVER ...

   Start running section :  PRASMB ...

   Time spent in section :  PRASMB;  CPU=       0.08;  WALL=      0.08

   Start running section :  SPASMB ...
 
***** datamanager call insert force for array size 2147483640 using 32-bit

   Time spent in section :  SPASMB;  CPU=     127.78;  WALL=    127.71

   Solving subcase         11

   Solving subcase         11

   Solving subcase         11

   Start running section :  PREFCT ...

   Time spent in section :  PREFCT;  CPU=       0.32;  WALL=      0.32

   Start running section :  INPTVM ...

   Time spent in section :  INPTVM;  CPU=       8.34;  WALL=      8.34

   Start running section :  FCTMTX ...

   Time spent in section :  FCTMTX;  CPU=    3660.07;  WALL=    520.04

   Start running section :  SOLRHS ...

   Time spent in section :  SOLRHS;  CPU=      81.98;  WALL=     11.57

   Start running section :  SAVFCT ...

   Time spent in section :  SAVFCT;  CPU=      32.89;  WALL=     31.11

   Start running section :  STRENR ...

   Time spent in section :  STRENR;  CPU=       4.14;  WALL=      3.73

   Start running section :  KINENR ...

   Time spent in section :  KINENR;  CPU=       0.00;  WALL=      0.00

   Start running section :  OBJFUN ...

   Time spent in section :  OBJFUN;  CPU=       0.00;  WALL=      0.00

TIME SPENT IN MODULE :  SOLVER;  CPU=    3916.96;  WALL=    704.95

START RUNNING MODULE :  STRESS ...

TIME SPENT IN MODULE :  STRESS;  CPU=     119.78;  WALL=    118.96

START RUNNING MODULE :  RESP 1 ...
 
***** datamanager call insert scratch1 for array size 2147483620 using 32-bit

TIME SPENT IN MODULE :  RESP 1;  CPU=       0.03;  WALL=      0.03

START RUNNING MODULE :  SCREEN ...

TIME SPENT IN MODULE :  SCREEN;  CPU=       0.02;  WALL=      0.02

START RUNNING MODULE :CONVERGE ...

TIME SPENT IN MODULE :CONVERGE;  CPU=       0.06;  WALL=      0.06

START RUNNING MODULE :   PRINT ...

TIME SPENT IN MODULE :   PRINT;  CPU=       0.01;  WALL=      0.01

START RUNNING MODULE : RESULTS ...
 
***** datamanager call insert sortres for array size 2147483640 using 32-bit

   Start running section :     OP2 ...
CASECC data block written to the .op2 file for     1 SUBCASEs.  Length =    14820 words.
GEOM1S data block written to the .op2 file.
GEOM2S data block written to the .op2 file.
GEOM3S data block written to the .op2 file.
GEOM4S data block written to the .op2 file.
EPTS data block written to the .op2 file.
MPTS data block written to the .op2 file.
OUGV1 data block written to the .op2 file for    2108395 grids.
OES1X data block written to the .op2 file for    1954188 elements.

   Time spent in section :     OP2;  CPU=   39045.11;  WALL=  39026.59

   Start running section :    MISC ...

   Time spent in section :    MISC;  CPU=       0.64;  WALL=      1.29

 
 LIMIT ON ITERATIONS REACHED.
  
 INFEASIBLE DESIGN (AT LEAST ONE CONSTRAINT VIOLATED).
  

TIME SPENT IN MODULE : RESULTS;  CPU=   39045.76;  WALL=  39027.88

START RUNNING MODULE :FINISHUP ...

TIME SPENT IN MODULE :FINISHUP;  CPU=       4.36;  WALL=      4.37

 Detailed disk usage: 
    High usage overall   :96709.6 MiB,
    Solver scratch       :41605.0 MiB,
    Internal database    : 4109.0 MiB,
    Out-of-core assembly :    0.0 MiB,
    Temporary files      :51183.6 MiB.
AVERAGE SYSTEM LOAD                                      5.14                   

SOLUTION COMPLETED, THE FOLLOWING IS CPU/WALL CLOCK SUMMARY (IN SECONDS):

TOTAL TIME SPENT IN MODULE :READ    ;  CPU=     133.56;  WALL=    133.51
TOTAL TIME SPENT IN MODULE :PREPR1  ;  CPU=       4.13;  WALL=      4.18
TOTAL TIME SPENT IN MODULE :DESIGN  ;  CPU=       6.48;  WALL=      6.47
   Total time spent in section :PRPDDM  ;  CPU=       0.01;  WALL=      0.01
   Total time spent in section :PRPSL4  ;  CPU=     167.88;  WALL=    167.87
TOTAL TIME SPENT IN MODULE :PREPRO  ;  CPU=     314.15;  WALL=    310.87
TOTAL TIME SPENT IN MODULE :CHECKEL ;  CPU=       6.01;  WALL=      3.71
   Total time spent in section :PRASMB  ;  CPU=       0.08;  WALL=      0.08
   Total time spent in section :SPASMB  ;  CPU=     127.78;  WALL=    127.71
   Total time spent in section :PREFCT  ;  CPU=       0.32;  WALL=      0.32
   Total time spent in section :INPTVM  ;  CPU=       8.34;  WALL=      8.34
   Total time spent in section :FCTMTX  ;  CPU=    3660.07;  WALL=    520.04
   Total time spent in section :SOLRHS  ;  CPU=      81.98;  WALL=     11.57
   Total time spent in section :SAVFCT  ;  CPU=      32.89;  WALL=     31.11
   Total time spent in section :STRENR  ;  CPU=       4.14;  WALL=      3.73
   Total time spent in section :KINENR  ;  CPU=       0.00;  WALL=      0.00
   Total time spent in section :OBJFUN  ;  CPU=       0.00;  WALL=      0.00
TOTAL TIME SPENT IN MODULE :SOLVER  ;  CPU=    3916.96;  WALL=    704.95
TOTAL TIME SPENT IN MODULE :STRESS  ;  CPU=     119.78;  WALL=    118.96
TOTAL TIME SPENT IN MODULE :RESP 1  ;  CPU=       0.03;  WALL=      0.03
TOTAL TIME SPENT IN MODULE :SCREEN  ;  CPU=       0.02;  WALL=      0.02
TOTAL TIME SPENT IN MODULE :CONVERGE;  CPU=       0.06;  WALL=      0.06
TOTAL TIME SPENT IN MODULE :PRINT   ;  CPU=       0.01;  WALL=      0.01
   Total time spent in section :OP2     ;  CPU=   39045.11;  WALL=  39026.59
   Total time spent in section :MISC    ;  CPU=       0.64;  WALL=      1.29
TOTAL TIME SPENT IN MODULE :RESULTS ;  CPU=   39045.76;  WALL=  39027.88
TOTAL TIME SPENT IN MODULE :FINISHUP;  CPU=       4.36;  WALL=      4.37

CUMULATIVE RUN TIME:     CPU=    43551.31;     WALL=    40315.02

TIME SPENT IN MODULE :   CLOSE;  CPU=       0.46;  WALL=      0.98
 

 

 

Rahul Rajan_21763

Looks like you are running optimization run and it end up giving infeasible design. What type of analysis you are running?