S1111: insufficient Virtual Memory Problem 1

[darcyo]

Hi, i am receiving a S1111: insufficient Virtual Memory Failure. i have read thread825-135347 and have followed the steps suggested by dcmacs:

1. Set DECOMPMETHOD to the PCGLSS solver.
2. Set SPARSEITERMODE to 3. This uses the least memory and but you may not get SPC forces, not sure.
3. Windows XP has the ability to access 3GB of RAM and I know NEiNastran supports this mode. I think the process is as simple as putting a /3GB in your boot.ini file but I would ask Noran support about this as they have a file that talks about it. Maybe you can find something on the Microsoft website about the 3GB switch.
and so on...

THE question is: do i need to set my boot.ini file to 3Gb switch, or should my analysis not run. i have set my page file to 4Gb. i have set RAM to 1100MB and 0MB. here is the final bit of my output file:


MEMORY ALLOCATED = 2215305 WORDS 270.4 MEGABYTES
DISK SPACE ACCESSED = 2129567 WORDS 260.0 MEGABYTES

NUMBER OF SOLUTION ITERATIONS = 723
FINAL CONVERGENCE FACTOR = 9.688E-06

SOLUTION TIME FOR 1519686 DEGREES OF FREEDOM = 661.4 SECONDS

MAXIMUM DISPLACEMENT MAGNITUDE = 2.928038E-03 AT GRID 535277
MAXIMUM ROTATION MAGNITUDE = 0.000000E+00 AT GRID 541170

EPSILON = 9.385623E-11
STRAIN ENERGY = 2.750800E+00

MAXIMUM SINGLE POINT CONSTRAINT FORCE MAGNITUDE = 0.000000E+00 AT GRID 541170
MAXIMUM SINGLE POINT CONSTRAINT MOMENT MAGNITUDE = 0.000000E+00 AT GRID 541170

WARNING S1111: INSUFFICIENT MEMORY AVAILABLE TO COMPLETE CURRENT OPERATION

CALCULATING RESULTS FOR SUBCASE 1

CALCULATING HEX ELEMENT RESULTS IN VOLUME 0 FOR SUBCASE 1

MAXIMUM HEX ELEMENT PRINCIPAL STRESS = 1.321945E+08 AT ELEMENT 426956
MINIMUM HEX ELEMENT PRINCIPAL STRESS = -1.304368E+08 AT ELEMENT 426740
MAXIMUM HEX ELEMENT SHEAR STRESS = 7.631613E+07 AT ELEMENT 426737
MAXIMUM HEX ELEMENT VON MISES STRESS = 1.365883E+08 AT ELEMENT 426737

CALCULATING PENT ELEMENT RESULTS IN VOLUME 0 FOR SUBCASE 1

MAXIMUM PENT ELEMENT PRINCIPAL STRESS = 1.481991E+08 AT ELEMENT 426741
MINIMUM PENT ELEMENT PRINCIPAL STRESS = -1.967385E+08 AT ELEMENT 426726
MAXIMUM PENT ELEMENT SHEAR STRESS = 7.894290E+07 AT ELEMENT 426726
MAXIMUM PENT ELEMENT VON MISES STRESS = 1.415459E+08 AT ELEMENT 426726

CALCULATING TET ELEMENT RESULTS IN VOLUME 0 FOR SUBCASE 1

MAXIMUM TET ELEMENT PRINCIPAL STRESS = 5.133748E+08 AT ELEMENT 256579
MINIMUM TET ELEMENT PRINCIPAL STRESS = -5.807417E+08 AT ELEMENT 256579
MAXIMUM TET ELEMENT SHEAR STRESS = 3.791114E+08 AT ELEMENT 358035
MAXIMUM TET ELEMENT VON MISES STRESS = 6.573231E+08 AT ELEMENT 358035

CALCULATING ELEMENT RESULTS ON CONTACT SURFACE 1 FOR SUBCASE 1
...


and that is it. it looks like it has found a solution, but then just stops. any help/assistance would be much appreciated.

hope this post isn't too long.
darcyo

 

[fkmeyers]

I would recommend the /3GB switch. I think you do need to modify the boot.ini. I would check the noran website knowledge base and see if they have any papers on this part. Then you can see if this model will run with SPARSEITERMODE set to AUTO, which would calculate the SPC forces. What was the matrix size? Paste the beginning of the .log file.

 

[darcyo]

thanks very much, the /3Gb switch worked really well. nastran handles a 1.5 mil DOF analysis no problem. very nice!