CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   CD-adapco (https://www.cfd-online.com/Forums/cd-adapco/)
-   -   Proam ERROR:Could not allocate more memory (https://www.cfd-online.com/Forums/cd-adapco/54148-proam-error-could-not-allocate-more-memory.html)

Erika November 8, 2004 11:17

Proam ERROR:Could not allocate more memory
 
Hi,

When I try to run Tetfix to fix cpmatches or unresolved cells for mesh with almost 800000 trimmed cells, it often happens that it doesn't work and the BAMM Online Help appears with the following error:

**ERROR - COULD NOT ALLOCATE MORE MEMORY

**ERROR - COULD NOT ALLOCATE MEMORY TO INCREASE PARAMETER MAXSC2 TO 413009760 MAXSC2 IS CURRENTLY 43696879

I'm using a PC with 1G of RAM, but I don't think the reason is the RAM dimension beacuse I monitored the RAM status during the TetFix Run and it doesn't use the whole memory (almost 40 %).

Does anyone knows which could be the reason for this error and how can I fix it?

Thank you!

Brian November 8, 2004 14:21

Re: Proam ERROR:Could not allocate more memory
 
It is most certainly the RAM - you don't have enough for that operation.

Jihwan November 8, 2004 20:10

Re: Proam ERROR:Could not allocate more memory
 

Jihwan November 8, 2004 20:14

Re: Proam ERROR:Could not allocate more memory
 
I had the same experience. Even if you have enough memory for generating trimmed cells, STAR use much more memory for TetFix operation. I increased the RAM from 3G to 6G and finaly got no error message. Good luck Erika

Erika November 9, 2004 07:17

Re: Proam ERROR:Could not allocate more memory
 
I cannot increase the memory up to 3 or 6 G!!! How many cells did you have when you needed a 6G RAM?

Thank you anyway!

Jihwan November 9, 2004 20:32

Re: Proam ERROR:Could not allocate more memory
 
To my memory, I had 1.5 M trimmed cells when I cope with the problem. Anyway, you would need 64 bit CPU machine in order to fully take advantage of larger memory than 4G.

Kevin November 9, 2004 22:17

Re: Proam ERROR:Could not allocate more memory
 
Erika,

Sometimes it is helpful to close the model and prosize again. The MAXSC2 can be very difficult I have problems when using the ezip command.

For some reason I thought your custom mesh had cpmatch errors because you said during the generation of the trimmed mesh. Now I get it, you have unresolved cells and cpmatch erros after the trimming process.

When you did the trim, did you have the AFT and GROW TET FIX activiated also cell-division should be activiated? These are found in the advanced options.

I would probably fix by hand between 20 - 30 unresolved cells, for the cpmatches if you are O.K. with partial boundaries, then leave them, Did you run the following cpcheck.

cpcheck,,all,,,news,,fix

Kevin

Erika November 10, 2004 04:41

Re: Proam: Memory allocation and cpmatches errors
 
Ok, For MAXSC2 I will try to run prosize again.

For cpmatches errors, sure, I find it after the trimming! I usually have the ATF and GROW TETFIX activated and the CELL DIVISION not disabled (then activated) as default.

I tried to modify this option:

-disabling ATF and TETFIX I found more unresolved cells and sometime (but only sometimes) less cpmatches errors

-disabling CELL DIVISION I found concave cells

Yesterday I try to use your command cpcheck,,all,,,news,,fix but one cpmatches error remain and the programm stalled, then it has a segmentation fault and it closed itself. When I opened it again I found the database locked.

Don't worry, It was not the first time that it happens this "database locking": anyway, do you know how to fix this problem? I saw that the file .dbs is not empty as it seems, but I cannot unlock it!

For cpmatches errors I'm trying to use the TetFix procedure.

Thank you!

Kevin November 10, 2004 22:07

Re: Proam: Memory allocation and cpmatches errors
 
Erika,

I usually delete the files associated with *.lck*, I believe there should be two. When these are deleted the "dbase list" command will show all your databases.

I am suprised that is crashed during the cpcheck command.

By the way what is the geometry you are trying to mesh?

Kevin

Erika November 16, 2004 13:10

Re: Proam: Memory allocation and cpmatches errors
 
Hi Kevin,

I'm trying to process a pump ciruit from an inlet to an outlet, full of simply water with no slip walls. Unfortunatley I have problem also with the solution convergence in Comet!

Last week I obtained a good mesh with a boundary layer and no poorly formed cells (I fixed the cpmatches errors and Poorly Formed Cells by running TetFix), but I can't get the convergence in Comet!

I tryed several things by changing under relaxation factors, using pseudostationary model, trying different boundary conditions,... the mass residue decreases at the beginning, but diverges at the end!

Probably the geometry is too complicated, but it's the real one and I need to use it!

Do you use Comet to do CFD analysis? Any suggest?

Thank you!

Kevin November 16, 2004 21:42

Re: Proam: Memory allocation and cpmatches errors
 
Dear Erika,

I am impressed that you got a trimmed mesh into Comet without problems. I use Comet, but mainly for VOF, free surface flows. If you only want the solution for water than I suggest going back to star because it is more robust for simple flows. If you want free surface or caviatation, stick with Comet.

I suppose you already knocked down the under relaxation factor for pressure to be 0.02.

If it is a moving mesh, I can not help, but if you are looking for the stationary flow in a pump I can help, Is there a way for me to download the *.mdl file? Send me an e-mail.

Kevin


All times are GMT -4. The time now is 14:12.