CFD Online Discussion Forums

CFD Online Discussion Forums (
-   FLUENT (
-   -   floating point error during grid adaptation (

Nurul Murad September 3, 2001 00:18

floating point error during grid adaptation
hey guys..

i'm trying to refine my grid size by using the adapting the grid by using pressure gradient, velcoty gradient, volume change and also boundary condition...

however, i get a floating point error during the adaptation. usually after getting the error..the adaption register will still have the gradient, volume and boudary functions that i specified earlier.

the thing is...the grid size is refined. the thing that i'm not sure of do fluent distribute this grids eventhough an error message pops out and also what is the cause for that error message to pop up. thanks guys...


John C. Chien September 3, 2001 00:46

Re: floating point error during grid adaptation
(1). You probably need to track down the person who wrote that section of the code in the first place. (2). You might want to find out whether it is caused by the precision of math used, single precision vs double precision math. Is you code using double precision math. This is usually the first step in the debugging. (3). And it is also likely that the programmer forgot to check the denominator of a division operator in the code. Floating point error is just a Fortran error.(I am not sure whether you will get the same error message using other languages, such as c, C++.)

david September 5, 2001 19:15

Re: floating point error during grid adaptation
These might help A) if you get the error for single precision, try double precision next time and vice versa B) Adapt based on ONE solution parameter at a time, then run few iterations and adapt again based on other C) Limit the refinement, so that it's not getting too big.


All times are GMT -4. The time now is 14:01.