Dear all,

Hi, I would like to ask what the meanings of following error messages in Solver are.

ERROR #001100279 has occured in subroutine ErrAction. Message: c_fpx_handler: Floating point exception: Invalid Operand

ERROR #001100279 has occured in subroutine ErrAction. Message: Stopped in routine c_fpx_handler.

Anyone can kindly help me to fix this problem? Thank you in advance.

Sara

You have gotten a floatin point exception. In your case you've gotten an invalid operand. This could mean several things. For example, square root of a negative number, or a ieee non-conformant number. The release notes for 5.5 talk about not mixing DEC with some other platforms, did you do this? If so, don't mix them.

Neale.

Hi Neale,

Thank you for your reply. Actually, I'm a beginner of CFX. So, I do not quite familiarise with the thing you mentioned, e.g. I don't know what the DEC is, could you kindly tell me in more details?

By the way, do you think whether my problem is due to some wrong settings in the part of "Solver Control" of CFX-Build (I'm using CFX 5.4.1)?? However, it can write a definition file properly. So, can you give me advice to check what the problem is? Thank you very much.

regards, Sara

The DEC is a UNIX computer platform. Don't worry about it if you don't know what it is. What computer platform are you running on?

You should really upgrade to 5.5, it's been out for 3-4 months now. At any rate though, does your problem run for a while, then do this, or does it do it right away. Other questions would be: What are you modelling? What models do you have turned on? Did you try turning things off or changing boundary conditions until it didn't floating point? More details are necessary.

Neale.

Hi Neale,

The computer platform I'm running on is SUN workstation, i.e.UNIX platform. In fact, what I'm trying is to model the dispersion of pollutant. And this problem does not happen so often. When I change the sizes of the geometry to a certain value, the problem happen. But of course I really don't know what the critical value of the geometry size will cause the problem. So, any suggestion to me to find out the problem? and how do the boundary condition related to the floating point?

by the way, could you kindly tell me the definition of "floating point" in CFX? I really don't know what it is indeed.

Thank you very much in advance. I look forward to hearing from you.

Sara

It's hard to say what's going on with your case without actually seeing it. So, all I can do is give you suggestions. If the problem floating points when you change the geometry then the mesh has probably changed in a way which is problematic. You might load it into CFX-Visualise and see how many bad tets you have for example. If there are a large number then you might need to add mesh controls so that the mesh generator makes nicer tets in the problem areas.

You should also carefully review your problem setup for errors in boundary condition setup, and physical problem definition. I'm not sure what physics you all have on, but you might also try switching off certian models and see if it works without them. At least then you have a hint of what exactly is causing the problem.

One other trick is to set the expert parameters "backup frequency = 1" and "delete backup files = f" so that you get a series of backup files which you can load into visualise or CFX-Post sequentially and see what is happening transiently as the solution converges.

A floating point exception occurs when a real number becomes unrepresentable by the IEEE standard. This occurs when software attempts to perform operations which would necessarily generate an invalid result which can't be represented, an overflow which is out of range, and an underflow which is also out of range. For example, operations such as 1/0, 0/0, sqrt(-1) would all generate a floating point exception. Whether or not the software stops when one occurs is up to the developers, in the case of CFX-5 it will stop.

Hi Neale,

Yes, I think the problem may be due to lack of mesh controls. However, can you share with me, from your past experience, when or under what circumstance that we need to add mesh controls? and how to determine whether the point or line mesh controls should be used? and how to determine the values of length scale, radius and expansion factor?

Thank you very much for your helping.

Sara

[sunilkonatham]

Originally Posted by Neale
;63678

The DEC is a UNIX computer platform. Don't worry about it if you don't know what it is. What computer platform are you running on?

You should really upgrade to 5.5, it's been out for 3-4 months now. At any rate though, does your problem run for a while, then do this, or does it do it right away. Other questions would be: What are you modelling? What models do you have turned on? Did you try turning things off or changing boundary conditions until it didn't floating point? More details are necessary.

Neale.

Hi Neale,
I got same error as posted above. The thing is, it is showing this error only when I run it on cluster (which is of linux). It is running without any error on my local workstation. You have said about mixing of platforms I guess it could be related to something like that, can you please help me resolve this?

[tsibel]

Originally Posted by sunilkonatham

Hi Neale,
I got same error as posted above. The thing is, it is showing this error only when I run it on cluster (which is of linux). It is running without any error on my local workstation. You have said about mixing of platforms I guess it could be related to something like that, can you please help me resolve this?

Hi


I am running two phase flow (boiling) and I have the same error when I am using cluster. How did you solve this problem?


Thanks