Hi, I am wondering if it is possbile to use an interpreted UDF (simple function describing a time-dependent velocity inlet) instead of compiled UDF while running a case on a parallel computing cluster? I'm having problems compiling that UDF but I thinking that this struggle might be avoidable? Any comment is welcome.

Hi, do you have problems when you interprete your UDF in the serial solver? Luca

Hi Luca, Thanks for responding. To answer your question, no I do not have problem interpreting my UDF in the serial solver. My problem at the moment is that I cannot compile it in the serial solver although I have set the environment variables as described on the Fluent online help site. I might be able to avoid this problem... Do you know if it is possible to use such a interpreted UDF on a parallel solver? Thanks again David

I not sure you can use interpreted UDF in the parallel solver. Have you set the environmet variables for visual C++?Luca

Hi Luca, I tried to set up the Visual C++ environment variable s as described in the following link:

http://www.fluentusers.com/support/i.../udfguides.htm

After having set the variables, I typed cl in a command prompt window. Ok. I get a similar message but the version of the compiler is slightly different (Compiler Version 12.00.8186 instead of 12.00.8804.)

When I type link/version in a new prompt window I unfortunately do not get what I should. I get the following:

Microsoft (R) Incremental Linker Version 6.00.8168 Copyright (C) Microsoft Corp 1992-1998. All rights reserved. LINK : fatal error LNK146: no argument specified with option "/version"

Note that the linker version is not the one suggested by Fluent. I downloaded Microsoft Visual Studio 6 full version of Service Pack 6 and installed it so I'm surprised to see that the versions (both for compiler and linker) differ from what is suggested on the Fluent website.

Any suggestions? Let me know if you'd like to see the log file generated by FLuent when attempting to build the library. Many thanks David

Yes send me the log file. Luca

Will do. I will send it by email.

Hi all, just to follow up with this problem. I turned out that the problem was within my UDF code. When interpreted, I had no problem. When compiled, the compiler picked up syntax errors that were causing this problem. Regards, David

[amindavari]

could anyone help me: i provided an udf that set a profile on the boundary, at first i couldnt even interprete this udf, however by adding two lines in code i finally could interprete it. now i encounter a new problem, when i start iterating, an error entitled "Floating point error: invalid number" appears. what should i do?
thank you in advance


#include "udf.h"
DEFINE_PROFILE(heatflux, thread, position){
#if !RP_HOST
real y[ND_ND]; /* this will hold the position vector */
real z;
face_t f;
begin_f_loop(f, thread)
{
F_CENTROID(y,f,thread);
z = y[2];
F_PROFILE(f, thread, position) = 283889.07*cos((4.0*atan(1.0)/0.815)*(z+0.01005));
}
end_f_loop(f, thread)
#endif
}