HI,there! Does anybody ever do some researches using parallized UDF? I try it the whole day, but can't tackle it! Can anybody show me the step of using UDF in parallel version. (I'm using windows network for parallized running) Thx!

Hi

Don't you have documentation for parallel UDF's. What kind of udf are you writing. I mean for what purpose. If you don't have documentation, I can send it to you in pdf format.

Vinod Dhiman

HI,Vinod Dhiman I do have the udf documentation, but I can't load the compiled libudf. The following is a simple parallized udf I copied from the documentation. I want to use it to test the procedure of using parallized udf. First, I launch the fluent in parallel version.(Windows Network) Then I load the case file in the fluent host. After that, I compile the following udf and get the library named "libudf".(the directory containing the libudf and case file have already been shared). Finally, i load the "libudf" . An error appears.

so, could you give me some advise!! Thx! #include "udf.h" DEFINE_ADJUST(where_am_i, domain) { #if RP_HOST Message("I am in the host process\n"); #endif /* RP_HOST */

#if RP_NODE Message("I am in the node process with ID %d\n",myid); /* myid is a global variable which is set to the multiport ID for

each node */ #endif /* RP_NODE */

#if !PARALLEL Message("I am in the serial process\n"); #endif /* !PARALLEL */ }

What kind of error do you get? Are you sure you're compiling the UDF for the parallel solver. Go to your UDF directory/ntx86: if you see 3ddp_host and 3ddp_node or (3d_host and 3d_node depending of your analuysis) the UDF is for the parallel. Luca

Yes, can you tell what is the error you are getting

Vinod Dhiman