Skip to end of metadata
Go to start of metadata

How to Compile UDFs

Compiled method adopts more features of C than the interpreted method does, e.g. implementation of multidimensional array data.

NOTE: Do not include any SPACE in the name of the source file!

  1. Put both the C source file and header files (if used) into the same folder as the case and data files.

    1. If you're running from WorkBench, put these files into dp0\FFF\Fluent.
       

  2. Open FLUENT.
     

  3. Unload previously compiled UDFs first to avoid potential errors. (This step can be skipped if the compiled method in the case did not use a UDF.)

    1. Define > Define UDF > Manager, unload previous compiled UDF library.
       

  4. Define >Define UDF > User Defined scaler, User Defined memory.  Give Fluent the correct number of UDS and UDMI if used. Just like what is done in interpreted method.

     

  5. Build a UDF library.

    1. Define-> Define UDF -> Compiled. Select the C source and header files, click “Build”. 

    2. Fluent will then start compiling. Errors will be reported if there are any. Warnings can be ignored.
       

  6. Click “Load”. All the variables that are returned, executed, initialized, or etc. should be shown to be successfully loaded.
     

  7. The remaining steps are identical to compiling functions from an interpreted method:

    1. Choose the conditions you want to use the UDF. Select the right functions.

    2. Go to Define > Define UDF > Function Hooks. Choose and hook the functions properly.
       

  8. The remaining steps are identical to compiling functions from an interpreted method.

Note that functions from compiled UDF are distinguished by the appendage ::libudf. Functions from interpreted methods do not have that.

We would like to credit Shengzhi He for writing this guide.