f1 process could not be started in parallel mode of fluent

- July 1, 2019 at 2:00 am
  
  Weiqiang Liu
  Subscriber
  
  Hi all,
  
  In my case, I have 20 macros in source code. I can run under serial mode without any error. However, when I build, load and hook my source code under parallel mode, fluent complains about error 'f1 process could not be started'
  
  I wonder what is the reason for this. Do I have to modify my UDF under parallel mode? Or some other reasons for this?
  
  Best.
  
  Weiqiang
- July 1, 2019 at 4:59 am
  
  DrAmine
  Ansys Employee
  
  Actually parallelization is generally required whenever threads are passed or you want to read or write something or you rely on global reductions. Can you debug each macro or part alone? Check especially profile aNd adjust macros.
- July 1, 2019 at 3:35 pm
  
  Weiqiang Liu
  Subscriber
  
  Hi Amine,
  
  I debugged my code line by line. I found udf returned right value under serial mode. However, fluent complained error like 'f1 process could not be started'
  
  Do I have to modify my source code because of parallel reason?
  
  Best.
  
  Weiqiang.
- July 1, 2019 at 3:41 pm
  
  Rob
  Forum Moderator
  
  Maybe: are any of the macros/terms listed in the Parallelisation chapter of the manual?
- July 1, 2019 at 3:55 pm
  
  Weiqiang Liu
  Subscriber
  
  Hi ,
  
  do you mean I have some macros or items in the source code that are customized for parallelisation? I have the following macros in the code:
  
  DEFINE_SOURCE
  
  DEFINE_PROFILE
  
  DEFINE_PROPERTY
  
  By the way, I have UDS in my model and I use IF(THREAD_ID(t)==zone_ID) to recognize different zones.
  
  Best.
  
  Weiqiang