Can someone please tell me if it is NECESSARY to parallelize UDF codes if you are using a cluster to run your case??
My problem is that my case and UDF code run flawlessly on my desktop but as soon as I run the case on our computing center everything falls apart and I get very bad results.
Any input is appreciated.
it depends. do you compile your udf or interpret it? does it return any errors when reading the case on cluster or when reading C file? what is the DEFINE macro you used?
make the problem more clear.
Thanks for the quick response,
No I dont have any problems reading the case and data files or the UDFs, but the results that I get are totally different( AND WRONG) from what I get on my desktop (WHICH ARE MOST DEFINITELY CORRECT).
macros that I use include execute at end, CG-motion, define-geometry, define-profile.
Can you explain to me what exactly you mean when you say it depends?
|All times are GMT -4. The time now is 22:23.|