beginner's questions about fluent paralle processing
My computer has a core i7 920 (quad core) cpu, 12gb memory, and vista 64bit. Now I am using fluent 6.3 64bit to perform paralle computing. The grid info is as below:
Level Cells Faces Nodes Partitions 0 57756 121860 62462 1 The computation time for serial computing is about 9 h. The wall-clock time for 2-processer paralle computing is about 5.5 h. If I use more than 2 processors, the wall-clock time does not change. I am wondering if it is because the nodes are two few. Plus, I noticed that when fluent was running the page file usage increased. Can I increase the usage of memory and reduce that of page files? I think that would make the computation faster. Another question is that it is better to turn on SMT or not? Thanks in advance. |
Your mesh is too small.
Try with a mesh with 1 million cells you will see the difference ;) |
All times are GMT -4. The time now is 00:51. |