Nastran DMP

[MertGumus]

Hello everyone,

My question is about the MSC Nastran DMP
(Distributed Memory Parallel)

I have a model with 400000 Elements and 4.5 Million DOF's.
And I also have about 4000 loadcases. The solution is linear static, SOL 101. One LC takes about 6 minutes with DMP module.

What is the fastest way to solve these 4000 LC's with a 24 core, 64 bit server?

Do you have any advice about processor, memory, bulk file settings? Should I distribute the LC's per 1 core, or use all the cores for 1 LC.

Thanks in advance for your comments.

 

[loki3000]

running one loadcase per core would be the best imo. but you will probably need a scheduler for that or a script so that two won't be running simultaneously (i.e. fighting for resources on the same core)