[makkks]

Hi all,

I am meshing some designs which contains 30M to 50M and are very complex in nature, hence it takes ages to mesh it.
Can some1 help me explain how parallel meshing works?
I have heard that the mesh size increses ridiculously if we use parallel meshing.
ALso can some1 explain advantages and dis advanatges of parallel meshing?

THanks a ton in advance

Cheerz
MAK

[PSYMN]

At 13.0, there is lots of distributed parallel solver tech at ANSYS, but all the meshing methods are limited to "shared memory parallel".

The speedup is not linear... Your second CPU may gain you 50 to 60% speed improvement, but then your third will only gain you an extra 25 to 40% of the first depending on the method, etc. By the time you add your 8th CPU, the increase is probably less than 10 percent. There are not exact numbers for this sort of thing because the exact speedup is very model specific.

Also note that not all meshing methods are SMP...

SMParallel meshing is a nice bonus (you will get your mesh sooner), but it isn't going to change your whole world...

Mesh size is not at all related to parallel meshing... It is dependent on your mesh parameters.

I don't know of any disadvantages, if you have a multi-CPU machine, then you might as well use at least 2 or 4 for meshing...

Our roadmap does call for a lot of distributed meshing coming with future releases...