Discussion Groups

Simulation CFD

Reply
*Expert Elite*
OmkarJ
Posts: 437
Registered: ‎10-02-2012

Multi threaded meshing

308 Views, 4 Replies
02-14-2013 04:06 AM

By default, while meshing, are all the cores on the system used or only one?

If I enable mesh_multicore in Flag manager and use 4 cores, will performance improve?

 

It would be handy since we often deal with meshes from 3 million even upto 9 million!

 

Regards

Omkar

Please use plain text.
*Expert Elite*
OmkarJ
Posts: 437
Registered: ‎10-02-2012

Re: Multi threaded meshing

02-15-2013 11:23 AM in reply to: OmkarJ

Someone?

Please use plain text.
Product Support
Royce_adsk
Posts: 632
Registered: ‎08-24-2011

Re: Multi threaded meshing

02-15-2013 02:37 PM in reply to: OmkarJ

You will only see 1 meshing process in task manager.  Personally I haven't done any benchmarks with the flag and don't generally use it, but I was told to expect about a max 2x speedup with it turned on.

 

What sort of speedup are you seeing? 

 

 



Royce Abel
Please use plain text.
Active Member
Erik.Gifford
Posts: 10
Registered: ‎04-29-2010

Re: Multi threaded meshing

02-19-2013 01:49 PM in reply to: Royce_adsk

Royce,

 

Checking this thread http://forums.autodesk.com/t5/Autodesk-Simulation-CFD/Multithreading-for-meshing/m-p/3777459/highlig... a user is indicating multi-core meshing could slow down the mesh process in some cases.  We're experiencing a very slow "...creating volume mesh" with a simulation utilizing the multi-core option set to 4 and it's only using approximately 25% across all all cores. Is this normal?  Would it  be better to revert to a lower core setting?

 

Thanks,

 

 

Please use plain text.
*Expert Elite*
OmkarJ
Posts: 437
Registered: ‎10-02-2012

Re: Multi threaded meshing

02-20-2013 10:58 AM in reply to: Erik.Gifford

Alright, here are the results: 

 

multithread.jpg

Clarifications:

 

1) The time was measured from pressing the Solve buttom until "Meshing complete" appeared in the message pane at the bottom. The process goes on till BC restart, Analysis completed successfully etc. but since the multithreading is used ONLY for meshing, I measured till meshing complete.

 

2) There were no extruded meshes and hence multithread meshing was not aborted.

 

3) Both the cases were done on the same part and same physics. To make sure the mesh elements remain close to 5.9 or 1.4 million, I used to coarsen by the factor of 1.05 and again refine by the factor of 0.95 for every measurement.

 

4) Solver computer was my own PC : 4 cores with 3.4 GHz, 16 GB RAM

 

5) The coarse and fine mesh was chosen to see if there is any marked benefit of multithreading for larger meshes. 

 

In a nutshell, I do not understand rationale behind multithreading since  I see no benefit at least in this benchmarking. I am inclined to believe that if I use cluster with 16 cores for meshing, I infact will end up slowing the process significantly. 

 

Guess the old ways are the best ways.

 

Cheers

OJ

Please use plain text.