CFD Online Logo CFD Online URL
www.cfd-online.com
[Sponsors]
Home > Forums > FLUENT

How to distribute the case in parallel computing?

Register Blogs Members List Search Today's Posts Mark Forums Read

Reply
 
LinkBack Thread Tools Display Modes
Old   October 14, 2010, 10:55
Default How to distribute the case in parallel computing?
  #1
Senior Member
 
lakeat's Avatar
 
Daniel WEI (老魏)
Join Date: Mar 2009
Location: South Bend, IN, USA
Posts: 688
Blog Entries: 9
Rep Power: 12
lakeat is on a distinguished road
Send a message via Skype™ to lakeat
I have a large unsteady LES case, my grid number would be around 10 million, that is 10,000,000, so I donít know how to make the best use of the parallel computing. How to distribute the grids to each cpu so to make the simulation most efficient? How many grids a single cpu and a single node bear with highest efficiency?

I know the question is kind of general, but could you give me also both general guidelines for it, and also your detailed experience, for example, one of your large cases, how you distribute it.

Thank you so much.
__________________
~
Daniel WEI
-------------
NatHaz Modeling Laboratory
Department of Civil & Environmental Engineering & Earth Sciences
University of Notre Dame, USA
Email || My Personal CFD Blog
lakeat is offline   Reply With Quote

Old   October 14, 2010, 12:52
Default
  #2
New Member
 
Join Date: Sep 2010
Posts: 12
Rep Power: 6
csmith_PADT is on a distinguished road
Fluent will distribute the model for you, as long as you start it in parallel mode. From the Fluent "startup" panel, you have the option to select a parallel solution as well as the number of processors.

There are ways within Fluent to decompose the model across processors manually, but typically the automatic mesh partitioner does the job very well.
csmith_PADT is offline   Reply With Quote

Old   October 14, 2010, 13:05
Default
  #3
Senior Member
 
lakeat's Avatar
 
Daniel WEI (老魏)
Join Date: Mar 2009
Location: South Bend, IN, USA
Posts: 688
Blog Entries: 9
Rep Power: 12
lakeat is on a distinguished road
Send a message via Skype™ to lakeat
Quote:
Originally Posted by csmith_PADT View Post
Fluent will distribute the model for you, as long as you start it in parallel mode. From the Fluent "startup" panel, you have the option to select a parallel solution as well as the number of processors.

There are ways within Fluent to decompose the model across processors manually, but typically the automatic mesh partitioner does the job very well.
But I mean how many grid points for each cpu, any guidelines?
__________________
~
Daniel WEI
-------------
NatHaz Modeling Laboratory
Department of Civil & Environmental Engineering & Earth Sciences
University of Notre Dame, USA
Email || My Personal CFD Blog
lakeat is offline   Reply With Quote

Old   October 14, 2010, 15:05
Default
  #4
New Member
 
Join Date: Sep 2010
Posts: 12
Rep Power: 6
csmith_PADT is on a distinguished road
The answer depends on the memory associated with each processor, and the nature of the physics you are solving. The more complicated the physics, the more memory overhead. You could probably solve your LES model with 10 million cells on a machine with around 24 GB of RAM.
csmith_PADT is offline   Reply With Quote

Reply

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are On
Pingbacks are On
Refbacks are On


Similar Threads
Thread Thread Starter Forum Replies Last Post
Superlinear speedup in OpenFOAM 13 msrinath80 OpenFOAM Running, Solving & CFD 18 March 3, 2015 06:36
Performance of GGI case in parallel hannes OpenFOAM Running, Solving & CFD 26 August 3, 2011 03:07
About Gambit and Parallel Computing yuki FLUENT 2 September 7, 2009 01:13
parallel computing rui CD-adapco 0 June 25, 2007 06:18
DHCP, Cluster computing, Parallel processing ? Jihwan CD-adapco 1 January 2, 2005 18:32


All times are GMT -4. The time now is 17:56.