Parallel running optimitzation utility

dzordz · November 23, 2016, 03:52

Dear foamers,

I am using many different solver for my simulations and I also use each time different meshing, thus my number of cells varies a lot.

I am wondering if there exists a utility that would do all the hard work of figuring out what decomposition and on how many cores it is optimal to run the sim for a specific solver for a specific mesh?

Thanks

Ardali · November 23, 2016, 05:03

Hi,
Actually not.
Try to scotch to decompose your domain.
Try to stay higher than 40000 cells per processor.
Ardalan

dzordz · November 23, 2016, 05:11

Hi Ardali,

will for sure do that.

Extra question. Is there any way to know how much time (for each time step) the calculation took for each specific cell?

Ardali · November 23, 2016, 05:18

Hi,
That depends on very many things and not easy to use.
For incompressible single phase high quality structured mesh simply divide the wallclock time per time-step to the number of cells.
That is very close to impossible in other cases. Skewness, multiphase, compresiblity and so on matter.
Ardalan

November 23, 2016, 03:52	Parallel running optimitzation utility	#1
dzordz Member Join Date: May 2016 Posts: 39 Rep Power: 9	Dear foamers, I am using many different solver for my simulations and I also use each time different meshing, thus my number of cells varies a lot. I am wondering if there exists a utility that would do all the hard work of figuring out what decomposition and on how many cores it is optimal to run the sim for a specific solver for a specific mesh? Thanks

Thread Tools	Search this Thread
Show Printable Version Email this Page	Search this Thread: Advanced Search
Display Modes
Linear Mode Switch to Hybrid Mode Switch to Threaded Mode

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
error while running in parallel using openmpi on local mc 6 processors	suryawanshi_nitin	OpenFOAM	10	February 22, 2017 21:33
Running mapFields with Parallel Source and Parallel Target	RDanks	OpenFOAM Pre-Processing	4	August 2, 2016 05:24
Running AMI case in parallel	Kaskade	OpenFOAM Running, Solving & CFD	3	March 14, 2016 15:58
error when running in parallel, but succeed in a single core.	sharonyue	OpenFOAM Running, Solving & CFD	1	December 12, 2014 13:24
Problems running in parallel - missing controlDict	Argen	OpenFOAM Running, Solving & CFD	4	June 7, 2012 03:50

November 23, 2016, 05:03		#2
Ardali Member Ardalan Join Date: Jul 2012 Location: Atlanta, USA Posts: 77 Rep Power: 13	Hi, Actually not. Try to scotch to decompose your domain. Try to stay higher than 40000 cells per processor. Ardalan

November 23, 2016, 05:11		#3
dzordz Member Join Date: May 2016 Posts: 39 Rep Power: 9	Hi Ardali, will for sure do that. Extra question. Is there any way to know how much time (for each time step) the calculation took for each specific cell?

November 23, 2016, 05:18		#4
Ardali Member Ardalan Join Date: Jul 2012 Location: Atlanta, USA Posts: 77 Rep Power: 13	Hi, That depends on very many things and not easy to use. For incompressible single phase high quality structured mesh simply divide the wallclock time per time-step to the number of cells. That is very close to impossible in other cases. Skewness, multiphase, compresiblity and so on matter. Ardalan