parallel and multiphase

Zaktatir · September 8, 2010, 07:24

hello,
i have a problem,
my calculation on 4 cpu's differs two much from this on 8 cpu's with the same setup? what did go wrong? (multiphase flow)

Claudia · September 8, 2010, 16:52

What differs exactly? Did you use the same setup? Did the simulation reach convergence?

Zaktatir · September 8, 2010, 17:17

Thanks for the reply.

I am using the same setup, setting a transient simulation with normal convergene criteria..

when i ran the simulation on 8 cpus the solver command told me that libmp_hp was ignored and it worked and reach the final time.

Even if am monitoring the sims with a couple of Points, there is difference ( in the Rms values too)..

I am using a cluster with different nodes, every node has an other amount of cpus ( 4 and 8)..

thanks beforehand

Zaktatir · September 16, 2010, 08:45

Is there any idea why differences exist

joey2007 · September 21, 2010, 13:56

Did you checked the usual suspects? Mesh quality and interfaces? Did you used double precision?

Zaktatir · September 21, 2010, 15:22

yes off course. No Improve
But i as turned Turbulence Numerics to High Resolution, all the Simulations are the same.
Ansys Support are trying to resolve the Problem too.

September 8, 2010, 07:24	parallel and multiphase	#1
Zaktatir Senior Member Join Date: Jan 2010 Location: Germany Posts: 268 Rep Power: 17	hello, i have a problem, my calculation on 4 cpu's differs two much from this on 8 cpu's with the same setup? what did go wrong? (multiphase flow)

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
Parallel Processing ocean wave multiphase model	robb	CFX	1	July 26, 2010 14:37
multiphase multicomponent physics	ckleanth	CFX	3	June 4, 2009 20:15
Problems with MFIX code and Parallel Processing.	Fernando Pio	Main CFD Forum	4	August 29, 2006 14:33
cfx4 parallel sover Problem!! HELP	Richard.A	CFX	2	August 3, 2002 03:54
parallel PHOENICS	George Bergantz	Phoenics	9	November 16, 2000 15:57

September 8, 2010, 16:52		#2
Claudia Member Join Date: Mar 2009 Posts: 38 Rep Power: 17	What differs exactly? Did you use the same setup? Did the simulation reach convergence?

September 8, 2010, 17:17		#3
Zaktatir Senior Member Join Date: Jan 2010 Location: Germany Posts: 268 Rep Power: 17	Thanks for the reply. I am using the same setup, setting a transient simulation with normal convergene criteria.. when i ran the simulation on 8 cpus the solver command told me that libmp_hp was ignored and it worked and reach the final time. Even if am monitoring the sims with a couple of Points, there is difference ( in the Rms values too).. I am using a cluster with different nodes, every node has an other amount of cpus ( 4 and 8).. thanks beforehand

September 16, 2010, 08:45		#4
Zaktatir Senior Member Join Date: Jan 2010 Location: Germany Posts: 268 Rep Power: 17	Is there any idea why differences exist

September 21, 2010, 13:56		#5
joey2007 Senior Member Join Date: Mar 2009 Location: Europe Posts: 169 Rep Power: 17	Did you checked the usual suspects? Mesh quality and interfaces? Did you used double precision?

September 21, 2010, 15:22		#6
Zaktatir Senior Member Join Date: Jan 2010 Location: Germany Posts: 268 Rep Power: 17	yes off course. No Improve But i as turned Turbulence Numerics to High Resolution, all the Simulations are the same. Ansys Support are trying to resolve the Problem too.