simpleFoam divergence due to hardware/compilation?

wildfire230 · May 16, 2018, 16:17

Hi all,

I have a case I am working with using simpleFoam with OpenFOAM 5.0. The problem has a lot of interesting features including variable viscosity and coupled transport equations. I have the same version of OpenFOAM compiled on my Ubuntu 16.04 laptop as well as on the Titan supercomputer at Oak Ridge National Lab. Basically on my laptop, the code seems to execute perfectly with nice convergence, but on Titan, after about 30 seconds of simulation time, simpleFoam suddenly diverges.

I have tried debugging every possible thing I can think of, but when I finally just ran an identical case on two different machines and found divergence on one (after running smoothly for a while), and no divergence on the other machine, I became very concerned. To be honest I do not know enough details about hardware/compilations to debug this problem.

Does anyone have any idea what could be going wrong???

Thanks

piu58 · May 17, 2018, 00:29

Assuming that there is no failure in the large machine, one has to ask What is different? I think that the floating point operating units may be different and with it the internal representation of floating point numbers. A setup at the limits of stability may crashes with some less digits in floating point representation.

May 16, 2018, 16:17	simpleFoam divergence due to hardware/compilation?	#1
wildfire230 Senior Member Join Date: Jul 2013 Posts: 124 Rep Power: 12	Hi all, I have a case I am working with using simpleFoam with OpenFOAM 5.0. The problem has a lot of interesting features including variable viscosity and coupled transport equations. I have the same version of OpenFOAM compiled on my Ubuntu 16.04 laptop as well as on the Titan supercomputer at Oak Ridge National Lab. Basically on my laptop, the code seems to execute perfectly with nice convergence, but on Titan, after about 30 seconds of simulation time, simpleFoam suddenly diverges. I have tried debugging every possible thing I can think of, but when I finally just ran an identical case on two different machines and found divergence on one (after running smoothly for a while), and no divergence on the other machine, I became very concerned. To be honest I do not know enough details about hardware/compilations to debug this problem. Does anyone have any idea what could be going wrong??? Thanks

May 17, 2018, 00:29		#2
piu58 Senior Member Uwe Pilz Join Date: Feb 2017 Location: Leipzig, Germany Posts: 744 Rep Power: 15	Assuming that there is no failure in the large machine, one has to ask What is different? I think that the floating point operating units may be different and with it the internal representation of floating point numbers. A setup at the limits of stability may crashes with some less digits in floating point representation. __________________ Uwe Pilz -- Die der Hauptbewegung überlagerte Schwankungsbewegung ist in ihren Einzelheiten so hoffnungslos kompliziert, daß ihre theoretische Berechnung aussichtslos erscheint. (Hermann Schlichting, 1950)

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Floating Point Exception Error	nyox	FLUENT	11	November 30, 2018 12:31
Two Phase Flow In Vertical Pipe	stonepreston	FLUENT	2	October 31, 2017 08:35
kklOmega SimpleFoam divergence, kt kl omega	NRA	OpenFOAM Running, Solving & CFD	9	February 18, 2017 08:54
PimpleFoam Womersley type Flow: Divergence due to underdetermined cells	90nash	OpenFOAM Running, Solving & CFD	4	July 1, 2014 01:14
solution divergence due to turbulence	Haider	FLUENT	1	October 18, 2006 07:27