Dear all, When running parallel (72 core) aggregate quantitities such as body forces (particularly pressure forces) seem to include a certain degree -low amplitude- of randomness in their evolution. The high frequency random oscillations seem to be of the order of the sampling period. This suggests some kind of issue with the gathering of data from all parallel processes for aggregate quantity computation. Is there some precaution that must be taken upon compilation or in defining the cluster parallel environtment to avoid this issue? I attach one such forces time-series to illustrate this. Viscous forces seem to be just fine, while pressure (and total) forces oscillate at high frequency. Cheers
Hi Mellibovsky, I am not an expert, but the same issue bothered me for a while. I finally managed to solve it by adding the <GLOBALSYSSOLNINFO> into the xml file (page 31, User Guide 4.4.0). This does not only smooth the force signal, but also appears to speed up the simulation by a magnitude. Hopefully some people can give a good explanation on it. Kind regards, Feifei -----Original Message----- From: nektar-users-bounces@imperial.ac.uk [mailto:nektar-users-bounces@imperial.ac.uk] On Behalf Of F Mellibovsky Sent: Friday, 12 January 2018 4:47 PM To: nektar-users Subject: [Nektar-users] forces computation in parallel computations Dear all, When running parallel (72 core) aggregate quantitities such as body forces (particularly pressure forces) seem to include a certain degree -low amplitude- of randomness in their evolution. The high frequency random oscillations seem to be of the order of the sampling period. This suggests some kind of issue with the gathering of data from all parallel processes for aggregate quantity computation. Is there some precaution that must be taken upon compilation or in defining the cluster parallel environtment to avoid this issue? I attach one such forces time-series to illustrate this. Viscous forces seem to be just fine, while pressure (and total) forces oscillate at high frequency. Cheers
Hi all, Mostly likely this happens because by default parallel simulations use an iterative method to solve the implicit global systems. This is in contrast to serial simulations, which by default use a direct solver. Since the iterative solver proceeds just until satisfying a prescribed tolerance, it is reasonable to expect that the solution might show fluctuations which depend on this tolerance. I do not have much experience on this topic, but the speed up observed by Feifei could be related to a better choice of preconditioners, compared to the default settings. Cheers, Douglas 2018-01-12 7:18 GMT-02:00 Feifei Tong <feifei.tong@uwa.edu.au>:
Hi Mellibovsky,
I am not an expert, but the same issue bothered me for a while. I finally managed to solve it by adding the <GLOBALSYSSOLNINFO> into the xml file (page 31, User Guide 4.4.0). This does not only smooth the force signal, but also appears to speed up the simulation by a magnitude. Hopefully some people can give a good explanation on it.
Kind regards, Feifei
-----Original Message----- From: nektar-users-bounces@imperial.ac.uk [mailto:nektar-users-bounces@ imperial.ac.uk] On Behalf Of F Mellibovsky Sent: Friday, 12 January 2018 4:47 PM To: nektar-users Subject: [Nektar-users] forces computation in parallel computations
Dear all,
When running parallel (72 core) aggregate quantitities such as body forces (particularly pressure forces) seem to include a certain degree -low amplitude- of randomness in their evolution. The high frequency random oscillations seem to be of the order of the sampling period. This suggests some kind of issue with the gathering of data from all parallel processes for aggregate quantity computation.
Is there some precaution that must be taken upon compilation or in defining the cluster parallel environtment to avoid this issue?
I attach one such forces time-series to illustrate this. Viscous forces seem to be just fine, while pressure (and total) forces oscillate at high frequency.
Cheers _______________________________________________ Nektar-users mailing list Nektar-users@imperial.ac.uk https://mailman.ic.ac.uk/mailman/listinfo/nektar-users
Dear all, This is not an answer, nor a solution. Just wanted to share an observation from running a parallel simulation, which should converge to a stationary solution I observe a kind of random variation in the kinetic energy - which matches the "low amplitude, random" description of Fernando. I thought it comes from the outflow condition I am using, or the high order modes, and I ignored it. Best, Stan data:image/png;base64,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 On 12.01.2018 09:46, F Mellibovsky wrote:
Dear all,
When running parallel (72 core) aggregate quantitities such as body forces (particularly pressure forces) seem to include a certain degree -low amplitude- of randomness in their evolution. The high frequency random oscillations seem to be of the order of the sampling period. This suggests some kind of issue with the gathering of data from all parallel processes for aggregate quantity computation.
Is there some precaution that must be taken upon compilation or in defining the cluster parallel environtment to avoid this issue?
I attach one such forces time-series to illustrate this. Viscous forces seem to be just fine, while pressure (and total) forces oscillate at high frequency.
Cheers
_______________________________________________ Nektar-users mailing list Nektar-users@imperial.ac.uk https://mailman.ic.ac.uk/mailman/listinfo/nektar-users
participants (4)
-
Douglas Serson
-
F Mellibovsky
-
Feifei Tong
-
Stanisław Gepner