[Nek5000-users] turbChannel compiling issue after adding more elements

[nek5000-users at lists.mcs.anl.gov]

The local problem size (defined by the local number of elements) is important when it comes to time to solution and memory consumption.

Typically the 2GB limit is reached for lx1=8 and 500-1000 elements. Yes, this limit can be relaxed by e.g. -mcmodel=medium but this is not needed.  The 2GB limit is not a real limitation as in almost all production runs the number of elements per MPI rank is way lower (reasonable time to solution).

In your case you compiled Nek5000 for a local problem size of lelg/lpmin = 36864 elements. This is a pretty huge number!
No way this fits into the 2GB memory limit. Moreoever, your computation would take forever if you would run Nek5000 with just lpmin MPI ranks.
  

Hope this helps,
Stefan

-----Original message-----
> From:nek5000-users at lists.mcs.anl.gov <nek5000-users at lists.mcs.anl.gov>
> Sent: Thursday 25th January 2018 18:21
> To: nek5000-users at lists.mcs.anl.gov
> Subject: Re: [Nek5000-users] turbChannel compiling issue after adding more elements
> 
> You need a way larger lpmin for 73728 elements say 128!
> 
> On 25 Jan 2018, at 18:05, "nek5000-users at lists.mcs.anl.gov <mailto:nek5000-users at lists.mcs.anl.gov>" <nek5000-users at lists.mcs.anl.gov <mailto:nek5000-users at lists.mcs.anl.gov>> wrote:
> 
> Hello,
> Thank you all for your suggestions. However they did not really work either. The current work on the turbulent channel is in fact based on the original turbChannel in the example. The current
> lelg number is the actual number of elements (73728) I used. I also tried to increase the
> lelg a bit to 73750, which did not make any difference. I decreased lpmin from 4 to 2 and unfortunately this did not work either. I wonder if there is any actual limitation to the number of elements that shall be used provided the physical dimentions
>  defined in the code (2 pi delta * pi delta * 2 delta).
> Thank you all for your time and suggestions!
> Regards,
> Ming
> _______________________________________________
> Nek5000-users mailing list
> Nek5000-users at lists.mcs.anl.gov <mailto:Nek5000-users at lists.mcs.anl.gov>
> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users <https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users>
> _______________________________________________
> Nek5000-users mailing list
> Nek5000-users at lists.mcs.anl.gov
> https://lists.mcs.anl.gov/mailman/listinfo/nek5000-users