open TELEMAC-MASCARET The mathematically superior suite of solvers

Hi,

I am trying to get a 2D model up and running that has 10 tracers. I am using version v7p3r1 and attempting to run in parallel.

With 5 tracers the model runs fine but I am getting the following error with 10 tracers:

USING STREAMLINE VERSION 7.3 FOR CHARACTERISTICS
PARALLEL::ORG_CHARAC_TYPE1:: NOMB NOT IN RANGE [0..MAX_BASKET_SIZE]
MAX_BASKET_SIZE, NOMB: 10 14application called MPI_Abort(MPI_COMM_WORLD, 2) - process 0


I have tried to search this on the forum, I have read the related posts and tried many things but still haven't found the solution. Things I have tried include:

1. Attempting to change the MAX_BASKET_SIZE key word to 15 by copying streamline.f (from svn.opentelemac.org/svn/opentelemac/tags.../sources/utils/bief/)
editing MAX_BASKET_SIZE and the copying into my .f file.

2. Attempting to change the MAXIMUM NUMBER OF TRACERS. I attempted to do this by changing this as a keyword in declarations_telemac2d.f however I noted that MAXSCE is no longer a parameter that can be altered in V7. I noticed that on the release notes to V7 it states; "MAXIMUM NUMBER OF TRACERS These 4 new keywords are used to give an a priori dimension to a number of allocatable files, they give the values of respectively MAXFRO, MAXBLB, MAXSCE and MAXTRA." I assumed this meant that including MAXIMUM NUMBER OF TRACERS = 30;30;100;100 in my cas file would set the maximum number of tracers to 100 so I tried this.

Unfortunately neither of these worked and the error was unchanged. Note that I intend to introduce the tracers through water sources but for now I have just commented out the water sources in the cas file to try and get the model working without first.

My CAS file, .f and output file is attached. Any help would be much appreciated.

Many thanks,

Neil

Hi,

Here is my output file showing the error. The upload failed first time due to the type.

Hello Neil,

You should read this topic: www.opentelemac.org/index.php/kunena/16-...dd-salt-temp-tracers

Moreover, you should use the latest release v8p1r1.

Hope this helps,

Chi-Tuan

Thanks for the link Chi-Tuan, I had already read this and tried the suggestions regarding the advection of tracers.

SCHEME FOR ADVECTION OF TRACERS = 5;5;5;5;5;5;5;5;5;5
SCHEME OPTION FOR ADVECTION OF TRACERS = 4;4;4;4;4;4;4;4;4;4

Unfortunately this didn't resolve the issue regarding MAX_BASKET _SIZE and the output still reads:

USING STREAMLINE VERSION 7.3 FOR CHARACTERISTICS
application called MPI_Abort(MPI_COMM_WORLD, 2) - process 0
PARALLEL::ORG_CHARAC_TYPE1:: NOMB NOT IN RANGE [0..MAX_BASKET_SIZE]
MAX_BASKET_SIZE, NOMB: 10 14


Regarding V8 I have requested that this is installed on our cluster so hopefully I can test this out soon - do you think this could solve my problem?

Is it the case that I need to change the MAX_BASKET_SIZE somewhere other than in the streamline module?

Any other ideas very welcome.

Many thanks,

Neil

Hello,

Are sure you changed the two keywords and only them as you wrote?
Because if yes, you should only use streamlines (characteristics in parallel) for 2 variables (U and V) and NOMB should be equal to 2.
It is really strange that your error message speaks about 14 variables to be advected with this method.

Anyway, we only assist people using the latest tagged release (currently v8p1r1) or trunk version. If your issue still occurs, you know what to do.

Chi-Tuan

Thanks Chi-Tuan;

Yes I have triple checked everything and definitely have they keywords as I wrote. I attempted to re-run several times and tried various things such as changing the treatment of negative depths which it suggests in the manual but I still got the same result so I am a bit lost.

In any case I am hoping that I will be able to attempt to run on V8 at some point this week so I will let you know how I get on.

Hello,

Can you send your whole case (steering file, FORTRAN file + all input files).
If it can be run on the trunk without too much effort, I can try to have a look.

Chi-Tuan

Thank you Chi-Tuan, that is very kind of you.

I have tried lots of different things things but here is the case that I modified following your advice after my initial post.

As I said before I suspect that the TREATMENT OF NEGATIVE DEPTHS and TYPE OF ADVECTION keywords may no longer be appropriate following the changes you suggested to INITIAL VALUES OF TRACERS AND PRESCRIBED TRACERS VALUES. I have experimented with these trying to follow the advice given in the manual/user reference but I haven't had any success!

Many thanks,

Neil

Hello Neil,

Would it also be possible to have your GEOMETRY FILE (GEOMETRYcamel.slf) + LIQUID BOUNDARY FILES (LBcamel_20171202.qsl) please? I cannot see them.

Chi-Tuan

Sure - I guess they didn't upload due to the file extension. i'LL try changing them to .txt.

Thanks