[MITgcm-support] Re：Re: OBCS errors in OB location vs Mask

[Jean-Michel Campin]

Hi Liu,

> The message in STDERR.0000 is
> *** ERROR *** OBCS_CHECK:   222 errors in OB location vs Mask
> *** ERROR *** S/R ALL_PROC_DIE: ending the run
You can check the other STDERR.00??, and you should find a total of
222 error messages telling what is wrong.

And you may want to check some of the OB indices values in STDOUT.00?? files:
1) in global domain map (in you case, 640 x 640 map size),
  just after "OBCS_READPARMS: finished reading data.obcs"
2) in local processor tile domain (sNx x sNy, in your case, 80 x 80 ),
  after "OBCS_CHECK: start summary:"
  you should be able to find some repport for each tile that looks like:
(PID.TID 0000.0001) ======== Tile bi=   1 , bj=   1 ========
(PID.TID 0000.0001)  OB_Jn = /* Northern OB local indices */
(PID.TID 0000.0001)    46 @      -99                            /* I = -2: 43 */
(PID.TID 0000.0001)  OB_Js = /* Southern OB local indices */
(PID.TID 0000.0001)    46 @        1                            /* I = -2: 43 */
(PID.TID 0000.0001)  OB_Ie = /* Eastern OB local indices */
(PID.TID 0000.0001)    27 @      -99                            /* J = -2: 24 */
(PID.TID 0000.0001)  OB_Iw = /* Western OB local indices */
(PID.TID 0000.0001)    27 @        1                            /* J = -2: 24 */
(PID.TID 0000.0001) ======== Tile bi=   2 , bj=   1 ========

Otherwise, are you compiling pkg/exch2 (and using it, since the switch
is at compile time) ? And do you specify an "interior mask" file,
i.e., insideOBmaskFile= 'some_file_name',  in parameter file "data.obcs" ?

Cheers,
Jean-Michel

On Sun, Sep 14, 2014 at 11:32:42AM +0800, 刘成彦 wrote:
> 
> Dear Carroll and Michael
>  
> Thank you for your help so much.
>  
> Re Carroll:
> I have checked my values for sNx,sNy & nPx,nPy in the SIZE.h file.
> sNx =  80,
> sNy =  80,
> nPx =  8,
> nPy =  8,
> It means that I would use 64 CORES to run this experiment with grids points of 640*640.
> I also set useOBCS =.TRUE. In data.pkg
>  
> Re Michael:
> I do think your suggestions are proper answer to my problem. But my model cannot been
> integrated for only step due to this problem here. So I cannot get the hFac[WS]. Indeed, I want
>  to overcome this problem by a simple way, setting the southern boundary as a solid wall!
> In fact, the topography in Prydz bay in the southern boundary is also depicts as land entirely
> without ocean.
> In my data.obcs, I have set the solid wall property for the Southern boundary as follows:
> 
> OB_Jsouth =   640*0,
> 
> (I think ’640*0’ denotes there is no open boundary here, but it seems does work.
> If I am wrong, how to set it as solid wall?)
>  
> Again, thanks both of you a lot!
> Cheers
> Yours sincerely Liu
> 
> 
> 
> ----------------
> *********************************************
> Dr. Chengyan Liu
> Nanjing University of Information
> Science & Technology
> No.219, Ningliu Road, Nanjing, Jiangsu,
> China(Postcode: 210044)
>   Phone: 18251854735
> E-Mail: killy at nuist.edu.cn
> 
> ********************************************* 
>  

> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-support