TELEMAC-MASCARET Forum: T2DTomSis 3-way coupling: tom I/O error with Computation Continued (1/1)

Welcome, Guest

TOPIC: T2D<>Tom<>Sis 3-way coupling: tom I/O error with Computation Continued

T2D<>Tom<>Sis 3-way coupling: tom I/O error with Computation Continued 6 years 10 months ago #28532

ag_davies
OFFLINE
Junior Boarder
Posts: 36

Hi Thierry and others,

I have been attempting a 3-way coupling for a large coastal domain, with tides, waves and sediments including bed-roughness feedback, using v7p2r0 on HPC Wales.

My initial run (12 hr duration) ends correctly and the solution files for T2D, Tom & Sis are fine. [The only caveat here is that any attempt to initialise Sisyphe with a PREVIOUS SEDIMENTOLOGICAL COMPUTATION FILE appears to be ignored.]

However, my query here concerns continuation of the solution (COMPUTATION CONTINUED, NEXT COMPUTATION) which has been unsuccessful so far. On continuation the resumption time is identified correctly (i.e. at 12hr, see output error 1 attached for a sequential continuation run). However, the run fails during its call from wac.F to ecrspe.f, and specifically, I think, at the call made from here to P_SYNCH.

The same error involving I/O to unit 17 occurs both for a sequential run (1 processor - see error 1 attached) and also for parallel continuation (e.g. with 16 processors). There are no external I/O files involved in any of T2D, Tom or Sis, though Sisyphe uses 4 Private Arrays.

If I comment out the calls to ecrspe.f, the continuation run then does make some progress:-

- For a sequential run (1 processor), three correct solution files are produced (t2d, tom & sis). But the global results file WACBRI is empty, meaning that no further continuation is possible.

- For a parallel run (16 processors), the continuation run again ends correctly. However, the results files are not moved out of the working directory and an error is generated ('Attempting to use an MPI routine before initializing MPI' - see error 2 attached). In this case, the sub-domain results are merged correctly for t2d (T2DRES) and tom (WACRES). However the gretel operation does not happen for Sisyphe (i.e. no SISRES), though the subdomain SIS solutions are available in the working directory. The global results file (WACBRI) is again empty following an error with runGRETEL. So in this parallel case there is no merged Sisyphe solution and again further continuation is not possible for tomawac.

I wonder whether anyone else has encountered this kind of issue, possibly involving the numbering of (tomawac) channel units in coupled mode? Any insights will be very much appreciated.

Kind regards, Alan

File Attachment:

File Name: Errors12_Forum_jan18.doc
File Size: 16 KB

The administrator has disabled public write access.

T2D<>Tom<>Sis 3-way coupling: tom I/O error with Computation Continued 6 years 10 months ago #28539

cyamin
OFFLINE
openTELEMAC Guru
Posts: 997
Thank you received: 234

Hello Alan,

I have been working successfully with a similar setup like yours except TOMAWAC is not coupled, but chained and no tide. I do continuation for all models and it works fine. I do remember that I had overcome a few issues related to TOMAWAC continuation/global results file (Thierry made corrections to the trunk) and now everything works fine for me with the trunk.

I would strongly encourage to try at least with the latest tag, or even the trunk. The corrections may apply to your case as well. Or it could be that 3way coupling and continuation with TOMAWAC is not thoroughly tested.

Regarding SIS continuation, it works fine for me:

================================================================================
 ITERATION        0    TIME:  5 H 20 MIN   0.0000 S   (    19200.0000 S)
 PRERES: MAXIMUM COURANT NUMBER:    0.2070578E-01
 TELEMAC2D COUPLED WITH: SISYPHE
 TELEMAC-2D: INTERNAL COUPLING WITH SISYPHE
 PART 0 : INITIALISING SISYPHE
 INBIEF (BIEF): NOT A VECTOR MACHINE (ACCORDING TO YOUR DATA)

 LIQUID BOUNDARIES:           1
 FONSTR : FRICTION COEFFICIENTS READ IN THE
          GEOMETRY FILE
 STRCHE (BIEF): NO MODIFICATION OF FRICTION

 TITLE OF PREVIOUS COMPUTATION: [112.5_157.5)_[0.5_1)

 NAME: WATER DEPTH       UNIT: M
 NAME: BOTTOM            UNIT: M
 NAME: RIGID BED         UNIT: M
 NAME: SOLID DISCH       UNIT: M2/S
 NAME: SOLID DISCH X     UNIT: M2/S
 NAME: SOLID DISCH Y     UNIT: M2/S
 NAME: EVOLUTION         UNIT: M
 NAME: RUGOSITE TOTALE   UNIT: M
 NAME: MEAN DIAMETER M   UNIT:
 NAME: FRACLAY 1  CL 1   UNIT:
 NAME: FRACLAY 2  CL 1   UNIT:
 NAME: FRACLAY 1  CL 2   UNIT:
 NAME: FRACLAY 2  CL 2   UNIT:
 NAME: FRACLAY 1  CL 3   UNIT:
 NAME: FRACLAY 2  CL 3   UNIT:
 NAME: FRACLAY 1  CL 4   UNIT:
 NAME: FRACLAY 2  CL 4   UNIT:
 NAME: FRACLAY 1  CL 5   UNIT:
 NAME: FRACLAY 2  CL 5   UNIT:
 NAME: FRACLAY 1  CL 6   UNIT:
 NAME: FRACLAY 2  CL 6   UNIT:
 NAME: FRACLAY 1  CL 7   UNIT:
 NAME: FRACLAY 2  CL 7   UNIT:
 NAME: FRACLAY 1  CL 8   UNIT:
 NAME: FRACLAY 2  CL 8   UNIT:
 NAME: FRACLAY 1  CL 9   UNIT:
 NAME: FRACLAY 2  CL 9   UNIT:
 NAME: LAYER1  THICKNES  UNIT: M
 NAME: LAYER2  THICKNES  UNIT: M
 NAME: LAYER1  CONC      UNIT: KG/L
 NAME: LAYER2  CONC      UNIT: KG/L

 READ_DATASET : READ OF RECORD     4

Regards,
Costas

The administrator has disabled public write access.

The following user(s) said Thank You: ag_davies