intro story D-Flow FM

 

D-Flow Flexible Mesh

D-Flow Flexible Mesh (D-Flow FM) is the new software engine for hydrodynamical simulations on unstructured grids in 1D-2D-3D. Together with the familiar curvilinear meshes from Delft3D 4, the unstructured grid can consist of triangles, pentagons (etc.) and 1D channel networks, all in one single mesh. It combines proven technology from the hydrodynamic engines of Delft3D 4 and SOBEK 2 and adds flexible administration, resulting in:

  • Easier 1D-2D-3D model coupling, intuitive setup of boundary conditions and meteorological forcings (amongst others).
  • More flexible 2D gridding in delta regions, river junctions, harbours, intertidal flats and more.
  • High performance by smart use of multicore architectures, and grid computing clusters.
An overview of the current developments can be found here.
 
The D-Flow FM - team would be delighted if you would participate in discussions on the generation of meshes, the specification of boundary conditions, the running of computations, and all kinds of other relevant topics. Feel free to share your smart questions and/or brilliant solutions! 

 

=======================================================
We have launched a new website (still under construction so expect continuous improvements) and a new forum dedicated to Delft3D Flexible Mesh.

Please follow this link to the new forum: 
/web/delft3dfm/forum

Post your questions, issues, suggestions, difficulties related to our Delft3D Flexible Mesh Suite on the new forum.

=======================================================

** PLEASE TAG YOUR POST! **

 

 

Sub groups
D-Flow Flexible Mesh
DELWAQ
Cohesive sediments & muddy systems

 


Message Boards

Issues with domain decomposition

JV
Jay Veeramony, modified 6 Years ago.

Issues with domain decomposition

Youngling Posts: 13 Join Date: 3/25/11 Recent Posts
When I try to run the example dd case, I get the following:

-----------------------------------------------------------------------------


Part VIII - Start Simulation...

Time to finish 0s, 0.0% completed, time steps left 60
Time to finish 0s, 0.0% completed, time steps left 60
Time to finish 0s, 0.0% completed, time steps left 60
forrtl: severe (40): recursive I/O operation, unit 35, file unknown
Image PC Routine Line Source
libintlc.so.5 00007FEBC0372A1E Unknown Unknown Unknown
libintlc.so.5 00007FEBC03714B6 Unknown Unknown Unknown
libifcore.so.5 00007FEBBBC0322E Unknown Unknown Unknown
libifcore.so.5 00007FEBBBB72C4E Unknown Unknown Unknown
libifcore.so.5 00007FEBBBBD5110 Unknown Unknown Unknown
libflow2d3d.so 00007FEBBCD3F456 Unknown Unknown Unknown
libflow2d3d.so 00007FEBBCDB9FC0 Unknown Unknown Unknown
libflow2d3d.so 00007FEBBC7E7C6D Unknown Unknown Unknown
libflow2d3d.so 00007FEBBC7D75B4 Unknown Unknown Unknown
libflow2d3d.so 00007FEBBC7DA3FC Unknown Unknown Unknown
libflow2d3d.so 00007FEBBC7DB2B6 Unknown Unknown Unknown
libflow2d3d.so 00007FEBBC7A778C Unknown Unknown Unknown
libflow2d3d.so 00007FEBBC797055 Unknown Unknown Unknown
libpthread.so.0 0000003E048079D1 Unknown Unknown Unknown
libc.so.6 0000003E044E8B6D Unknown Unknown Unknown
-----------------------------------------------------------------

This happens with the trunk version and the tagged version 3574. There are no issues with the other examples I tried (01, 02 03 04 and 07), including running the first example in parallel.

Any help would be welcome!

Thanks
jay
Adri Mourits, modified 6 Years ago.

RE: Issues with domain decomposition

Yoda Posts: 1224 Join Date: 1/3/11 Recent Posts
Hi Jay,

I know this problem may occur when compiling with the Gnu Fortran compiler and when several subdomains refer to the same input file(s). But "libifcore.so" suggests that you are using the Intel Fortran compiler. And the "example 2" testcase does not contain files being shared by subdomains.

Just to get things clear: You are compiling with the Intel compiler, both for the C and the Fortran code?

Regards,

Adri
JV
Jay Veeramony, modified 6 Years ago.

RE: Issues with domain decomposition

Youngling Posts: 13 Join Date: 3/25/11 Recent Posts
Hi Adri,

Yes, I am using the intel compiler suite, version 13.1

I tried with a number of other tagged releases and the last version successful in running the dd example was 5.01.00.2163.

Using the v3574 and compiling with debugging turned on gives this diagnostic:

veeramon@pope [09:57]:cat tri-diag.zz1

********************************************************************************
***
*** Deltares, FLOW2D3D Version 6.01.07.3675, Apr 30 2014, 14:02:24
*** built from : https://svn.oss.deltares.nl/repos/delft3d/tags/3574
***
*** runid : zz1
*** date,time : 2014-05-01, 09:55:14
***
********************************************************************************

*** MESSAGE Executable for 64-bits platform.
*** MESSAGE Double precision computation using reals of kind 8
*** MESSAGE Using UNESCO density formulation by default
*** MESSAGE Using default momentum solver cyclic
*** MESSAGE Transport solver cyclic-method method is specified
*** WARNING DPUOPT not prescribed, set to MEAN
*** WARNING Latitude of your model is 0 degrees, model located at equator
*** ERROR Error occured while reading file TMP_<runid>.bcc
Last time read: -Infinity
*** ERROR Flow exited abnormally


Thanks,
jay
Adri Mourits, modified 6 Years ago.

RE: Issues with domain decomposition

Yoda Posts: 1224 Join Date: 1/3/11 Recent Posts
Hi Jay,

Between 5.01.00.2163 and (6.00.00.)2367 is exactly the point where we introduced two flags to avoid this type of problems when using the Intel compilers:
-threads
-lifcoremt

There are two options I'm thinking about:
  • Try to leave a subset of these flags out (3 possible combinations). May be Intel changed the meaning of these flags
  • Check the Intel documentation related to these flags. May be Intel renamed them in version 13.x. I had a look myself on the Intel site but was not successful.


Please report back your findings.

Thanks.

Regards,

Adri
JV
Jay Veeramony, modified 6 Years ago.

RE: Issues with domain decomposition

Youngling Posts: 13 Join Date: 3/25/11 Recent Posts
Hi Adri,

I believe I found the issue. The flag "-fopenmp" was not being added to the FFLAGS and FCFLAGS variables. It was instead being placed in the variables OPENMP_FCFLAGS and OPENMP_FFLAGS.

Once I explicitly added this -fopenmp to FFLAGS and FCFLAGS, I was able to compile and run the dd example.

Thanks
jay



Adri Mourits:
Hi Jay,

Between 5.01.00.2163 and (6.00.00.)2367 is exactly the point where we introduced two flags to avoid this type of problems when using the Intel compilers:
-threads
-lifcoremt

There are two options I'm thinking about:
  • Try to leave a subset of these flags out (3 possible combinations). May be Intel changed the meaning of these flags
  • Check the Intel documentation related to these flags. May be Intel renamed them in version 13.x. I had a look myself on the Intel site but was not successful.


Please report back your findings.

Thanks.

Regards,

Adri
Adri Mourits, modified 6 Years ago.

RE: Issues with domain decomposition

Yoda Posts: 1224 Join Date: 1/3/11 Recent Posts
Hi Jay,

Thanks for sharing this.

I'm surprised that this helps. Openmp is not used by Delft3D-FLOW, only by DELWAQ.

Regards,

Adri