[CIG-SEISMO] MPI setting (?) in SPECFEM3D

[Kenichi Tsuda]

Dear Surendra

 

Thanks for the comments.

 

I think I have used the latest version of the script for Cubit (I use Trelis 15.1.5).

But I’ll send it to you later.

 

Regards

 

Kenichi

 

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

Kenichi Tsuda, Ph.D.

Earthquake Science & Engineering Group

Center for Safety and Reliability Engineering

Institute of Technology Shimizu Corporation

Phone: +81-3-3820-5652

fax  : +81-3-3820-5959

 

From: Surendra Nadh Somala [mailto:surendra at iith.ac.in] 
Sent: Monday, February 29, 2016 12:04 AM
To: Kenichi Tsuda
Cc: cig-seismo at geodynamics.org
Subject: Re: [CIG-SEISMO] MPI setting (?) in SPECFEM3D

 

Dear Kenichi,

Is this a mesh that used to work with an earlier version of code but not with the latest?

While running fault kinematic/dynamics the edges of the fault need to be glued (except for the edge that intersects the free surface, if such an edge exists).  Multiple faults should be handled such that the nodes at intersection belong to only one of the two faults.  If these aren't handled properly you would encounter the kind of errors you are seeing.

 

If you can email me your input files (CUBIT mesh scripts), I might be able to figure out what's going wrong.

Best regards,

Surendra




___________

Surendra Nadh Somala

Assistant Professor
Department of Civil Engineering (Block-E, 208)
Indian Institute of Technology (IIT) Hyderabad

Kandi, Sangareddy
Telangana, India - 502285

Phone : +91 (0)40 2301-8457 

http://www.iith.ac.in/~surendra/ 

 

On Sat, Feb 27, 2016 at 1:53 PM, Kenichi Tsuda <kenichi.tsuda at shimz.co.jp> wrote:

Dear

My name is Kenichi Tsuda.
I'm working for the private company in Japan
and involving the project using SPECFEM3D related to fault dynamics.

Recently I have updated the latest version of SPECFEM3D package.
Then I tried to run the code and xdecompose_mesh looked working.

However, when I have run the generate_database, I met some errors
showing as follows and stopped the calculation.

--------------------------------------------
 application called MPI_Abort(MPI_COMM_WORLD, 30) - process 57
 application called MPI_Abort(MPI_COMM_WORLD, 30) - process 17
and
　Error detected, aborting MPI... proc           51
  error sorting MPI interface points:          46
    interface:           6       24207       24192
  error sorting MPI interface
  Error detected, aborting MPI... proc           46
 APPLICATION TERMINATED WITH THE EXIT STRING: Hangup (signal 1)
--------------------------------------------

Even after I updated some codes (using development version) such as the
parallel.f90, the situation is not changed.

When I do the configuration, I set the option
”MPI_INC=/opt/intel/impi/4.0.3.008/intel64/include”

Also the output_mesher.txt has stopped on the following spot
----------------------------------------------------------------------
create regions:

   ...allocating arrays
   ... reading            1  faults from file DATA/Par_file_faults
   ...setting up jacobian
   ...indexing global points
   ... resetting up jacobian in fault domains
   ...preparing MPI interfaces
----------------------------------------------------------------------

If you have any idea to be corrected or to be checked, please let me know.
I really appreciate your help.

Regards

Kenichi
----------------------------------------
Kenichi Tsuda, Ph.D.
Earthquake Science & Engineering Group
Center for Safety and Reliability Engineering
Institute of Technology Shimizu Corporation
Phone: +81-3-3820-5652
fax  : +81-3-3820-5959


_______________________________________________
CIG-SEISMO mailing list
CIG-SEISMO at geodynamics.org
http://lists.geodynamics.org/cgi-bin/mailman/listinfo/cig-seismo

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.geodynamics.org/pipermail/cig-seismo/attachments/20160229/a13822ba/attachment.html>