Creative meshing

[TidbitSoftware]

NOTE: This discussion was originally posted to the ISSM Forum, which has been decommissioned. It is reproduced here for reference. Please start a new Discussion or Issue if you wish to ask something related to this discussion.

---

lkaluzi
I'm using bamg mesh to adapt the mesh of my domain using both ice velocity and bed topography. There's also another area of my domain that I want to refine as well that doesn't doesn't have a strong gradation in velocity or topography (or at least the gradation isn't as pronounced as other areas of the domain). I tried specifying a different hmax for the area of interest, but the result is jarring (all of the cells in the area of interest are the same size). I was trying to find another way to go about this, perhaps defining a different 'err' value within that particular region, or extracting a subdomain for the area of interest and remeshing further but ran into issues with each of these ideas. Has anyone tried something like this?

---

mathieumorlighem
Hi Lynn,
it may be easier to answer your question with a plot, but you should definitely use both 'err' for the entire mesh and 'hMaxVertices' for this specific area (with NaN everywhere else), and maybe set it to different values depending on what you want. Sorry it's hard to give you another answer, 'err' cannot vary in space right now.
Best
Mathieu

---

lkaluzi
Thanks Mathieu. Here is my original mesh.

I'm using the following values
hinit=30000; % element size for the initial mesh
hmax=25000; % maximum element size of the final mesh
hmin=1000; % minimum element size of the final mesh
gradation=1.7; % maximum size ratio between two neighboring elements
err=7; % for bed topography
err2=10; % for velocity
I'm also applying an hmin values of 50000 for regions I'm not interested in.

Here is what happens when I apply hMax of 3000 in my ROI.

Ideally, there'd be a way to bump up the resolution in this region without over doing it.

---

mathieumorlighem
Hi Lynn, I looked a bit into it today and it should be possible to provide an error per vertex instead of one single value. Do you have access to the development branch (trunk-jpl) or do you use the public branch (trunk)? I will be away for the rest of the week but I can try to make the necessary changes next week?
Best
Mathieu

---

lkaluzi
Hi Mathieu, I don't have access to the development branch, how do I go about getting access?
This not a pressing issue so next week would be great. Thank you!

---

mathieumorlighem
Did you install from source? or did you use the binaries directly?

---

lkaluzi
I installed from the binaries. I probably need to redownload soon anyway. I'm getting a weird PETSC error when using the requested_output option for a transient run. Should I try to download from source?

---

mathieumorlighem
it is just that it takes much longer (months) for changes to "propagate" from the developer branche to the public branch so you may need to wait a little longer. Out of curiosity, what error message do you get with PETSc?

---

lkaluzi
Here's the error. Only when I try to request outputs such as ice volume above flotation. Otherwise the transient solution seems to work if I comment out that line

[0]PETSC ERROR: ------------------------------------------------------------------------
[0]PETSC ERROR: Caught signal number 11 SEGV: Segmentation Violation, probably memory access out of range
[0]PETSC ERROR: Try option -start_in_debugger or -on_error_attach_debugger
[0]PETSC ERROR: or see http://www.mcs.anl.gov/petsc/documentation/faq.html#valgrind
[0]PETSC ERROR: or try http://valgrind.org/ on GNU/linux and Apple Mac OS X to find memory corruption errors
[0]PETSC ERROR: configure using --with-debugging=yes, recompile, link, and run
[0]PETSC ERROR: to get more information on the crash.
[1]PETSC ERROR: ------------------------------------------------------------------------
[1]PETSC ERROR: Caught signal number 11 SEGV: Segmentation Violation, probably memory access out of range
[1]PETSC ERROR: Try option -start_in_debugger or -on_error_attach_debugger
[1]PETSC ERROR: or see http://www.mcs.anl.gov/petsc/documentation/faq.html#valgrind
[1]PETSC ERROR: or try http://valgrind.org/ on GNU/linux and Apple Mac OS X to find memory corruption errors
[1]PETSC ERROR: configure using --with-debugging=yes, recompile, link, and run
[1]PETSC ERROR: to get more information on the crash.
[2]PETSC ERROR: ------------------------------------------------------------------------
[2]PETSC ERROR: Caught signal number 11 SEGV: Segmentation Violation, probably memory access out of range
[2]PETSC ERROR: Try option -start_in_debugger or -on_error_attach_debugger
[2]PETSC ERROR: or see http://www.mcs.anl.gov/petsc/documentation/faq.html#valgrind
[2]PETSC ERROR: or try http://valgrind.org/ on GNU/linux and Apple Mac OS X to find memory corruption errors
[2]PETSC ERROR: configure using --with-debugging=yes, recompile, link, and run
[2]PETSC ERROR: to get more information on the crash.
[0]PETSC ERROR: --------------------- Error Message --------------------------------------------------------------
[0]PETSC ERROR: Signal received
[0]PETSC ERROR: See http://www.mcs.anl.gov/petsc/documentation/faq.html for trouble shooting.
[0]PETSC ERROR: Petsc Release Version 3.7.6, Apr, 24, 2017
[0]PETSC ERROR: [1]PETSC ERROR: --------------------- Error Message --------------------------------------------------------------
[1]PETSC ERROR: Signal received
[1]PETSC ERROR: See http://www.mcs.anl.gov/petsc/documentation/faq.html for trouble shooting.
[1]PETSC ERROR: Petsc Release Version 3.7.6, Apr, 24, 2017
[1]PETSC ERROR: [2]PETSC ERROR: --------------------- Error Message --------------------------------------------------------------
[2]PETSC ERROR: Signal received
[2]PETSC ERROR: See http://www.mcs.anl.gov/petsc/documentation/faq.html for trouble shooting.
[2]PETSC ERROR: Petsc Release Version 3.7.6, Apr, 24, 2017
[2]PETSC ERROR: /Users/lynnkaluzienski/Documents/MATLAB/ISSM/bin/issm.exe on a arch-darwin-c-opt named Lynns-MacBook-Pro.local by lynnkaluzienski Mon Aug 10 16:29:06 2020
[2]PETSC ERROR: /Users/lynnkaluzienski/Documents/MATLAB/ISSM/bin/issm.exe on a arch-darwin-c-opt named Lynns-MacBook-Pro.local by lynnkaluzienski Mon Aug 10 16:29:06 2020
[0]PETSC ERROR: Configure options --prefix=/Users/jenkins/workspace/macOS_binaries/externalpackages/petsc/install --with-single-library=1 --PETSC_DIR=/Users/jenkins/workspace/macOS_binaries/externalpackages/petsc/src --with-shared-libraries=0 --with-debugging=0 --with-mpi-dir=/Users/jenkins/workspace/macOS_binaries/externalpackages/mpich/install --download-fblaslapack=1 --download-metis=1 --download-parmetis=1 --with-x=0 --download-scalapack=1 --download-mumps=1 --with-ssl=0 --with-valgrind=0 --with-pic=1 --COPTFLAGS=-fPIC --CXXOPTFLAGS=-fPIC --FOPTFLAGS=-fPIC
[0]PETSC ERROR: #1 User provided function() line 0 in unknown file
/Users/lynnkaluzienski/Documents/MATLAB/ISSM/bin/issm.exe on a arch-darwin-c-opt named Lynns-MacBook-Pro.local by lynnkaluzienski Mon Aug 10 16:29:06 2020
[1]PETSC ERROR: Configure options --prefix=/Users/jenkins/workspace/macOS_binaries/externalpackages/petsc/install --with-single-library=1 --PETSC_DIR=/Users/jenkins/workspace/macOS_binaries/externalpackages/petsc/src --with-shared-libraries=0 --with-debugging=0 --with-mpi-dir=/Users/jenkins/workspace/macOS_binaries/externalpackages/mpich/install --download-fblaslapack=1 --download-metis=1 --download-parmetis=1 --with-x=0 --download-scalapack=1 --download-mumps=1 --with-ssl=0 --with-valgrind=0 --with-pic=1 --COPTFLAGS=-fPIC --CXXOPTFLAGS=-fPIC --FOPTFLAGS=-fPIC
[1]PETSC ERROR: #1 User provided function() line 0 in unknown file
Configure options --prefix=/Users/jenkins/workspace/macOS_binaries/externalpackages/petsc/install --with-single-library=1 --PETSC_DIR=/Users/jenkins/workspace/macOS_binaries/externalpackages/petsc/src --with-shared-libraries=0 --with-debugging=0 --with-mpi-dir=/Users/jenkins/workspace/macOS_binaries/externalpackages/mpich/install --download-fblaslapack=1 --download-metis=1 --download-parmetis=1 --with-x=0 --download-scalapack=1 --download-mumps=1 --with-ssl=0 --with-valgrind=0 --with-pic=1 --COPTFLAGS=-fPIC --CXXOPTFLAGS=-fPIC --FOPTFLAGS=-fPIC
[2]PETSC ERROR: #1 User provided function() line 0 in unknown file
application called MPI_Abort(MPI_COMM_WORLD, 59) - process 2
application called MPI_Abort(MPI_COMM_WORLD, 59) - process 0
application called MPI_Abort(MPI_COMM_WORLD, 59) - process 1
Error using loadresultsfromdisk (line 17)

============================================================
Binary file Ross.test.outbin not found

This typically happens when the run crashed.
Please check for error messages above or in the outlog

Error in loadresultsfromcluster (line 38)
md=loadresultsfromdisk(md,[md.miscellaneous.name '.outbin']);

Error in solve (line 170)
md=loadresultsfromcluster(md);

Error in runme_annotated_Bedmachine_8_19_19 (line 968)
md=solve(md,'Transient');

---

mathieumorlighem
Hi lkaluzi, could you send me your model that crashed when you get a chance? (save model md, and then send it to me through dropbox or GoogleDrive). I would like to investigate the cause of the crash. Even though it says "PETSc", it is not petsc. Also, I just finished implementing the non uniform error threshold. That will be pushed to the public branch within a couple of weeks hopefully.
Thanks
Mathieu

---

wadewei
Hi Mathieu,

I encountered the same problem with 4.21 using one of my previous model scripts. Interestingly it worked fine with ISSM4.17+PETsc3.12 on my previous cluster.

Cheers,
Wade

iteration 1/20 time [yr]: 0.05 (time step: 0.05)
updating boundary conditions...
computing enthalpy
** PETSc DEPRECATION WARNING ** : the option -pc_factor_mat_solver_package is deprecated as of version 3.9 and will be removed in a future release. Please use the option -pc_factor_mat_solver_type instead. (Silence this warning with -options_suppress_deprecated_warnings)

solver residue: norm(KU-F)/norm(F)=3.39587e-16
mechanical equilibrium convergence criterion 0.890683 < 5 %
Convergence criterion: norm(du)/norm(u) 1.94976 > 1 %
Convergence criterion: max(du) 38858.7
number of unstable constraints: 0

solver residue: norm(KU-F)/norm(F)=3.56696e-16
mechanical equilibrium convergence criterion 7.06241e-08 < 5 %
Convergence criterion: norm(du)/norm(u) 1.69643e-06 < 1 %
Convergence criterion: max(du) 0.48986
number of unstable constraints: 0

total number of iterations: 2
depth averaging WaterfractionDrainage
extruding WaterfractionDrainageIntegrated from base...
extruding BasalforcingsGroundediceMeltingRate from base...
computing smb
call Delta18opdParameterization module
call positive degree day module
computing new velocity
[20]PETSC ERROR: ------------------------------------------------------------------------
[20]PETSC ERROR: Caught signal number 11 SEGV: Segmentation Violation, probably memory access out of range
[20]PETSC ERROR: Try option -start_in_debugger or -on_error_attach_debugger
[20]PETSC ERROR: or see https://petsc.org/release/faq/#valgrind
[20]PETSC ERROR: or try http://valgrind.org/ on GNU/linux and Apple MacOS to find memory corruption errors
[20]PETSC ERROR: configure using --with-debugging=yes, recompile, link, and run
[20]PETSC ERROR: to get more information on the crash.
[20]PETSC ERROR: [17]PETSC ERROR: ------------------------------------------------------------------------
[17]PETSC ERROR: Caught signal number 11 SEGV: Segmentation Violation, probably memory access out of range
[17]PETSC ERROR: Try option -start_in_debugger or -on_error_attach_debugger
[17]PETSC ERROR: or see https://petsc.org/release/faq/#valgrind
[17]PETSC ERROR: or try http://valgrind.org/ on GNU/linux and Apple MacOS to find memory corruption errors

---

mathieumorlighem
wadewei This is a bad segmentation fault, it says "PETSc" but it is not (it is just that PETSc has a wrapper to catch these errors). Send me your model ready to run (i.e. I want to load it, change the cluster, and run md=solve(md,'tr');) and I will investigate.
Mathieu

---

wadewei
mathieumorlighem
I've emailed you the model.