Crash when try an example

Hi,
when I try any of example (for time of writing I try:
-plate with hole
-spanner
-heat transfer)
I get crash on solve with MATLAB error. Crash report on the end.
I work on Debian

uname -a
Linux desktop 4.19.0-6-amd64 #1 SMP Debian 4.19.67-2+deb10u2 (2019-11-11) x86_64 GNU/Linux

If you need I can check something more.
Any idea how to solve this issue?

Have a nice dat
Piotr

--------------------------------------------------------------------------------
           Illegal instruction detected at 2024-01-31 22:39:32 +0100
--------------------------------------------------------------------------------

Configuration:
  Crash Decoding           : Disabled - No sandbox or build area path
  Crash Mode               : continue (default)
  Default Encoding         : UTF-8
  Deployed                 : true
  Desktop Environment      : XFCE
  GNU C Library            : 2.28 stable
  Graphics Driver          : Brian Paul Mesa X11 Version 2.1 Mesa 17.1.3
  Graphics card 1          : 0x1002 ( 0x1002 ) 0x683d Version 5.2.0.19 (0-0-0)
  Java Version             : Java 1.8.0_202-b08 with Oracle Corporation Java HotSpot(TM) 64-Bit Server VM mixed mode
  MATLAB Architecture      : glnxa64
  MATLAB Entitlement ID    : Unknown
  MATLAB Root              : /storage/projekty/FEM/matlab_runtime/R2023a
  MATLAB Version           : 9.14.0.2337262 (R2023a) Update 5
  OpenGL                   : hardware
  Operating System         : Debian GNU/Linux 10 (buster)
  Process ID               : 31074
  Processor ID             : x86 Family 6 Model 45 Stepping 7, GenuineIntel
  Session Key              : afb6733f-fd53-4c59-a6d6-d3f15cec50df
  Window System            : The X.Org Foundation (12004000), display :0.0

Fault Count: 1


Abnormal termination:
Illegal instruction

Current Thread: 'MCR 0 interpret' id 140172598388480

Register State (from fault):
  RAX = 0000000000000001  RBX = 00007f7c79ee7e88
  RCX = 00007f7b03cda000  RDX = 0000000000000002
  RSP = 00007f7c79ee0000  RBP = 0000000000000001
  RSI = 00007f7b03dda060  RDI = 00007f7b03cda080

   R8 = 00007f7b03dda000   R9 = 00007f7b1a96b878
  R10 = 0000000000000090  R11 = 0000000000000004
  R12 = 0000000000000001  R13 = 0000000000000011
  R14 = 0000000000000002  R15 = 00007f7b1a96b758

  RIP = 00007f7b1029a854  EFL = 0000000000010202

   CS = 0033   FS = 0000   GS = 0000

Stack Trace (from fault):
[  0] 0x00007f7b1029a854 /home/sasza/.mcrCache9.14/FEAToo0/FEATool_Mult/build/featool_pkg/app/lib/mumps/dmumpsmex_st.mexa64+02046036


This error was detected while a MEX-file was running. If the MEX-file
is not an official MathWorks function, please examine its source code
for errors. Please consult the External Interfaces Guide for information
on debugging MEX-files.

Hi,

The issue seems to be that the Mumps linear solver mex (Matlab extension) file has crashed. These files are compiled and can sometimes causes crashes if for some reason they are not compatible with the system (possibly your Debian version is older than the system it was built for Ubuntu 20.04 I think).

If you open the "Solver Settings" dialog box and change the default "Linear solver" setting from "mumps" to "backslash" (the built-in in Matlab solver) it will probably work for your models. It might still crash for the automated tutorials if they select mumps anyway, in this case you can go to the folder and rename (or delete) the problematic mex file (in this case "/home/sasza/.mcrCache9.14/FEAToo0/FEATool_Mult/build/featool_pkg/app/lib/mumps/dmumpsmex_st.mexa64"). In case other mex files also have issues (such as fsparse) you can try to rename and delete them as well.

Thank you so much :)
You are right, after changing solver it works.

Can you tell me which libraries in my system should be update to make it work also with mumps?

sasza wrote

Can you tell me which libraries in my system should be update to make it work also with mumps?

Compatibility issues like this are often due to the c library "libc". Unfortunately, as almost everything is compiled and linked against it, it usually can't be updated by itself, only by updating the Linux distribution/version.

I think Mumps in general is only about 5-10% faster than the built-in solver Umfpack (backslash), so it might not be worth going through the trouble of updating your whole system unless you really need all performance.