Simulation with svSolver finishing before reaching specified number of timesteps

[Patryk Rygiel]

Hi,

I'm running pulsatile simulation for 4 heart cycles with svSolver with following time parameters - 500 steps per one cycle:

Code: Select all

Number of timesteps: 2000
Time Step Size: 0.002
Number of Timesteps between Restarts: 25

I see that the simulation ends at the timestep 1938:

Code: Select all

  
  1937 2.934E+04 4.498E-02  (  -9)   2.106E-02   1.534E-05  <  8112-15|  13> [   0 -   0]
  1937 2.934E+04 3.317E-03  ( -20)   2.234E-03   8.964E-06  <  8112-15|  16> [   0 -   0]
  1937 2.935E+04 3.263E-04  ( -30)   4.344E-03   3.426E-02  < 11088-31|  15> [   0 -   0]
  1938 2.935E+04 4.408E-02  (  -9)   2.042E-02   2.909E-04  <  7304-18|  10> [   0 -   0]
  1938 2.935E+04 3.077E-03  ( -21)   2.256E-03   1.503E-05  <  8112-15|  12> [   0 -   0]
 

I would like the simulation to be resolved for all 2000 steps. I'm wondering what could be the cause of the simulation finishing early - I don't see any error being thrown or logged. Could be so that there is some trigger terminating the simulation upon reaching some convergence criterion?

Best,
Patryk

[David Parker]

Hi Patryk,

What OS and SV version are you using. Are you running the solver from SV or on an HPC cluster?

Cheers,
Dave

[Patryk Rygiel]

Hi Dave,

Hereby the version of software I use:
OS: Ubuntu 20.04.6
svSolver: 2021-09-30 release

I'm running solver via terminal with following command on a single computing node with 64 cores (using 32 with MPI):

Code: Select all

mpiexec -np 32 "$sv_path/svsolver"

Best,
Patryk

[David Parker]

Hi Patryk,

I don't know what might be going on here, the history file residuals look fine.

Please email me at davep@stanford.edu the solver input files (no result files) and I'll have a look.

Cheers,
Dave

[Ignacio Bartol]

Hey! Got into a similar problem today using 48 cores. I was trying to asses how well was scaling the computing time with the number of cores and while trying 48 after ~300 iterations the simulation stopped while in total there were 6000 iterations. It just happened that I clean the simulation directory, ran the simulation again (was not too long) and it worked fine.