Skip to content

GitLab

Open
Created by Valeria Rizzoglio@rizzoglio_v

Segmentation fault - Degrader 70 MeV

I am trying to test the influence of the time step on the results of the OPAL Monte Carlo using the Multi-Slabs degrader for 70 MeV (Degrader_70.in)

I set the time step to 1e-10 s and I got a segmentation fault. So I did few tests, trying different configurations of time steps, n. of cores and options (ENABLERUTHERFORD = TRUE/FALSE or with/without GPU)

-- Configuration 1

  • protons = 1e5, DT = 1e-10 s, cores = 4, with dks and ENABLERUTHERFORD = TRUE
  • result: segmentation fault Config1.out

-- Configuration 2

  • protons = 1e5, DT = 1e-10 s, cores = 4, with dks and ENABLERUTHERFORD = FALSE
  • result: OK Config2.out

-- Configuration 3

  • protons = 1e5, DT = 1e-10 s, cores = 4, without dks and ENABLERUTHERFORD = TRUE
  • result: segmentation fault Config3.out

-- Configuration 4

  • protons = 1e7, DT = 1e-10 s, cores = 4, without dks and ENABLERUTHERFORD = TRUE
  • result: OPAL stops at 4.4 mm with 4 protons while the ZSTOP is 4.3 m Config4.out

-- Configuration 4.b

  • protons = 1e5, DT = 1e-10 s, cores = 8, without dks and ENABLERUTHERFORD = TRUE
  • result: segmentation fault Config4b.out

-- Configuration 5 (on Merlin)

  • protons = 1e7, DT = 1e-10 s, cores = 32, without dks and ENABLERUTHERFORD = FALSE
  • result: OK

-- Configuration 6

  • protons = 1e5, DT = 1e-11 s, cores = 4, with dks and ENABLERUTHERFORD = FALSE
  • result: OK Config6.out

-- Configuration 7

  • protons = 1e5, DT = 1e-11 s, cores = 4, with dks and ENABLERUTHERFORD = TRUE
  • result: OK Config7.out

Run on opalrunner and Merlin with these settings:

Currently Loaded Modulefiles:
  1) gcc/5.4.0           4) OPAL/1.6            7) Tcl/8.6.4          10) boost/1.62.0
  2) openmpi/1.10.4      5) root/6.08.02        8) Tk/8.6.4           11) gsl/2.2.1
  3) OPAL/1.6.0rc3       6) openssl/1.0.2j      9) Python/2.7.12      12) H5root/1.3.2rc4-1