Field ionization + collisions

[vbrsi]

I'd like to ionize a H target with a laser pulse. Depending on the density and the thickness of the target only a small part of it will be ionized. In this layer I would have e-, p, and then H neutral.
I have different questions:

• Is it possible to include both collisions, ionizing collisions and field ionization?
• How do I keep track of p+ ? Using the same block for p as it is done for el does not work: the diagnostics remain empty.
• Once ionization takes place, what is the meaning of the H-related diagnostics? Do they track the H neutral, H+, the partially ionized H, both.. ?
• As far as I understand from a previous discussion I can not use neutrals in the collision block, is that correct? However, the sim. runs if I add intra-species collisions H-H (to be cleared what it means depending on what H 'becomes' after partial ionization) or e-e, but it does not run for floating point exception including e-H collisions.

Thanks for helping!

[mccoys]

1. Yes you can use modules together
2. Please be more specific about the diagnostics you are using. There is no a priori distinction between ions of different charges. You have to isolate them but this depends on the diagnostic
3. Electron collisional ionization with neutrals should be ok. What is not available now is the correct collision cross section with neutrals. It should be available in a few weeks though, but does not impact ionization

[vbrsi]

Same:

The following have been reloaded with a version change:

1. GCCcore/12.2.0 => GCCcore/13.3.0
2. UCX/1.13.1-GCCcore-12.2.0 => UCX/1.16.0-GCCcore-13.3.0
3. numactl/2.0.16-GCCcore-12.2.0 => numactl/2.0.18-GCCcore-13.3.0
4. zlib/1.2.12-GCCcore-12.2.0 => zlib/1.3.1-GCCcore-13.3.0

Shell debugging restarted

• unset __lmod_sh_dbg
• return 0
• total_mpi_processes=4
• echo .
• srun -n 4 -c 16 /Smilei_dev/smilei input_test.py
  [cn0245:3226922:0:3227018] Caught signal 11 (Segmentation fault: address not mapped to object at address 0x1c3b30d38)
  [cn0245:3226923:0:3227019] Caught signal 11 (Segmentation fault: address not mapped to object at address 0x94c11c0)
  [cn0235:3912651:0:3913092] Caught signal 11 (Segmentation fault: address not mapped to object at address 0x2b61ca4c)
  [cn0235:3912650:0:3913109] Caught signal 11 (Segmentation fault: address not mapped to object at address 0xef17160)
  ==== backtrace (tid:3227018) ====
  0 0x0000000000012d10 __funlockfile() :0
  1 0x00000000009886fc std::vector<int, std::allocator >::operator /cvmfs/sling.si/modules/el7/software/GCCcore/12.2.0/include/c++/12.2.0/bits/stl_vector.h:1124
  2 0x0000000000897ea9 SyncVectorPatch::finalizeExchParticlesAndSort() /Smilei_dev/src/Patch/SyncVectorPatch.cpp:57
  3 0x000000000082a938 VectorPatch::finalizeExchParticlesAndSort() /Smilei_dev/src/Patch/VectorPatch.cpp:448
  4 0x0000000000962614 main._omp_fn.4() /Smilei_dev/src/Smilei.cpp:633
  5 0x000000000002138e gomp_thread_start() /cvmfs/sling.si/modules/el7/build/GCCcore/13.3.0/system-system/gcc-13.3.0/stage3_obj/x86_64-pc-linux-gnu/libgomp/../../../libgomp/team.c:129
  6 0x00000000000081ca start_thread() ???:0
  7 0x00000000000398d3 __GI___clone() :0
  =================================
  Stack trace (most recent call last) in thread 3227018:
  add new temporal profile tsin2plateau #7 Object "[0xffffffffffffffff]", at 0xffffffffffffffff, in
  Params.cpp - typo lxmax -> least #6 Object "/lib64/libc.so.6", at 0x7f9c157dc8d2, in clone
  ExtField incorrect example #5 Object "/lib64/libpthread.so.0", at 0x7f9c163671c9, in
  Odd explosion - energy conservation problem #4 Object "/cvmfs/sling.si/modules/el7/software/GCCcore/13.3.0/lib64/libgomp.so.1", at 0x7f9c1726e38d, in
  Conversion of outputs to OpenPMD #3 Object "/Smilei_dev/smilei", at 0x962613, in
  abs(coeff2*qqm*term3*term5) #2 Object "/Smilei_dev/smilei", at 0x82a937, in VectorPatch::finalizeExchParticlesAndSort(Params&, SmileiMPI*, SimWindow*, double, Timers&, int)
  Compile error #1 Object "/Smilei_dev/smilei", at 0x897ea8, in SyncVectorPatch::finalizeExchParticlesAndSort(VectorPatch&, int, Params&, SmileiMPI*)
  #0 Object "/Smilei_dev/smilei", at 0x9886fc, in SpeciesV::sortParticles(Params&)
  Segmentation fault (Signal sent by tkill() [0x75df135b00313d2a])
  srun: error: cn0245: task 2: Exited with exit code 1
  slurmstepd: error: mpi/pmix_v3: _errhandler: cn0245 [1]: pmixp_client_v2.c:212: Error handler invoked: status = -25, source = [slurm.pmix.54246201.0:2]
  slurmstepd: error: *** STEP 54246201.0 ON cn0235 CANCELLED AT 2025-03-06T11:11:20 ***
  srun: Job step aborted: Waiting up to 32 seconds for job step to finish.
  srun: error: cn0245: task 3: Killed
  srun: error: cn0235: tasks 0-1: Killed

Looks like a problem with some openmpi libraries but it crashes also with mpirun -np 1.
Smilei version is : 5.1-126-gd77bad2e6-develop

[vbrsi]

I am trying on tst1d_05_tunnel_ionisation.py adding collisions between electron and hydrogen

[vbrsi]

The segmentation fault seems to happen here:
vecPatches(ipatch)->importAndSortParticles(ispec, params);

in SyncVectorPatch (line 57 I think):

I added if (vecPatches(ipatch) == nullptr) {
std::cerr << "ERROR: vecPatches(" << ipatch << ") is null!" << std::endl;
continue;
}

    std::cerr << "Accessing vecPatches(" << ipatch << ")..." << std::endl;)

And this is the output in err:

Accessing vecPatches(0)...
[cn0730:190704:0:190731] Caught signal 11 (Segmentation fault: address not mapped to object at address 0x10fa03b0)

[mccoys]

It really means that there is a segmentation fault somewhere in the sorting of particles. It is hard to find out what is the problem here. In a few days a new version should bring a bunch of modifications in collisions and we will see if the error persists

[vbrsi]

Ok, thank you!!