You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I am currently attempting to re-run continental rifting models that were previously run with older versions of ASPECT (v2.4.0-pre, commit 70a2ec3) and deal.II (v9.3.3) with newer versions (ASPECT v2.6.0-pre, commit 845dd6d; deal.II v9.5.1) in order to take advantage of recently-added Fastscape coupling. I am finding that a model that runs successfully on the same HPC system using the old versions fails when using the newer versions with the following particle deregistration error:
terminate called after throwing an instance of 'dealii::StandardExceptions::ExcMessage'
what():
--------------------------------------------------------
An error occurred in line <62> of file </cluster/home/dvasey01/vaseylab/shared/software/dealii/v9.5.1/candi/install/tmp/unpack/deal.II-v9.5.1/source/particles/property_pool.cc> in function
void dealii::Particles::PropertyPool<dim, spacedim>::clear() [with int dim = 2; int spacedim = 2]
The violated condition was:
n_open_handles == 0
Additional information:
This property pool currently still holds 1 open handles to memory that
was allocated via allocate_properties_array() but that has not been
returned via deregister_particle().
Stacktrace:
-----------
#0 /cluster/home/dvasey01/vaseylab/shared/software/dealii/v9.5.1/candi/install/deal.II-v9.5.1/lib/libdeal_II.so.9.5.1: dealii::Particles::PropertyPool<2, 2>::clear()
#1 /cluster/home/dvasey01/vaseylab/shared/software/dealii/v9.5.1/candi/install/deal.II-v9.5.1/lib/libdeal_II.so.9.5.1: dealii::Particles::ParticleHandler<2, 2>::clear_particles()
#2 /cluster/home/dvasey01/vaseylab/shared/software/dealii/v9.5.1/candi/install/deal.II-v9.5.1/lib/libdeal_II.so.9.5.1: dealii::Particles::ParticleHandler<2, 2>::~ParticleHandler()
#3 /cluster/home/dvasey01/vaseylab/shared/software/aspect/v2.6.0-pre-845dd6d/aspect/build/./aspect: aspect::Particle::World<2>::~World()
#4 /cluster/home/dvasey01/vaseylab/shared/software/aspect/v2.6.0-pre-845dd6d/aspect/build/./aspect: aspect::Particle::World<2>::~World()
#5 /cluster/home/dvasey01/vaseylab/shared/software/aspect/v2.6.0-pre-845dd6d/aspect/build/./aspect: aspect::Simulator<2>::~Simulator()
#6 /cluster/home/dvasey01/vaseylab/shared/software/aspect/v2.6.0-pre-845dd6d/aspect/build/./aspect: void run_simulator<2>(std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, bool, bool, bool)
#7 /cluster/home/dvasey01/vaseylab/shared/software/aspect/v2.6.0-pre-845dd6d/aspect/build/./aspect: main
--------------------------------------------------------
SIGABRT received
This error occurs after the model has been running on a single node for a couple hours (after ~3 Myr of model time), so it does not initiate immediately. Since the same model runs successfully on the same cluster with the older ASPECT/deal.II versions, it does not appear to be an issue (at first glance) with the system. @naliboff and I discussed this briefly and he has not seen this error before, so any insight into what is going on would be helpful. Perhaps I should be using a different version of deal.II?
The text was updated successfully, but these errors were encountered:
Hi Dylan, can you check how much your problem overlaps with the one discussed in #5594? They may be related or the same, in which case either one of the following options should fix the crash:
running the model in release mode should work just fine (the problem is an actual bug, but it doesnt affect the solution of the model)
using a recent development version of deal.II should fix it in both debug and release mode
Hi Rene, thanks for pointing me to that issue - the error certainly looks very similar, and the version of deal.II I am using predates the pull request you put in to address this (dealii/dealii#16709). I will try compiling a recent deal.II and see if that fixes the problem. This error is actually occurring in release mode (I haven't tried to reproduce it in debug mode yet). Thanks!
It looks like the pull request mentioned above fixes the issue. I was unable to easily compile a development version of deal.II (I am using recent Intel compilers but got an error about GNU compilers being out of date that I don't fully understand, but that is a separate issue), but I was able to install v9.5.1 with dealii/dealii@9d90510 cherry-picked. Thanks for the help!
I am currently attempting to re-run continental rifting models that were previously run with older versions of ASPECT (v2.4.0-pre, commit 70a2ec3) and deal.II (v9.3.3) with newer versions (ASPECT v2.6.0-pre, commit 845dd6d; deal.II v9.5.1) in order to take advantage of recently-added Fastscape coupling. I am finding that a model that runs successfully on the same HPC system using the old versions fails when using the newer versions with the following particle deregistration error:
This error occurs after the model has been running on a single node for a couple hours (after ~3 Myr of model time), so it does not initiate immediately. Since the same model runs successfully on the same cluster with the older ASPECT/deal.II versions, it does not appear to be an issue (at first glance) with the system. @naliboff and I discussed this briefly and he has not seen this error before, so any insight into what is going on would be helpful. Perhaps I should be using a different version of deal.II?
The text was updated successfully, but these errors were encountered: