CoreNeuron
CoreNeuron copied to clipboard
BlueBrain
Simulator optimized for large scale neural network simulations.
**Description** Added asynchronous execution of kernels in multiple GPU streams. Still there are some constructs that the compiler doesn't support: 1. `#pragma omp target update to() depend(inout: stream) nowait` is...
Various changes (including temporary) to make XL OpenMP offload build working * todo: need to rebase this branch on latest hackathon_master * todo: temporary changes to OpenAccHelper.cmake, needs refinement *...
Some mpiexec version does not have the option -n, so it should be configurable
**Motivation** As a follow up to #252, we want CoreNeuron to be able to create checkpoints right before an allocation expires. Since most job schedulers send a SIGTERM before sigkill,...
**Describe the issue** If CoreNEURON is configured with `CORENRN_ENABLE_SPLAYTREE_QUEUING=OFF`, the following NEURON tests segfault: ``` 46 - coreneuron_modtests::test_netmove_py_cpu (Failed) 54 - coreneuron_modtests::inputpresyn_py_cpu (Failed) 63 - coreneuron_modtests::test_netmove_py_gpu (Failed) 71 - coreneuron_modtests::inputpresyn_py_gpu...
In https://github.com/BlueBrain/CoreNeuron/pull/834 we will increase the minimum C++ standard to C++17. This should allow TODO items like https://github.com/BlueBrain/CoreNeuron/blob/0574812f772d6dde595f33913452f798bba5166e/coreneuron/utils/memory.cpp#L39-L40 to be addressed. This should be helpful (also in NEURON) for structures...
In https://github.com/BlueBrain/CoreNeuron/pull/834 we will require C++17. This will allow TODOs like https://github.com/BlueBrain/CoreNeuron/blob/0574812f772d6dde595f33913452f798bba5166e/coreneuron/mpi/nrnmpi.h#L71-L73 to be addressed, simplifying the code. There is an [internal reference](https://bbpgitlab.epfl.ch/hpc/personal/coreneuron-dynamic-mpi-sandbox/-/merge_requests/1) to consult too.
**Describe the issue** coreneuron with GPU support does not compile successfully. **To Reproduce** ``` cmake module load profile/archive module load profile/base module load profile/advanced module load pgi/19.10--binary gnu/8.4.0 cuda/10.1 module...
**Describe the issue** If we install OpenMP runtime via `brew` and try to build CoreNEURON with AppleClang then building as well as running tests fail. **To Reproduce** ```bash brew install...
**Describe the issue** In GPU-enabled builds that combine NMODL, GPU offload using OpenACC, and OpenMP host-side threading, the `coreneuron_modtests::datareturn_py_gpu` test hangs. The hang occurs at ```c++ void nrn_init_pas(NrnThread* nt, Memb_list*...