neuronsimulator / testcorenrn

Description

This repository includes tests related to CoreNEURON and NEURON intergration.

More information about NEURON and CoreNEURON can be found here:

• NEURON
• CoreNEURON

Compatibility

• This version of the repository is compatible with NEURON versions newer than 8.1 (#10)

How to install NEURON with CoreNEURON

More information and details how to do this you can find in the above links. A typical way to install them to run the tests is the following:

git clone https://github.com/neuronsimulator/nrn.git
cd nrn
mkdir build && cd build
cmake .. -DCMAKE_INSTALL_PREFIX=./install -DNRN_ENABLE_CORENEURON=ON
make -j8
make install

For GPU execution set the following CMake variable:

-DCORENRN_ENABLE_GPU=ON

Note: For GPU execution make sure that you use a compiler that supports OpenACC compilation and CUDA is also available in your system. See more details here.

How to run the tests

All the tests can be run with NEURON or CoreNEURON and with multiple configurations depending on the provided options to special or special-core. The available options are described bellow:

• coreneuron: Enables the execution of the test with CoreNEURON using the on-line mode
• gpu: Enables the execution of the test with CoreNEURON on-line mode on GPU.

Note: for this option you need to make sure that NEURON is installed with the CMake option CORENRN_ENABLE_GPU enabled.

• filemode: Runs CoreNEURON from NEURON in off-line mode. NEURON automatically generates the CoreNEURON dataset and runs the simulation with CoreNEURON. To use this coreneuron option must be enabled.
• dumpmodel: Creates the CoreNEURON dataset. This enables CoreNEURON to run separately after the initial simulation.

Example

On-line mode

The following commands runs one of the included test with CoreNEURON and with the GPU backend enabled.

nrnivmodl -coreneuron mod
mpirun -n 2 ./x86_64/special -mpi -c sim_time=100 -c coreneuron=1 -c gpu=1 testkin.hoc

Off-line mode

The following commands runs one of the included tests with NEURON, generates the CoreNEURON dataset, runs the CoreNEURON simulation and compares the spikes between the two.

nrnivmodl -coreneuron mod
mpirun -n 2 ./x86_64/special -mpi -c sim_time=100 -c dumpmodel=1 -c gpu=1 testkin.hoc
cat outkin.dat | sort -k 1n,1n -k 2n,2n > out_nrn_kin.spk
mpirun -n 2 ./x86_64/special-core --mpi --tstop 100 --datpath coredat
cat out.dat | sort -k 1n,1n -k 2n,2n > out_cn_kin.spk
diff -w -q out_nrn_kin.spk out_cn_kin.spk

Comparison

To make sure that the NEURON and CoreNEURON used to run the tests are running properly it's also needed to compare their generated spikes with the ones existing in the reference folder for each test.

How to run all the tests automatically

To run all the tests you can use the run.sh file which runs all the tests included in this repository with NEURON and CoreNEURON using some default parameters and then compares them with the reference spikes. You can select which tests to run by setting the spike_comparison_tests, direct_tests and gpu_tests.

Note: Make sure that you compiled the included mod files in the mod directory before running the script. To do this you need to run nrnivmodl -coreneuron mod.

There are 3 kinds of tests:

• spike_comparison_tests: Those tests are used only to compare the generated spikes from NEURON and CoreNEURON to the reference spikes
  • bbcore
  • conc
  • deriv
  • gf
  • kin
  • patstim
  • vecplay
  • watch
  • vecevent
• direct_tests: Those test compare internally the voltage and currents between NEURON and CoreNEURON and generate a spike output file for comparison with the reference spikes
  • netstimdirect
• gpu_tests: Same as spike_comparison_tests but if set they compare the spikes generated by NEURON and CoreNEURON using the GPU backend with the reference spikes.