www.neuron.yale.edu

I'm currently searching about how to use NEURON with GPU machine.
From my understanding, the Parallel Network Manager allow you to paralleling your NEURON processing on multi-core CPU through MPI.
The technique is good but it could be better to use NEURON with GPU for large network simulation.
If such support is available, would you please suggest me where to find tutorials and further information about NEURON on GPU.
I'm the first member of the lab to use NEURON. Therefore, I've no one to help me about this matter.I'm sorry to ask some simple questions.
Thank you in advance!

The NEURON+GPU project is under development. At present the only thing ready for public use is
http://bitbucket.org/nrnhines/nrngpu
which computes gating states and currents from nmodl channel descriptions on the gpu. The G*V=I matrix
is still solved on the cpu and all network events are handled on the cpu. A research version which is not yet public
does the G*V=I gaussian elimination part on the gpu as well. But we still have not ported event handling and
NET_RECEIVE block computations to the gpu yet. Building for the gpu requires an --enable-cuda configure argument.
If you decide to experiment with the nrngpu version and have questions it is best to correspond via email.
michael dot hines at yale dot edu

Thank you very much for your prompt response!

Updating this old post with the latest information (for future references): If you have questions related to GPU support in NEURON, I would suggest looking at the documentation of the latest release, e.g. https://nrn.readthedocs.io/en/latest/de ... sting.html.