Search found 41 matches

by sgratiy
Mon Jul 25, 2016 5:26 pm
Forum: Parallel NEURON
Topic: NEURON on GPUs?
Replies: 3
Views: 5791

Re: NEURON on GPUs?

@ Michael
Could you clarify, what do you mean by
large memory bandwidth limited models
?
by sgratiy
Thu Jun 18, 2015 3:37 pm
Forum: NEURON + Python
Topic: Accessing section type (or segment type) information
Replies: 0
Views: 828

Accessing section type (or segment type) information

When I instantiating a morphology in NEURON, is there a way of getting section types (or segment types) in the same format as in the swc file? Basically, I would like to be able to have one of the following integer values for a give section/segment: type of neuronal compartment: 0 - undefined 1 - so...
by sgratiy
Thu Oct 02, 2014 4:23 pm
Forum: NEURON + Python
Topic: Accessing synapses on a section
Replies: 3
Views: 1870

Re: Accessing synapses on a section

Thanks for pointing out the get_loc() method. I followed your recommendation and it works.
by sgratiy
Wed Oct 01, 2014 8:59 pm
Forum: NEURON + Python
Topic: Accessing synapses on a section
Replies: 3
Views: 1870

Accessing synapses on a section

Is there a way of getting a list of synapses existing on a section? It is assumed that synapse were created, of course. I wanted to calculate x,y,z coordinates of synapses for visulaization purposes. So, I would like to be able to loop through sections in a particular cell and for each section get l...
by sgratiy
Sun Jun 22, 2014 1:53 pm
Forum: MSWin
Topic: Win installation does not recognize packages (numpy)
Replies: 1
Views: 5629

Win installation does not recognize packages (numpy)

I was trying to install NEURON and run it with python as an interpreter on my laptop just before the parallel course. However, my windos 32 installation is having issues as it refuses to recognize numpy and gives error message "ImportError: No module named testing". Please see below the output from ...
by sgratiy
Mon Jun 16, 2014 4:09 pm
Forum: NEURON + Python
Topic: Creating an array of membrane currents
Replies: 8
Views: 3663

Re: Creating an array of membrane currents

Thanks for clarification. I realize that this is what is happening. My confusion is with why this happens given that objects in python are passed by assignment not by value. So if the element of an array (or a list) were assigned to objects once, then the updates to the values stored in those object...
by sgratiy
Mon Jun 16, 2014 2:34 am
Forum: NEURON + Python
Topic: Creating an array of membrane currents
Replies: 8
Views: 3663

Re: Creating an array of membrane currents

Great, it works now. I would have never figured that out on my own. Could you please tell whether h.cvode.cache_efficient(1) works when using a constant time step? I am still puzzled with the conceptual issue of why simply populating a numpy array or (a python list) with segment currents as shown be...
by sgratiy
Thu Jun 12, 2014 3:45 pm
Forum: NEURON + Python
Topic: Creating an array of membrane currents
Replies: 8
Views: 3663

Re: Creating an array of membrane currents

I do have self.pVec initialized in the costructor and I ruturn numpy array as you recommended.
by sgratiy
Wed Jun 11, 2014 8:21 pm
Forum: NEURON + Python
Topic: Creating an array of membrane currents
Replies: 8
Views: 3663

Re: Creating an array of membrane currents

Thanks Michael, however I am still not able to make it work. Below is the code for my Cell class which basically only has functions for setting and getting i_membrane: getMemCurArrSlow() is a slow implementation where I loop through the tree at each time step. Following your suggestion I used the Pt...
by sgratiy
Tue Jun 10, 2014 1:14 pm
Forum: NEURON + Python
Topic: Creating an array of membrane currents
Replies: 8
Views: 3663

Creating an array of membrane currents

I would like to efficiently access membrane currents for each segment in the cell at each time step to multiply them by some matrix. Currently I populate a numpy array on each time step before performing matrix multiplication: jseg = 0 # iseg for the entire cell for sec in cell.hoc.all: # for seg in...
by sgratiy
Thu May 22, 2014 2:20 pm
Forum: Parallel NEURON
Topic: Saving vector at each time step - not associated with cell
Replies: 16
Views: 14740

Re: Saving vector at each time step - not associated with ce

I prefer to modify the proc advance() rather than using pybs.mod. Now I get the idea for accesing Python object from inside the hoc. However the code: h(''' objref po po = new PythonObject() proc advance() { fadvance() po.mypopulation.calc_csd() } does not seem to work because during the execution I...
by sgratiy
Wed May 21, 2014 3:03 pm
Forum: Parallel NEURON
Topic: Saving vector at each time step - not associated with cell
Replies: 16
Views: 14740

Re: Saving vector at each time step - not associated with ce

Thanks for the detailed response. Certainly I got confused in distinction between Python and the Hoc. I do programming in Python and access the Hoc with neuron import h as you mentioned. It still remains a mistery to me how to modify proc advance() from Python in order to insert in it the call to Py...
by sgratiy
Tue May 20, 2014 2:21 pm
Forum: Parallel NEURON
Topic: Saving vector at each time step - not associated with cell
Replies: 16
Views: 14740

Re: Saving vector at each time step - not associated with ce

Of course, I still need to call calc_csd, thanks for pointing that out. I did this in my serial implementation in the past, but got confused that with Vector.record it is not necessary because calling the member function csd_calc() once in the constuctor would reference self.csd to self.W*self.i_mem...
by sgratiy
Mon May 19, 2014 5:53 pm
Forum: Parallel NEURON
Topic: Saving vector at each time step - not associated with cell
Replies: 16
Views: 14740

Re: Saving vector at each time step - not associated with ce

Mike, thanks for pointing out the bug. I am glad I am on the right track, however, I got confused with your statement: Forget that suggestion and keep on with your calc_csd and modified proc advance(). I thougth that I do not need to modify the proc_advance() to store the csd vector since I can use ...
by sgratiy
Mon May 19, 2014 3:25 pm
Forum: Parallel NEURON
Topic: Saving vector at each time step - not associated with cell
Replies: 16
Views: 14740

Re: Saving vector at each time step - not associated with ce

Thanks for the detailed response. However, it is still unclear to me whether I must use the mod file to save my csd without modifying the proc advance() or I can get away withot inroducing a mod file.( I am reluctant to use mod file because it will be cell mechanism, whereas CSD is not a cell mechan...