#include <Neurons.h>

Inheritance diagram for Neurons:

Public Member Functions
	Neurons ()
virtual int	AddGroup (neuron_parameters_struct *group_params)
virtual void	allocate_device_pointers (int maximum_axonal_delay_in_timesteps, bool high_fidelity_spike_storage)
virtual void	copy_constants_to_device ()
virtual void	reset_neuron_activities ()
void	reset_current_injections ()
void	set_threads_per_block_and_blocks_per_grid (int threads)
Public Attributes
int	total_number_of_neurons
int	total_number_of_groups
int	number_of_neurons_in_new_group
int *	start_neuron_indices_for_each_group
int *	last_neuron_indices_for_each_group
int *	per_neuron_afferent_synapse_count
int **	group_shapes
int *	d_per_neuron_afferent_synapse_count
float *	d_current_injections
dim3	number_of_neuron_blocks_per_grid
dim3	threads_per_block

Detailed Description

This is the parent class for SpikingNeurons. It provides a set of default methods which are primarily used to add groups of neurons to the total population.

Constructor & Destructor Documentation

Neurons::Neurons ( )

Initializes pointers and variables.

Member Function Documentation

virtual int Neurons::AddGroup ( neuron_parameters_struct * group_params ) [virtual]

Determines the total number of neurons by which the simulation should increase. This is a virtual function to allow polymorphism in the methods of various SpikingNeuron implementations. Allocates memory as necessary for group size and indices storage.

Parameters:

group_params A neuron_parameters_struct instance describing a 2D neuron population size.

Returns:: The unique ID for the population which was requested for creation.

Reimplemented in InputSpikingNeurons, AdExSpikingNeurons, SpikingNeurons, GeneratorInputSpikingNeurons, IzhikevichSpikingNeurons, LIFSpikingNeurons, PoissonInputSpikingNeurons, and ImagePoissonInputSpikingNeurons.

virtual void Neurons::allocate_device_pointers	(	int	maximum_axonal_delay_in_timesteps,
		bool	high_fidelity_spike_storage
	)		`[virtual]`

Exclusively for the allocation of device memory. This class requires allocation of d_current_injections only.

Parameters:

maximum_axonal_delay_in_timesteps	The length (in timesteps) of the largest axonal delay in the simulation. Unused in this class.
high_fidelity_spike_storage	A flag determining whether a bit mask based method is used to store spike times of neurons (ensure no spike transmission failure).

Reimplemented in AdExSpikingNeurons, SpikingNeurons, GeneratorInputSpikingNeurons, IzhikevichSpikingNeurons, LIFSpikingNeurons, and PoissonInputSpikingNeurons.

virtual void Neurons::copy_constants_to_device ( ) [virtual]

Unused in this class. Allows copying of static data related to neuron dynamics to the device.

Reimplemented in AdExSpikingNeurons, SpikingNeurons, IzhikevichSpikingNeurons, LIFSpikingNeurons, and PoissonInputSpikingNeurons.

void Neurons::reset_current_injections ( )

A local, non-polymorphic function called by Neurons::reset_neuron_activities to reset Neurons::d_current_injections.

virtual void Neurons::reset_neuron_activities ( ) [virtual]

Resets any undesired device based data which is dynamically reassigned during a simulation. In this case, the current to be injected at the next time step.

Reimplemented in AdExSpikingNeurons, SpikingNeurons, IzhikevichSpikingNeurons, GeneratorInputSpikingNeurons, and PoissonInputSpikingNeurons.

void Neurons::set_threads_per_block_and_blocks_per_grid ( int threads )

A local, non-polymorphic function called in to determine the CUDA Device thread (Neurons::threads_per_block) and block dimensions (Neurons::number_of_neuron_blocks_per_grid).

Reimplemented in GeneratorInputSpikingNeurons, and PoissonInputSpikingNeurons.