Deprecated Propagators

Outdated signal propagation components

Types

Name	Description
InputBehavior	Determines how computed inputs are combined with existing buffers.

Functions

Name	Description
dense_p	Dense matrix propagation using cuBLAS sgemv.
forward_p	One-to-one propagation (diagonal weight vector).
sparse_p	Sparse propagation using explicit edge lists.

Function Details

dense_p

template<size_t stateVar, typename PreStateTypes, typename PostStateTypes> void dense_p( genericLayer<PreStateTypes>& preLayer, genericLayer<PostStateTypes>& postLayer, const thrust::device_vector<float>& W_flat, // column-major (N_post × N_pre) cublasHandle_t handle, InputBehavior behavior = INPUT_OVERRIDE, cudaStream_t stream = 0 )

Dense matrix propagation using cuBLAS sgemv.

stateVar: Index of the presynaptic state variable to propagate.
PreStateTypes: Tuple type for the presynaptic layer.
PostStateTypes: Tuple type for the postsynaptic layer.
preLayer: Source layer.
postLayer: Destination layer.
W_flat: Column-major weight matrix (N_post x N_pre).
handle: cuBLAS handle.
behavior: Input accumulation behavior.
stream: CUDA stream for execution.

forward_p

template<size_t stateVar, typename PreStateTypes, typename PostStateTypes> void forward_p( genericLayer<PreStateTypes>& preLayer, genericLayer<PostStateTypes>& postLayer, const thrust::device_vector<float>& W, InputBehavior behavior = INPUT_OVERRIDE, cudaStream_t stream = 0 )

One-to-one propagation (diagonal weight vector).

stateVar: Index of the presynaptic state variable to propagate.
PreStateTypes: Tuple type for the presynaptic layer.
PostStateTypes: Tuple type for the postsynaptic layer.
preLayer: Source layer.
postLayer: Destination layer.
W: Weight vector aligned with neuron indices.
behavior: Input accumulation behavior.
stream: CUDA stream for execution.

sparse_p

template<size_t stateVar, typename PreStateTypes, typename PostStateTypes> void sparse_p( genericLayer<PreStateTypes>& preLayer, genericLayer<PostStateTypes>& postLayer, const thrust::device_vector<int>& X, // presynaptic indices const thrust::device_vector<int>& Xn, // postsynaptic indices const thrust::device_vector<float>& W, // weights InputBehavior behavior = INPUT_OVERRIDE, cudaStream_t stream = 0 )

Sparse propagation using explicit edge lists.

stateVar: Index of the presynaptic state variable to propagate.
PreStateTypes: Tuple type for the presynaptic layer.
PostStateTypes: Tuple type for the postsynaptic layer.
preLayer: Source layer.
postLayer: Destination layer.
X: Presynaptic indices.
Xn: Postsynaptic indices.
W: Connection weights.
behavior: Input accumulation behavior.
stream: CUDA stream for execution.