sn-p-test-cases

Hello! This repo holds all test cases for evaluating Spiking Neural P (SN P) system simulations.

Downloading test cases

You can download all existing test cases by visiting the Releases section and downloading the {release}-sn-p-test-cases.zip archive attached to the latest release.

Test case format

Each format folder (xml, json, yaml) contains several systems. For each system, there's a file named $\verb!name.format!$ or $\verb!name(inputs).format!$ for the system itself and a file named $\verb!name.log!$ or $\verb!name(inputs).log!$ in the log folder to help track the system's behavior.

Test case list

All notes about probabilities below apply only for pseudorandom simulations (i.e., the simulator chooses the rule a neuron will apply uniformly at random).

Name	Function	Source	Notes
even_positive_integer_generator	generates $\{2k\mid k \ge 1\}$ using nondeterminism	Păun	probability of generating $2k$ is $\dfrac{1}{2^{k}}$
multiples_of(n)	generates $\{nk \mid k \ge 2\}$ using nondeterminism	Tim & Joshua	probability of generating $nk$ is $\dfrac{1}{2^{k-1}}$
boolean_function(f(b))	if input neuron $b_{i}$ is constantly provided with bit $b[i]$, there is an output spike at $t = 3$ iff $f(b) = 1$, where $f: \{0, 1\}^{|b|} \rightarrow \{0, 1\}$	generalization of Păun
increment(v)	increment the number $v$ in the register neuron $r$ by $1$ (a neuron containing $2k$ spikes represents number $k$); module passes control to one of $l_{j}$ and $l_{k}$ at random	Leporati et al.
decrement(v)	decrement the number $v$ in the register neuron $r$ by $1$ (a neuron containing $2k$ spikes represents number $k$); module passes control to $l_{j}$ if $v > 0$ and $l_{k}$ otherwise	Leporati et al.
bit_adder(L)	given the reversed binary representations of the elements in $L$, if $s$ is the sum of these elements, output spike train is the reversed binary representation of $s \times 2^{|L| - 2}$	generalization of WebSnapse v2 test cases
comparator(a,b)	given the unary representations of $a$ and $b$, the spike trains of output neurons $\min$ and $\max$ are the unary representations of $\min(a, b)$ and $\max(a, b)$, respectively	WebSnapse v2 test cases
subset_sum(L,s)	halts iff there is a subset of $L$ whose sum is $s$	Leporati et al. (with minor edits)	if $N$ is the number of subsets of $L$ whose sum is $s$, probability of halting is $\dfrac{N}{2^{|L|}}$
complete_graph(n)	a complete directed graph of neurons for stress testing; at time $t$, each neuron contains $2t + 1$ spikes	Louie's stress tester