||cirsiumNeuron is a neuronal circuit simulator that can directly and efficiently compute characterizations of stochastic behavior, i.e., noise, for multi-neuron circuits. In cirsiumNeuron, we utilize a general modeling framework for biological neuronal circuits which systematically captures the nonstationary stochastic behavior of the ion channels and the synaptic processes. In this framework, we employ fine-grained, discrete-state, continuous-time Markov Chain (MC) models of both ion channels and synaptic processes in a unified manner. Our modeling framework can automatically generate the corresponding coarse-grained, continuous-state, continuous-time Stochastic Differential Equation (SDE) models. In addition, for the stochastic characterization of neuronal variability and noise, we have implemented semi-analytical, non Monte Carlo analysis techniques that work both in time and frequency domains, which were previously developed for analog electronic circuits. In these semi-analytical noise evaluation schemes, (differential) equations that directly govern probabilistic characterizations in the form of correlation functions (time domain) or spectral densities (frequency domain) are first derived analytically, and then solved numerically. These semi-analytical noise analysis techniques correctly and accurately capture the second order statistics (mean, variance, autocorrelation, and power spectral density) of the underlying neuronal processes as compared with Monte Carlo simulations.