favorite0In this paper we presented the R package micompr, which implements a procedure for comparing multivariate samples associated with different factor levels or groups.
favorite0The package includes test data produced by several implementations of the Predator-Prey for High Performance Computing (PPHPC) simulation model (Fachada et al., 2015).
favorite0The procedure uses principal component analysis (PCA) (Jolliffe, 2002) to convert multivariate observations into a set of linearly uncorrelated statistical measures, which are then compared using a number of statistical methods, such as hypothesis tests and score plots.
favorite1The research goal is to differentiate among pre-specified, well-defined classes or groups of sampling entities generating highly multivariate observations in which the dimensions or dependent variables are correlated, and to test for significant differences among groups.
favorite0The aim of this paper is to present the micompr package for R (R Core Team, 2015), which implements a procedure for comparing multivariate samples associated with different factor levels or groups.
favorite17The latter, stats_compare_table, is a very versatile function which outputs a LATEX table with p-values resulting from statistical tests used to evaluate the alignment of model implementations.
favorite11Model comparison functions Utilities in the model comparison group aid the modeler in comparing and aligning simulation models through informative tables and plots, also producing publication quality LATEX tables containing p-values yielded by user-specified 5.
favorite8Distributional analysis functions Functions in the distributional analysis module generate tables and figures which summarize different aspects of the statistical distributions of FMs. The dist_plot_per_fm and dist_table_per_fm functions focus on one FM and provide a distributional analysis over several setups or configurations, i.e., over a number of model scales and/or parameter sets.
favorite2These utilities were originally developed to study the Predator-Prey for HighPerformance Computing (PPHPC) agent-based model , namely by statistically analyzing its outputs for a number of different parameters and comparing the dynamical behavior of different implementations [4, 5, 6].
favorite6(1) Overview Introduction SimOutUtils is a suite of MATLAB  functions for studying and analyzing time series-like output from stochastic simulation models, as well as for producing associated publication quality figures and tables.
favorite7Abstract SimOutUtils is a suite of MATLAB/Octave functions for studying and analyzing time series-like output from stochastic simulation models.
favorite12The alignment of two or more implementations can be assessed by analyzing the following information: 1) the p-values produced by the univariate and multivariate statistical tests, which should be above the typical 1% or 5% significance levels in case of implementation alignment; in the univariate case, it may be useful to adjust the p-values using the weighted Bonferroni procedure to account for multiple comparisons; 2) in case of misalignment, the total number of PCs required to explain a prespecified amount of variance should be lower than in case of alignment; also, more variance should be explained by the first PCs of the former than by the same PCs of the latter; and, 3) the scatter plot of the first two PC dimensions, which can offer visual, although subjective feedback on model alignment; e.g., in case of misalignment, points associated with runs from different implementations should form distinct groups..
favorite20This work is one of the main references in model replication, describing in detail the process of running two model implementations with different parameters, selecting comparison measures and performing adequate statistical tests.
favorite2We present a model comparison technique, which uses principal component analysis (PCA)  to convert simulation output into a set of linearly uncorrelated statistical measures, analyzable in a consistent, model-independent fashion.
favorite2ABMs are commonly implemented as a stochastic process, and thus require multiple runs (observations) with distinct pseudo-random number generator (PRNG) seeds in order to have appropriate sample sizes for testing hypotheses and differentiating multiple scenarios under distinct parameterizations .
favorite115In this paper, we present a model comparison technique, which uses principal component analysis to convert simulation output into a set of linearly uncorrelated statistical measures, analyzable in a consistent, model-independent fashion.
favorite647It features compile-time logical device configuration, management of several OpenCL object types, profiling capabilities, helper functions for manipulation of events, and a number of utilities for program manipulation.
favorite5It offers a very high-level abstraction, with a minimum of two types and three functions required to execute OpenCL code on a single compute device.
favorite65These libraries aim for a number of goals, such as rapid and/or simplified development of OpenCL programs, high-level abstractions for common computation and communication patterns, embedded OpenCL kernel code within C++ programs or handling of multiple OpenCL platforms and devices.
favorite17In this paper we present the C Framework for OpenCL, cf4ocl, a software library with the following goals: 1) promote the rapid development of OpenCL host programs in C (with support for C++), while avoiding the associated API verbosity; 2) assist in the benchmarking of OpenCL events, such as kernel execution and data transfers; and, 3) simplify the analysis of the OpenCL environment and of kernel requirements.
favorite4It aims to reduce the verbosity of the OpenCL API, offering straightforward memory management, integrated profiling of events (e.g., kernel execution and data transfers), simple but extensible device selection mechanism and user-friendly error management.
favorite1Data-parallel algorithms for agent-based model simulation of tuberculosis on graphics processing units.
favorite2Figure 6 displays, for the two parameter sets and selected model sizes (100, 400 and 1600), how the simulation time is affected by varying the number of worker threads in the multithreaded variants.
favorite0Implementations may differ in aspects such as the selected system architecture, choice of programming language and/or agent-based modeling framework, parallelization strategy, random number generator, and so forth.
favorite13. strategies offer specific trade-offs in terms of performance and simulation reproducibility; and, 3) PPHPC is a valid reference model for comparing distinct implementations or parallelization strategies, from both performance and statistical accuracy perspectives.
favorite1In this paper we present a multithreaded Java implementation of the PPHPC ABM, with two goals in mind: 1) compare the performance of this implementation with an existing NetLogo implementation; and, 2) study how different parallelization strategies impact simulation performance on a shared memory architecture.