#cuTauLeaping release 1.0.0
cuTauLeaping is a stochastic simulator of biological systems that exploits the remarkable memory bandwidth and computational capability of GPUs. cuTauLeaping allows to efficiently execute in parallel large numbers of stochastic simulations, which are usually required to investigate the emergent dynamics of a given biological system under different conditions. cuTauLeaping is based on Cao's improved version of tau-leaping (https://dx.doi.org/10.1063%2F1.2159468).
HOW TO CITE cuTauLeaping
Nobile M.S., Cazzaniga P., Besozzi D., Pescini D., Mauri G.: cuTauLeaping: a GPU-powered tau-leaping stochastic simulator for massive parallel analyses of biological systems, PLoS ONE, 9(3): e91963, 2014
Just the Nvidia CUDA library (version >7.0).
A cuTauLeaping binary can be compiled on any supported architecture (e.g., GNU/Linux, Microsoft Windows, Apple OS/X) using the following compilation command:
nvcc kernel.cu 2phase_n-tau-leaping.cu -gencode=arch=compute_20,code=compute_20 -O3 -o cuTauLeaping --use_fast_math
The command above would create a binary executable file runnable on GPUs with at least a compute capability equal to 2.0. Please note that a specific compute capability, supporting additional functionality, can be targeted by using the
gencode argument. For instance, to target the compute capability 3.5, the following argument can be passed to
cuTauLeaping is designed to be launched from the command line. The arguments are:
./cuTauLeaping input_folder threads blocks gpu offset output_folder prefix fitness force_ssa
input_folderis the path to the directory containing the input model;
threadsis the number of CUDA threads per blocks;
blocksis the number of CUDA blocks used to distribute the requested parallel threads;
output_folderis the path to the directory that will store the output dynamics of the simulations;
prefixis the file name of the output files. A number, corresponding the thread, will be automatically appended to the filename by cupSODA;
force_ssaforces the algorithm to rely only on exact SSA steps.
Further information about the
offset arguments, along with the specifications of the input files, can be found at the following address: