Skip to content

In current version of CAVIAR (https://github.com/fhormoz/caviar), it runs in a serial manner. To speed up the running, I convert it to a parallel version, which applies openmp to run with multi-thread. Simulation results demonstrate that when explored variants are massive, the running time is only 1/nthread of original version.

Notifications You must be signed in to change notification settings

jxzb1988/Parallel_CAVIAR

Repository files navigation

Parallel_CAVIAR

In current version of CAVIAR (https://github.com/fhormoz/caviar), it runs in a serial manner. To speed up the running, I convert it to a parallel version, which applies openmp to run with multi-thread. Simulation results demonstrate that when the explored variants are massive , the running time is only 1/nthread of original version. Most of the codes are still as same as the original,The main changes happen in PostCal.cpp. In previous version, when calculating the likelihood of the causal states, it was run one by one. In this parallel version, firstly, I store the causal status in a 2-dimensional vector, and then visit causal status in parallel. The parallel version can be much faster when multiple CPUs are available. To run it, please add "-t n", n is the number of threads.

About

In current version of CAVIAR (https://github.com/fhormoz/caviar), it runs in a serial manner. To speed up the running, I convert it to a parallel version, which applies openmp to run with multi-thread. Simulation results demonstrate that when explored variants are massive, the running time is only 1/nthread of original version.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages