Parallel Algorithm for Solving TOV Equations for Sequence of Cold and Dense Nuclear Matter Models

Alexander Ayriyan, Jan Busa Jr., Hovik Grigorian, and Gevorg Poghosyan. Parallel Algorithm for Solving TOV Equations for Sequence of Cold and Dense Nuclear Matter Models // European Physics Journal WoC, vol. 177, 07001 (2018)

Alexander Ayriyan1, Ján Buša Jr.1,2, Hovik Grigorian1,3 and Gevorg Poghosyan4

1 Laboratory of Information Technologies, JINR
2 Institute of Experimental Physics, SAS, Slovakia
3 Department of Theoretical Physics, YSU, Armenia
4 Karlsruhe Institute of Technology, Germany

We have introduced parallel algorithm simulation of neutron star configurations for set of equation of state models. The performance of the parallel algorithm has been investigated for testing set of EoS models on two computational systems. It scales when using with MPI on modern CPUs and this investigation allowed us also to compare two different types of computational nodes.

mpi_tov_time
Figure 1. The computational time of parallel computations vs number of processes
mpi_tov_speedup
Figure 2. The speedup of parallel computations vs number of processes
mpi_tov_efficiency
Figure 3. The efficiency of parallel computations vs number of processes