Alya Red CCM: HPC-Based Cardiac Computational Modelling - Selected Topics of Computational and Experimental Fluid Mechanics - page 191

Fig. 4 Strong scalability for the coupled electro-mechanical problem as compared with the ideal
linear scalability. The vertical axis shows the speedup, the lower horizontal axis is the number
of cores and the upper horizontal axis is the elements-per-core average. Tests were carried out in
NCSA's Blue Waters supercomputer. Reprinted from Vázquez et al. ( 2014 )
3 Parallel Computational Aspects
3.1 Parallel Efficiency
In this paper we only present parallel efficiency results for the tissue part, i.e., the
electro-mechanical coupling. The fully coupled problem scalability is still being
assessed.
Strong scalability measures the Cardiac Computational Model (CCM) parallel
efficiency for the coupled electro-mechanical problem. Alya has shown high paral-
lel efficiency up to several thousands of cores for different physical problems, either
coupled or single-physics (Houzeaux et al. 2008 , 2009 , 2011 ; Puzyrev et al. 2013 ). In
Vázquez et al. ( 2014 ), the authors have performed a scalability test assessing Alya's
parallel performance up to 100,000 cores in NCSA's Blue Waters supercomputer.