Skip to main content
. 2011 Apr 9;2:14. doi: 10.3389/fphys.2011.00014

Figure 3.

Figure 3

Benchmark results of strong scalability experiments. Shown are the effects of nodal-based (NBP) versus ParMeTis-based (PaBP) domain decomposition, implicit (IM) versus explicit (EX) solver strategy, and inline versus asynchronous (async) I/O strategy on strong scalability for a range of cores Nc between 128 and 16,384. The top panel shows parallel speedup and realtime lag factor ξr whereas the bottom panel shows parallel efficiency.