Skip to main content
. Author manuscript; available in PMC: 2015 Feb 1.
Published in final edited form as: J Comput Appl Math. 2013 Sep 6;257:195–211. doi: 10.1016/j.cam.2013.09.001

Table 6.

Per global iteration running times in milliseconds for all meshes: S1 and S2 are the speedups of patchPCGAMG compared to Hypre-PCGAMG and CUSP-PCGAMG. Speedup number is in parentheses when patchPCGAMG is slower.

meshes patch PCGAMG Hypre-PCGAMG S1 CUSP-PCGAMG S2
Regular 7.31 154 21 4.68 (1.6)
Irregular 5.40 104 19 3.86 (1.4)
Heart 10.9 361 33 13.7 1.3
Brain 8.71 288 33 9.60 1.1
Blobs 7.49 204 27 8.18 1.1