Parallel Implementation of FEM Solver for Shared Memory Using OpenMP

Table 13

Conjugate Gradient solver parallel timing and speedup for system of equations having 2099201 unknowns.

Threads	1	2	3	4	5	6	7	8
CG	1566	1214	997	931	721	687	670	634
Speedup	1.00	1.29	1.57	1.68	2.17	2.28	2.34	2.47
Threads	9	10	11	12	13	14	15	16
CG	520	519.95	479	449	399	366.93	366	312
Speedup	3.01	3.01	3.27	3.49	3.93	4.27	4.28	5.01