A New Algorithm for Solving Large-Scale Generalized Eigenvalue Problem Based on Projection Methods

Table 3

The minimum relative residual of Algorithm 1–4 and in Example 3.

N.A, N.B	m = 16, N = 64		M = 18, N = 64 m = 16\		M = 64, m = 4, l = 4		M = 4, L = 4, N = 64
					Threshold =
	SS	NE	CIRR	NE	BSS	NE	BSSRR	NE
n = 200	0.6600	16	5.7595E − 13	16	3.4279E − 11	8	6.3585E − 15	16
n = 400	0.6028	16	7.2639E − 11	16	6.1907E − 11	10	3.1440E − 15	16
n = 600	0.6509	16	3.4114E − 11	16	3.0774E − 09	11	3.3606E − 15	16
n = 800	0.5966	16	6.2826E − 11	16	3.0774E − 08	12	8.2243E − 15	16
n = 1000	0.6065	16	6.8571E − 10	16	9.7887E − 08	12	3.3916E − 15	16
n = 1500	0.6737	16	3.6078E − 12	16	7.9782E − 08	11	2.3400E − 15	16