Santa Fe Laser [15]
Test type	Best error	Mean error	Mean error bias	Worst error	Cost	Cost per thread	Mean (best validation error epoch)
A	0.0122094	0.0369875	5.63%	0.0647200	200,075	40,015	1619
B	0.0151571	0.0430133	−17.93%	0.0624001	195,075	39,015	1742
C	0.0138906	0.0323462	11.01%	0.0553674	500,150	50,015	1715
D	0.0186784	0.0606052	38.98%	0.1560946	200,075	40,015	1509
E	0.0203611	0.0656812	54.02%	0.2174987	200,075	40,015	1269
Sequential [9]	0.0571243	0.0695330	−65.25%	0.0721420	269,421	-	-
Concurrent [9]	0.0634781	0.0768557	−68.66%	0.0793412	121,200	-	-
MCNE [8] [9]	0.1471420	0.1949820	−33.44%	0.2188464	100,000	-	-
Lorenz [13]
Test type	Best error	Mean error	Mean error bias	Worst error	Cost	Cost per thread	Mean validation error epoch
A	0.0027403	0.0052443	24.78%	0.0093983	200,075	40,015	1897
B	0.0022897	0.0096347	33.82%	0.0244856	195,075	39,015	1813
C	0.0019629	0.0038730	28.18%	0.0072823	500,150	50,015	1949
D	0.0022847	0.0144763	32.90%	0.0386242	200,075	40,015	1767
E	0.0037872	0.0146563	62.75%	0.0621459	200,075	40,015	1809
Sequential [9]	0.0713540	0.0731450	48.59%	0.0783210	260,668	-	-
Concurrent [9]	0.3214887	0.3445700	22.62%	0.3811421	121,200	-	-
MCNE [8] [9]	0.0747062	0.0753210	73.42%	0.0793321	100,000	-	-
Mackey Glass [14]
Test type	Best error	Mean error	Mean error bias	Worst error	Cost	Cost per thread	Mean validation error epoch
A	0.0027723	0.0037160	45.77%	0.0062529	200,075	40,015	1880
B	0.0027492	0.0037760	14.47%	0.0051503	195,075	39,015	1864
C	0.0023960	0.0033838	11.41%	0.0046261	500,150	50,015	1966
D	0.0010170	0.0040382	6.24%	0.0074617	200,075	40,015	1814
E	0.0020943	0.0043915	36.02%	0.0092757	200,075	40,015	1873
Sequential [9]	0.0019264	0.0045463	−37.10%	0.0057482	271,031	-	-
Concurrent [9]	0.0032004	0.0059527	−56.75%	0.0067121	121,200	-	-
MCNE [8] [9]	0.0123215	0.0252556	−16.57%	0.0345122	100,000	-	-