Table 6

Table 6. The CPU Time Consumption (in Minutes) of Each Test Job Concurrently Executed in Duplicate in the E7505 System.

System	E7505 (CL2)
Compiler	ifc 7.1						PGI 5.0-2
Arch. opt.^a	(a)						(b)
Library	BLAS			MKL 6.0			GOTO 0.6p
Tasking	Single	Dup. 1	Dup. 2	Single	Dup. 1	Dup. 2	Single	Dup. 1	Dup. 2
322-1	2.15	3.37	3.38	2.13	3.34	3.35	2.69	3.70	3.68
322-2	56	68	68	55	74	73	92	113	113
322-3	80	99	100	80	95	95	121	136	136
338	13	17	17	13	16	16	20	24	24
339	15	19	18	15	18	18	22	27	27
364	23	24	24	23	24	24	30	47	47
397	220	315	325	216	339	340	316	415	423
415-1	10	12	12	9	12	11	12	16	16
415-2	22	28	28	20	27	27	25	33	34
420-1	46	68	70	46	80	81	58	80	80
420-2	47	75	72	45	69	69	59	76	76
424-1	32	45	46	26	35	35	33	47	48
424-2	32	43	44	26	34	33	32	48	49
438	17	21	20	16	18	18	18	22	24
439-1^b	42(2)	53(2)	53(2)	42(2)	56(2)	55(2)	67(2)	74(2)	76(2)
439-2	32	43	43	32	42	42	51	59	60
447-1	73	95	95	68	93	92	106	127	127
447-2	36	42	43	35	41	41	48	55	55
559-4	9	11	11	10	13	13	12	16	16
560-4	10	12	13	11	14	14	14	18	18
561-4	25	27	27	25	28	28	38	42	failed^c
Sum	800	1067	1079	773	1075	1073	1110	1405	-

^a The global options of ifc is set to "-unroll -O2", plus: (a) "-tpp7 -axW -ipo -ipo_obj". The global optimization to pgf77 is set to "-Munroll -O2", plus: (b) "-tp p7 -Mvect=cachesize:524288". ^bThe numbers in the parentheses are the step numbers required in the geometrical optimization calculations; this CPU time is excluded from sum. ^cReproducible failure while running two copies of test516 concurrently.