Scalable linkage-disequilibrium-based selective sweep detection: a performance guide

GigaScience

Table 2 Execution times (in seconds) for the analysis of binary datasets in different shapes

Sequences	100			1,000			10,000			50,000
Sites	10,000	50,000	100,000	10,000	50,000	100,000	10,000	50,000	100,000	10,000	50,000	100,000
OP (1)	151	2,024	2,467	241	2,653	2,535	438	3,442	4,433	777	5,636	11,485
F (4)	39	1,238	1,480	67	1,313	1,569	135	1,664	2,009	252	2,364	4,380
C (4)	64	1,017	1,052	114	1,015	991	195	1,265	1,528	469	2,627	4,360
M (4)	40	701	861	69	796	889	151	1,099	1,398	401	2,205	4,008
G (4)	40	521	624	65	545	652	115	924	1,249	250	1,975	4,395
F (8)	19	795	1,011	32	838	1,002	53	1,012	1,214	134	1,438	2,684
C (8)	36	464	488	64	478	487	132	709	869	373	1,701	2,857
M (8)	23	379	423	41	408	445	105	638	813	335	1,483	2,681
G (8)	20	254	312	32	274	328	59	481	679	144	1,091	2,570
F (16)	10	652	808	17	675	828	33	834	975	97	1,250	2,250
C (16)	20	258	264	39	266	280	95	463	585	320	1,255	2,109
M (16)	16	222	243	30	233	250	84	429	549	306	1,198	2,021
G (16)	10	128	157	16	140	169	35	275	411	97	685	1,664
F (32)	9	598	699	15	623	743	31	764	860	84	1,066	1,919
C (32)	14	147	166	24	158	185	76	338	446	325	1,193	1,990
M (32)	16	137	165	22	161	173	77	337	446	313	1,133	2,064
G (32)	7	86	95	11	91	104	26	193	317	85	504	1,229

OP indicates the sequential OmegaPlus implementation; F, C, M, and G indicate the fine-grained, coarse-grained, multi-grained, and generic parallel algorithms, respectively. The number in parentheses is the number of threads

ISSN: 2047-217X