SWAP-Assembler: scalable and efficient genome assembly towards thousands of cores

BMC Bioinformatics

Table 6 Scalability evaluation on parallel assemblers 1 (Time unit: seconds).

dataset	software	64	128	256	512	1024	2048	4096
S. aureus (2.87 Mb)	ABySS	248	269	334	554	831	_-2	-
	Ray	244	198	202	266	510	-	-
	SWAP-Assembler	23	15	8	5	7	13	-
R. sphaeroides (4.60 Mb)	ABySS	492	454	522	718	1312	-	-
	Ray	287	183	181	190	285	-	-
	SWAP-Assembler	43	29	15	7	7	7	-
Hg14 (88.29 Mb)	ABySS	6472	5299	6935	9045	16530	-	-
	Ray	2926	1746	1288	1517	2266	-	-
	SWAP-Assembler	585	428	203	128	59	67	-
Fish (1 Gb	ABySS	_* 3	₄	+	+	+	-	-
	Ray	*	+	+	+	+	-	-
	SWAP-Assembler	+	13941	8622	3263	962	971	2582
Yanhuang (3 Gb)	ABySS	*	+	+	+	+	-	-
	Ray	*	+	+	+	+	-	-
	SWAP-Assembler	*	11243	5761	4021	1783	1608	1778

1 This table records the time usage on assembling all five datasets with different number of cores, and the length of k-mer is set to be 23. For ABySS and Ray, the time is recorded until contigs are generated.
2 − denotes assembler with this parameter has not been tested.
3 ∗ denotes assembler with this parameter has run out of memory.
4 + denotes assembler with this parameter has run out of time, the time limit is 12 hours.

ISSN: 1471-2105