Table 1 Comparison of computational efficiency of Mash, BinDash, and Dashing at k=31 and various sketch sizes

From: Dashing: fast and accurate genomic distances with HyperLogLog

						Dashing	Dashing	Dashing
Phase	Measure	k	log2(size)	Mash	BinDash	Original	Ertl-MLE	Ertl-JMLE
Sketch	Wall clock (s)	31	10	1345	1157	273	271	277
			12	1349	1157	273	274	270
			14	1356	1159	286	289	278
			16	1400	1226	359	367	299
	Peak mem (MB)	31	10	17,720	141	12,683	12,721	12,644
			12	18,296	399	12,723	12,430	12,726
			14	19,706	1426	12,630	12,877	12,853
			16	25,127	5542	12,888	12,412	12,933
Distance	Wall clock (s)	31	10	1901	74	80	100	601
			12	2368	188	286	308	2139
			14	3446	672	1113	1137	8308
			16	8777	3603	6172	4251	30,506
	Peak mem (MB)	31	10	1120	409	116	116	116
			12	1380	673	371	371	372
			14	2785	1,709	1392	1392	1392
			16	10,776	5,816	5476	5476	5476
Both	Wall clock (s)	31	10	3246	1,231	345	365	870
			12	3717	1,345	557	579	2407
			14	4801	1,831	1390	1,408	8574
			16	10,177	4,829	4394	4,453	30,433
	Peak mem (MB)	31	10	17,720	409	12,468	12,950	12,988
			12	18,296	673	12,958	13,042	13,020
			14	19,706	1709	13,951	13,782	14,205
			16	25,127	5816	18,320	18,081	18,011

The log2(size) column reports the log2 of the sketch size in bytes. “Both” results obtained either by using a combined Sketch+Distance mode (for Dashing) or by combining results from separate sketching and distance-calculation invocations (for Mash and BinDash). Dashing was assessed using three estimation methods: Flajolet’s method using the harmonic mean (“Original”) and Ertl’s MLE and JMLE methods. Italicized entries indicate the lowest space or time for a given experiment

Back to article page

ISSN: 1474-760X

Contact us

Submission enquiries: editorial@genomebiology.com
General enquiries: info@biomedcentral.com