Skip to main content

Table 7 Runtimes and peak memory consumption for non-partitioned database construction (build) and querying for different data sets on a workstation with 512 GB RAM

From: A big data approach to metagenomics for all-food-sequencing

Data set

 

AFS-MetaCache

CLARK

Kraken2

Kraken2+Bracken

AFS20

Build time

1h 11m

15h 37m

1h 27m

5h 32m

 

Build memory

64 GB

428 GB

69 GB

147 GB

 

Query time

136 s

93 s

37 s

111 s

 

Query speed

11.5 MR/m

16.9 MR/m

43.2 MR/m

14.2 MR/m

 

Query memory

50 GB

152 GB

54 GB

54 GB

AFS31

Build time

1h 47m

-

3h 19min

11h 41min

 

Build memory

91 GB

-

107 GB

296 GB

 

Query time

175 s

-

44 s

58 s

 

Query speed

8.9 MR/m

-

35.9 MR/m

27.0 MR/m

 

Query memory

78 GB

-

72 GB

72 GB

AFS20RS90

Build time

1h 42m

-

2h 58m

8h 53m

 

Build memory

110 GB

-

94 GB

168 GB

 

Query time

180 s

-

43 s

117 s

 

Query speed

8.7 MR/m

-

37.0 MR/m

13.5 MR/m

 

Query memory

94 GB

-

79 GB

79 GB

AFS31RS90

Build time

3h 10m

-

5h 55min

17h 44min

 

Build memory

135 GB

-

134 GB

329 GB

 

Query time

217 s

-

49 s

61 s

 

Query speed

7.2 MR/m

-

32.1 MR/m

25.7 MR/m

 

Query memory

117 GB

-

97 GB

97 GB

  1. Query speeds are measured for the KAL_D dataset in terms of million reads per minute (MR/m). For the cases with “-” the corresponding program exceeds the main memory capacity of 512 GB. Fastest runtimes and lowest memory consumption for each dataset are indicated in bold