Skip to main content

Table 1 Comparison of the construction performance with the construction time and memory for all datasets

From: Pfp-fm: an accelerated FM-index

Dataset

n

Method

CONSTRUCT

INDEX

CONSTRUCT

   

MEMORY

SIZE

TIME

SARS-25k

751,526,774

RLCSA

9.90

0.026

322.85

RLFM

3.47

0.136

363.74

FIGISS

4.89

0.003

378.49

PFP-FM

12.99

4.318

117.29

  

PFP-FM-CSA

15.68

1.689

772.98

  

FM-index

13.35

4.399

120.08

  

Bowtie

3.55

0.47

7851.35

  

Bowtie2

3.54

0.59

6847.03

SARS-50k

1,503,252,577

RLCSA

19.88

0.051

679.89

RLFM

6.94

0.278

701.36

FIGISS

12.44

0.006

795.70

PFP-FM

26.12

8.763

233.04

  

PFP-FM-CSA

30.95

3.078

1546.75

  

FM-index

26.12

8.490

237.50

  

Bowtie

7.09

0.94

28238.74

  

Bowtie2

7.09

1.18

15242.00

SARS-100k

3,004,588,730

RLCSA

39.47

0.099

1690.22

RLFM

25.01

0.571

1432.16

FIGISS

25.57

0.009

1840.80

PFP-FM

53.90

18.156

489.45

  

PFP-FM-CSA

61.86

5.758

3150.72

  

FM-index

51.85

16.73

434.55

  

Bowtie

14.20

1.884

32143.48

  

Bowtie2

14.19

2.37

33914.46

GRCh38

3,189,750,467

RLCSA

45.45

2.022

924.60

RLFM

26.31

3.101

1839.25

FIGISS

34.65

1.538

1440.19

PFP-FM

71.13

37.862

1154.12

  

FM-index

70.93

32.54

877.43

  

PFP-FM-CSA

N/A

N/A

N/A

  

Bowtie

13.99

1.833

2160.76

  

Bowtie2

14.00

2.31

2170.32

  1. The number of characters in each dataset (denoted as n) is in the second column
  2. The construction time is reported in seconds (denoted as CONSTRUCT TIME)
  3. The construction memory is reported in gigabytes (denoted as CONSTRUCT MEM)
  4. The index size is reported in gigabytes (denoted as INDEX SIZE)
  5. The implementation of the FM-index that we used was sourced from the sdsl library