Table 2.
Comparisons of database size, running time for ProbMinHash, Densified MinHash, SuperMinHash and SetSketch with all other pieces of software
| Database size (indicates minimum memory required) | database loading time | Sketching & searching | Total Wall time | |
|---|---|---|---|---|
| GSearch(ProbMinHash) | 29GB (sketch + graph) | 57.2 s | 8 min 23 s | 9 min 20 s |
| GSearch(Densified MinHash) | 9.7GB | 27.2 s | 4 min 12s | 4 min 40 s |
| GSearch(SuperMinHash) | 25GB (sketch + graph) | 41 s | 24 min 13 s | 24 min 54 s |
| GSearch(SetSketch) | 2.5GB (sketch + graph) | 9.1 s | 12 min 14 s | 12 min 23 s |
| Mash | 19.8GB | 1 min 11 s | 182.1 min | 183 min |
| Sourmash (branchwater) | 7.9GB | ∼30 s | 64.3 min | 65 min |
| Dashing 1 | 2.7GB (sketch) | ∼61 s | 20.5 min | 21.5 min |
| Dashing 2 (ProbMinHash) | 4.7GB (sketch) | ∼61 s | 47.5 min | 48.5 min |
| skania | 30.1GB | - | 74.42 min | 74.42 min |
| BinDash | 16.2GB (Sketch) | 16 s | 51.4 min | 41.4 min |
aFor ∼8000 query genomes, skani total memory requirement is larger than 60GB, the largest among all pieces of software.
Results are based on searching 8466 query genomes against all NCBI/RefSeq genomes (∼318K) on a 24-thread machine. Times reported are average values from 3 runs.