Skip to main content

Table 2 Summary of memory and runtime performance for the three strategies on different datasets

From: Scalable transcriptomics analysis with Dask: applications in data science and machine learning

 

Peak memory (MiB)

Minimum runtime (s)

Framework

DT

DP

SPE

DT

DP

SPE

BRCA n1,205 \(\times\) f60,483 (713.3 MB)

12,380

14,039

OOM

984

1623

OOM

BRCA Coding n1,205 \(\times\) f19,564 (311.3 MB)

2897

5870

4489

903

1057

1183

LUAD/LUSC n911 \(\times\) f19,564 (194.9 MB)

2988

5204

3628

410

456

561

SYNTH n5,000 \(\times\) f20,000 (300.9 MB)

11,120

12,217

16,443

585

373

1552

BRCA Coding HPO n1,205 \(\times\) f19,564 (311.3 MB)

4208

6854

10,368

1469

1478

2324

  1. Tests were performed with and without intensive hyper-parameter optimization (exhaustive grid search). Bolded values represent the best-performing result for each dataset and metric pair. The data sizes shown in MB correspond to the uncompressed tabular files. OOM out of memory
Back to article page

BMC Bioinformatics

ISSN: 1471-2105

Contact us