Skip to main content

Table 1 Average speedup of CES over NPS for all PaCo tasks taking on average at least 15 min

From: Impact of concurrency on the performance of a whole exome sequencing pipeline

Task 4 processors 8 processors 16 processors
Mean ± Std (s) Mean ± Std (s) Mean ± Std (s)
(a) 2-samples execution
 Align to nuclear 1.75 ± 0.01 2.84 ± 0.02 3.19 ± 0.04
 Align to exome 1.76 ± 0.01 1.79 ± 0.01 2.87 ± 0.07
 Identify INDELs nuclear 1.95 ± 0.01 1.83 ± 0.01 2.15 ± 0.09
 Compute BQSR nuclear 2.38 ± 0.03 2.69 ± 0.02 2.91 ± 0.15
 Post BQSR nuclear 3.09 ± 0.01 3.17 ± 0.03 3.1 ± 0.04
Task 6 processors 12 processors 24 processors
Mean ± Std (s) Mean ± Std (s) Mean ± Std (s)
(b) 3-samples execution.
 Align to nuclear 2.05 ± 0.01 3.72 ± 0.01 3.91 ± 0.05
 Align to exome 1.61 ± 0.01 2.06 ± 0.01 3.62 ± 0.05
 Identify INDELs nuclear 1.88 ± 0.03 2.2 ± 0.06 2.75 ± 0.16
 Compute BQSR nuclear 2.96 ± 0.02 3.26 ± 0.03 3.96 ± 0.14
 Post BQSR nuclear 4.32 ± 0.03 4.37 ± 0.12 4.43 ± 0.04
  1. Number of processors stands for the number of available processors managed by both strategies