Table 4 Execution time for the second experiment on a standalone spark cluster with 2-VM workers
From: Instance segmentation on distributed deep learning big data cluster
Cluster Configuration: - Master VM in specs (7 GB RAM, Total processor cores = 8, HD = 130 GB)—Slave1 VM in specs (4.5 GB RAM, Total processor cores = 6, HD = 130 GB) And Spark cluster configured by driver-memory =3 GB, num-executors = 2, executor-cores = 4, driver-cores = 2, executor-memory = 3 GB | Spark cluster stand-alone stand-up time with 2 Executors | 14 s |
Total sum execution time for 26 images as one patch task job in 2 Executors | 1.28 min | |
Work Executors in Master VM processing time | 35.7 s | |
Work Executors in slave1 VM processing time | 36.7 s | |
One Frame/image Processing time | 36.7/26 =1.4 s it calculated by the long executor time | |
WIFI speed on the local network | 130 Mbps |