Table 26 Performance of scheduling algorithms on random task graphs on AWS EC2 Type-1, Type- 2, 2 Grids and 11 CPUs.
From: Multi-objective and multi constrained task scheduling framework for computational grids
Number of gridlets (n) | Scheduler | A task graph with 100% connectivity | |||
|---|---|---|---|---|---|
TAT (seconds) | Resource utilization (%) | Execution cost ($) | Communication cost ($) | ||
100 | Greedy (with fragmentation) scheduler | 223.11 | 99.6 | 41.35 | 222.2 |
Greedy communication cost scheduler | 286.28 | 77.62 | 39.68 | 0 | |
Random scheduler | 1098 | 20.24 | 44.44 | 243.75 | |
Greedy execution cost scheduler | 2000.08 | 11.11 | 39.68 | 0 | |
Greedy (without-Fragmentation) scheduler | 2000.08 | 11.11 | 39.68 | 0 | |
300 | Greedy (with fragmentation) scheduler | 669.34 | 99.6 | 124.06 | 666.6 |
Greedy communication cost scheduler | 858.84 | 77.62 | 119.04 | 0 | |
Random scheduler | 3232.57 | 20.62 | 133.75 | 728.72 | |
Greedy execution cost scheduler | 6000.25 | 11.11 | 119.05 | 0 | |
Greedy (without-fragmentation) scheduler | 6000.25 | 11.11 | 119.05 | 0 | |
500 | Greedy (with fragmentation) scheduler | 1115.57 | 99.6 | 206.77 | 1111 |
Greedy communication cost scheduler | 1431.41 | 77.62 | 198.4 | 0 | |
Random scheduler | 5536.25 | 20.07 | 222.33 | 1176.6 | |
Greedy execution cost scheduler | 10000.42 | 11.11 | 198.41 | 0 | |
Greedy (without-fragmentation) scheduler | 10000.42 | 11.11 | 198.41 | 0 | |
