Research Article
Decision Scheduling for Cloud Computing Tasks Relying on Solving Large Linear Systems of Equations
Table 3
Experimental data of Jordan elimination method.
| Number of parallel nodes | Dense matrix scale | 2002 | 4002 | 5002 | 6002 | 10002 |
| | Total time | 0.0443 | 0.1012 | 0.0583 | 0.0034 | 0.1169 | | Communication time | 0.0676 | 0.0461 | 0.1171 | 0.0917 | 0.0782 | 0 | Parallel computing time | 0.006 | 0.1163 | 0.1053 | 0.0995 | 0.04 | | Speedup ratio | — | — | — | — | — | | Parallel efficiency | — | — | — | — | — | | Total time | 0.0699 | 0.0479 | 0.0687 | 0.0279 | 0.0364 | | Communication time | 0.048 | 0.0936 | 0.0636 | 0.0055 | 0.0449 | 3 | Parallel computing time | 0.099 | 0.0802 | 0.0514 | 0.1047 | 0.1189 | | Speedup ratio | 0.39 | 0.43 | 0.57 | 0.45 | 0.4 | | Parallel efficiency | 0.55 | 0.77 | 0.38 | 0.61 | 0.45 | | Total time | 0.0935 | 0.0582 | 0.1137 | 0.0265 | 0.063 | | Communication time | 0.0772 | 0.0392 | 0.0976 | 0.0854 | 0.0878 | 4 | Parallel computing time | 0.119 | 0.0998 | 0.033 | 0.1005 | 0.0662 | | Speedup ratio | 0.48 | 0.62 | 0.97 | 0.83 | 0.7 | | Parallel efficiency | 0.43 | 0.63 | 0.49 | 0.68 | 0.73 |
|
|
The comparison of parallel efficiency in various situations is shown in Figure 4. |