Research Article
Distributed Parallel Endmember Extraction of Hyperspectral Data Based on Spark
Table 2
Execution time of the serial and parallel versions of the IEA method with Dataset 1.
| | Partition size (MB) | Initialization (s) | IEA (s) | Total (s) | Speedup (x) | Percentage of initialization |
| Serial | — | 30.17 | 2017.68 | 2047.85 | — | 1.47% | Parallel (1) | 269.17 | 12.33 | 415.80 | 428.13 | 4.78 | 2.88% | Parallel (2) | 134.58 | 9.00 | 210.00 | 219.00 | 9.35 | 4.11% | Parallel (4) | 67.29 | 8.67 | 106.13 | 114.80 | 17.84 | 7.55% | Parallel (8) | 33.65 | 7.00 | 58.40 | 65.40 | 31.31 | 10.70% |
|
|