Research Article
Analysis of Parallel Multidimensional Wave Digital Filtering Network on IBM Cell Broadband Engine
Table 3
Maximum CPU runtime for different types of model.
| | Maxi. CPU Time (sec) |
Field domain: (m2)
Scheduled Time Step (group) | | Grids | Model Type | 25 (G1) | 50 (G2) | 100 (G3) | 200 (G4) |
| | Full sequential | 15.75
(4.15×) | 32.33
(4.35×) | 60.77
(4.14×) | 121.3
(3.91×) | | Left-loop parallel | 13.71
(3.61×) | 28.21
(3.79×) | 56.75
(3.86×) | 117.7
(3.79×) | | Right-loop parallel | 12.93
(3.41×) | 26.26
(3.52×) | 52.76
(3.59×) | 110.25
(3.56×) | | Full parallel | 3.79 | 7.44 | 14.69 | 31.01 |
| | Full sequential | 2931
(4.14×) | 5337
(3.77×) | 10234
(3.72×) | 20622
(3.85×) | | Left-loop parallel | 4595
(6.49×) | 9141
(6.46×) | 17273
(6.27×) | 37791
(7.05×) | | Right-loop parallel | 4053
(5.73×) | 8465
(5.98×) | 16813
(6.11×) | 35026
(6.53×) | | Full parallel | 707 | 1415 | 2753 | 5356 |
|
|
