Table 1. Running time (in seconds) for the GPU implementation, Multi-core CPU implementation, Single-thread C/C++ implementation and Single-thread R implementation when 50% of the genes were filtered out during the data preprocessing stage (individual number = 590).
| Gene Number | GPU | Multi-core CPU | Single-thread C/C++ | Single-thread WGCNA | Single-thread R |
|---|---|---|---|---|---|
| 2k | 0.624 | 0.094 | 0.375 | 1.622 | 1.295 |
| 4k | 0.671 | 0.218 | 1.138 | 5.020 | 5.101 |
| 6k | 0.811 | 0.39 | 2.403 | 14.102 | 15.163 |
| 8k | 0.967 | 0.655 | 4.447 | 25.032 | 25.740 |
| 10k | 1.202 | 0.889 | 7.301 | 41.769 | 42.260 |
| 12k | 1.388 | 1.419 | 10.811 | 69.685 | 69.420 |
| 14k | 1.747 | 1.965 | 14.633 | 94.731 | 95.220 |
| 16k | 2.122 | 3.010 | 20.369 | 130.627 | 134.100 |
Multi-core CPU version ran on 16 CPU threads running on 16 CPU cores.